#iceshrimpjs — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #iceshrimpjs, aggregated by home.social.
-
CW: Misskey only allows for 512 characters of alt-text which is bad for my image posts; CW: long (over 8,600 characters), Fediverse meta, Fediverse-beyond-Mastodon meta, alt-text meta, character limit meta, content warning meta
Just the other day, I found something out. Something very inconvenient about Misskey and maybe also the Forkeys.
It should be commonly known that Misskey has a local limit of 3,000 characters for posts (which it refers to as "notes"). What is not so well-known is that Misskey has a limit of about 8,000 characters, probably 8,192 or so, for inbound messages, ironically fewer than this post is long. Also, it has a limit of 512 characters for alt-text, both locally and in-bound.
Mastodon has a character limit for in-bound content, too, at least for Note-type objects (not for Article-type objects because it refuses to render them fully and links to the original instead). To my best knowledge, it rejects messages with over 100,000 characters. As for its 1,500-charater limit for alt-text, it enforces that by truncating alt-text that's longer.
Misskey, in contrast, truncates everything that exceeds its limits while still letting it in. If your post is longer than the inbound limit of ca. 8,000, all excess characters are chopped off and thrown away. If your alt-text is longer than 512 characters, all excess characters are chopped off and thrown away.
I don't know which Forkey behaves how in this regard, seeing as all Forkeys I know about have a configurable local post character limit that can be adjusted to well over 8,000. But even if the inbound limit is configurable, too, I don't think any *key admin cranks it over 60,000 or over 70,000 or over 100,000. It's simply unimaginable that someone, anyone, could ever post that much at once if your idea of the Fediverse is pure microblogging.
Also, I don't know what *key users do when they come across a truncated post or what blind or visually-impaired *key users do when they come across a truncated alt-text. Do they even suspect that it's a truncated copy of something that's longer at its source and then go check the source? Either way, it's very inconvenient.
It's especially inconvenient for me. My longest posts by a gigantic margin are image posts with original images. They always have a long image description block in the post itself that tends to be tens of thousands of characters long. It contains highly detailed visual descriptions of all images in the post. It contains all explanations necessary to understand the post, the images and the descriptions. It contains verbatim transcripts of all bits of text within the borders of the image that I can read, no matter whether or not my audience can.
In addition, each image has a shorter description in the alt-text, along with a bit that announces the long description, including where to find it. I even used to explain how to get to that description for Mastodon users for whom the summary and content warning hides the post text, but not the images, depending on which Mastodon version and frontend they use. This alone took up several hundred characters in the alt-text. All in all, I got to a point in which my alt-texts always ended up either at precisely 1,500 characters or just a few characters short.
I myself am not really bound to character limits. I used to post images here on Hubzilla where I have over 16.7 million characters for the post, including all alt-texts. Now I post them on (streams) where I have over 24 million characters. I could theoretically write alt-texts as long as I want to, seeing as, unlike on Mastodon, they aren't separate text fields; instead, they're being woven into the image-embedding markup code in the post text.
Still, I stick to a maximum of 1,500 characters for alt-text to keep Mastodon from truncating it. If you post images into the Fediverse, the main audience for your alt-text is on Mastodon, and most of them don't understand that there's something, anything, out there in the Fediverse that does not work exactly like Mastodon. And 1,500 characters can be tight already.
But if I have to stay within Misskey's limits, I can hardly post images anymore. At least not with appropriate descriptions and explanations.
Since late 2024, I have been working on-and-off on a series of fairly simple avatar portraits or rather their image descriptions. The idea is for the long description to consist of a preamble that starts with a general summary, followed by explanations, then followed by visual descriptions of what all images in the post have in common. Next come the individual descriptions of each image. Each post shall have three or four images with three or four portraits each, all in the same pose, all with only minor differences in outfits, all with a neutral, bright white background.
In addition, of course, each image shall have an alt-text, and none of the alt-texts shall depend on each other.
Now, the problem is that I have to describe three or four individual portraits in each alt-text. I'm actually struggling to squeeze such a description plus the note that announces the long description into 1,500 characters, especially if I want to fulfill Veronica Lewis a.k.a. Veronica With Four Eyes' requirements for outfit descriptions to a tee in the alt-text as well (https://veroniiiica.com/how-to-write-alt-text-for-casual-outfits/, https://veroniiiica.com/writing-image-descriptions-for-red-carpet-outfits/; see also https://veroniiiica.com/how-to-write-alt-text-image-descriptions-visually-impaired/ and https://veroniiiica.com/how-to-create-visual-descriptions/).
But in 512 characters so that even Misskey users won't get a severely truncated version? This is absolutely impossible. Even if I limit the long description announcement to some 100 characters, even if I didn't walk people through how to get to the long description, I'd have fewer than 140 characters on average to describe each individual outfit.
The long description won't fare any better. Currently, the preamble starts with some 14,000 characters of explanations, most of which are necessary to understand the visual descriptions. But when Misskey goes and truncates the post at the 8,000-something mark, Misskey users won't even get to any visual description because all visual descriptions would be chopped off.
What makes matters worse is that the preamble grows the longer, the easier to understand I make it and the less I leave people with unexplained technical or jargon terms which you shouldn't use in image descriptions at all anyway. So the next time I go through it and rewrite it to make it easier to understand, I'll also make it even longer than it already is.
But what if I simply cut all the explanations? For one, I'd leave people to their own devices to understand extremely obscure niche content. They won't. My explanations aren't 14,000 characters long because I've artificially inflated them, but because there is so much to know before you understand the post and the images and the descriptions.
Besides, the visual descriptions alone won't fit into 8,192 characters either. What I currently have is over 5,000 characters of common visual description for all portraits in all images plus about 2,500 characters of individual visual description for the three portraits in the first image. That's over 7,500 characters altogether already. And I still have to describe nine portraits in another three images. The post will end up with some 15,000 characters of visual descriptions unless they grow longer when I simplify them again.
I guess users of Misskey or any Forkey will still have to put up with truncated alt-texts and truncated long descriptions in the future. But my future image posts will contain a paragraph at the beginning that explains that the post and/or the alt-text may be truncated on Misskey and the Forkeys, and that both are uncut at the source. Still, this means that *key users will have to put up with the extra hassle of opening my original post at a source with a quite cumbersome UI. And I've got my doubts that this UI is really accessible.
Unfortunately, this also means that *key users won't get any hashtags along with these posts. But then again, the handling of Identi.ca-style/Friendica-style hashtags with the number sign outside the link is broken on all *keys and will remain so for the foreseeable future.
#Long #LongPost #CWLong #CWLongPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #Misskey #Forkey #Forkeys #Calckey #Firefish #Sharkey #CherryPick #Iceshrimp #Iceshrimp-JS #AltText #AltTextMeta #CWAltTextMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta #CW #CWs #CWMeta #ContentWarning #ContentWarnings #ContentWarningMeta #CharacterLimit #CharacterLimits #CharacterLimitMeta #CWCharacterLimitMeta #A11y #Accessibility -
CW: Misskey only allows for 512 characters of alt-text which is bad for my image posts; CW: long (over 8,600 characters), Fediverse meta, Fediverse-beyond-Mastodon meta, alt-text meta, character limit meta, content warning meta
Just the other day, I found something out. Something very inconvenient about Misskey and maybe also the Forkeys.
It should be commonly known that Misskey has a local limit of 3,000 characters for posts (which it refers to as "notes"). What is not so well-known is that Misskey has a limit of about 8,000 characters, probably 8,192 or so, for inbound messages, ironically fewer than this post is long. Also, it has a limit of 512 characters for alt-text, both locally and in-bound.
Mastodon has a character limit for in-bound content, too, at least for Note-type objects (not for Article-type objects because it refuses to render them fully and links to the original instead). To my best knowledge, it rejects messages with over 100,000 characters. As for its 1,500-charater limit for alt-text, it enforces that by truncating alt-text that's longer.
Misskey, in contrast, truncates everything that exceeds its limits while still letting it in. If your post is longer than the inbound limit of ca. 8,000, all excess characters are chopped off and thrown away. If your alt-text is longer than 512 characters, all excess characters are chopped off and thrown away.
I don't know which Forkey behaves how in this regard, seeing as all Forkeys I know about have a configurable local post character limit that can be adjusted to well over 8,000. But even if the inbound limit is configurable, too, I don't think any *key admin cranks it over 60,000 or over 70,000 or over 100,000. It's simply unimaginable that someone, anyone, could ever post that much at once if your idea of the Fediverse is pure microblogging.
Also, I don't know what *key users do when they come across a truncated post or what blind or visually-impaired *key users do when they come across a truncated alt-text. Do they even suspect that it's a truncated copy of something that's longer at its source and then go check the source? Either way, it's very inconvenient.
It's especially inconvenient for me. My longest posts by a gigantic margin are image posts with original images. They always have a long image description block in the post itself that tends to be tens of thousands of characters long. It contains highly detailed visual descriptions of all images in the post. It contains all explanations necessary to understand the post, the images and the descriptions. It contains verbatim transcripts of all bits of text within the borders of the image that I can read, no matter whether or not my audience can.
In addition, each image has a shorter description in the alt-text, along with a bit that announces the long description, including where to find it. I even used to explain how to get to that description for Mastodon users for whom the summary and content warning hides the post text, but not the images, depending on which Mastodon version and frontend they use. This alone took up several hundred characters in the alt-text. All in all, I got to a point in which my alt-texts always ended up either at precisely 1,500 characters or just a few characters short.
I myself am not really bound to character limits. I used to post images here on Hubzilla where I have over 16.7 million characters for the post, including all alt-texts. Now I post them on (streams) where I have over 24 million characters. I could theoretically write alt-texts as long as I want to, seeing as, unlike on Mastodon, they aren't separate text fields; instead, they're being woven into the image-embedding markup code in the post text.
Still, I stick to a maximum of 1,500 characters for alt-text to keep Mastodon from truncating it. If you post images into the Fediverse, the main audience for your alt-text is on Mastodon, and most of them don't understand that there's something, anything, out there in the Fediverse that does not work exactly like Mastodon. And 1,500 characters can be tight already.
But if I have to stay within Misskey's limits, I can hardly post images anymore. At least not with appropriate descriptions and explanations.
Since late 2024, I have been working on-and-off on a series of fairly simple avatar portraits or rather their image descriptions. The idea is for the long description to consist of a preamble that starts with a general summary, followed by explanations, then followed by visual descriptions of what all images in the post have in common. Next come the individual descriptions of each image. Each post shall have three or four images with three or four portraits each, all in the same pose, all with only minor differences in outfits, all with a neutral, bright white background.
In addition, of course, each image shall have an alt-text, and none of the alt-texts shall depend on each other.
Now, the problem is that I have to describe three or four individual portraits in each alt-text. I'm actually struggling to squeeze such a description plus the note that announces the long description into 1,500 characters, especially if I want to fulfill Veronica Lewis a.k.a. Veronica With Four Eyes' requirements for outfit descriptions to a tee in the alt-text as well (https://veroniiiica.com/how-to-write-alt-text-for-casual-outfits/, https://veroniiiica.com/writing-image-descriptions-for-red-carpet-outfits/; see also https://veroniiiica.com/how-to-write-alt-text-image-descriptions-visually-impaired/ and https://veroniiiica.com/how-to-create-visual-descriptions/).
But in 512 characters so that even Misskey users won't get a severely truncated version? This is absolutely impossible. Even if I limit the long description announcement to some 100 characters, even if I didn't walk people through how to get to the long description, I'd have fewer than 140 characters on average to describe each individual outfit.
The long description won't fare any better. Currently, the preamble starts with some 14,000 characters of explanations, most of which are necessary to understand the visual descriptions. But when Misskey goes and truncates the post at the 8,000-something mark, Misskey users won't even get to any visual description because all visual descriptions would be chopped off.
What makes matters worse is that the preamble grows the longer, the easier to understand I make it and the less I leave people with unexplained technical or jargon terms which you shouldn't use in image descriptions at all anyway. So the next time I go through it and rewrite it to make it easier to understand, I'll also make it even longer than it already is.
But what if I simply cut all the explanations? For one, I'd leave people to their own devices to understand extremely obscure niche content. They won't. My explanations aren't 14,000 characters long because I've artificially inflated them, but because there is so much to know before you understand the post and the images and the descriptions.
Besides, the visual descriptions alone won't fit into 8,192 characters either. What I currently have is over 5,000 characters of common visual description for all portraits in all images plus about 2,500 characters of individual visual description for the three portraits in the first image. That's over 7,500 characters altogether already. And I still have to describe nine portraits in another three images. The post will end up with some 15,000 characters of visual descriptions unless they grow longer when I simplify them again.
I guess users of Misskey or any Forkey will still have to put up with truncated alt-texts and truncated long descriptions in the future. But my future image posts will contain a paragraph at the beginning that explains that the post and/or the alt-text may be truncated on Misskey and the Forkeys, and that both are uncut at the source. Still, this means that *key users will have to put up with the extra hassle of opening my original post at a source with a quite cumbersome UI. And I've got my doubts that this UI is really accessible.
Unfortunately, this also means that *key users won't get any hashtags along with these posts. But then again, the handling of Identi.ca-style/Friendica-style hashtags with the number sign outside the link is broken on all *keys and will remain so for the foreseeable future.
#Long #LongPost #CWLong #CWLongPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #Misskey #Forkey #Forkeys #Calckey #Firefish #Sharkey #CherryPick #Iceshrimp #Iceshrimp-JS #AltText #AltTextMeta #CWAltTextMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta #CW #CWs #CWMeta #ContentWarning #ContentWarnings #ContentWarningMeta #CharacterLimit #CharacterLimits #CharacterLimitMeta #CWCharacterLimitMeta #A11y #Accessibility -
CW: Misskey only allows for 512 characters of alt-text which is bad for my image posts; CW: long (over 8,600 characters), Fediverse meta, Fediverse-beyond-Mastodon meta, alt-text meta, character limit meta, content warning meta
Just the other day, I found something out. Something very inconvenient about Misskey and maybe also the Forkeys.
It should be commonly known that Misskey has a local limit of 3,000 characters for posts (which it refers to as "notes"). What is not so well-known is that Misskey has a limit of about 8,000 characters, probably 8,192 or so, for inbound messages, ironically fewer than this post is long. Also, it has a limit of 512 characters for alt-text, both locally and in-bound.
Mastodon has a character limit for in-bound content, too, at least for Note-type objects (not for Article-type objects because it refuses to render them fully and links to the original instead). To my best knowledge, it rejects messages with over 100,000 characters. As for its 1,500-charater limit for alt-text, it enforces that by truncating alt-text that's longer.
Misskey, in contrast, truncates everything that exceeds its limits while still letting it in. If your post is longer than the inbound limit of ca. 8,000, all excess characters are chopped off and thrown away. If your alt-text is longer than 512 characters, all excess characters are chopped off and thrown away.
I don't know which Forkey behaves how in this regard, seeing as all Forkeys I know about have a configurable local post character limit that can be adjusted to well over 8,000. But even if the inbound limit is configurable, too, I don't think any *key admin cranks it over 60,000 or over 70,000 or over 100,000. It's simply unimaginable that someone, anyone, could ever post that much at once if your idea of the Fediverse is pure microblogging.
Also, I don't know what *key users do when they come across a truncated post or what blind or visually-impaired *key users do when they come across a truncated alt-text. Do they even suspect that it's a truncated copy of something that's longer at its source and then go check the source? Either way, it's very inconvenient.
It's especially inconvenient for me. My longest posts by a gigantic margin are image posts with original images. They always have a long image description block in the post itself that tends to be tens of thousands of characters long. It contains highly detailed visual descriptions of all images in the post. It contains all explanations necessary to understand the post, the images and the descriptions. It contains verbatim transcripts of all bits of text within the borders of the image that I can read, no matter whether or not my audience can.
In addition, each image has a shorter description in the alt-text, along with a bit that announces the long description, including where to find it. I even used to explain how to get to that description for Mastodon users for whom the summary and content warning hides the post text, but not the images, depending on which Mastodon version and frontend they use. This alone took up several hundred characters in the alt-text. All in all, I got to a point in which my alt-texts always ended up either at precisely 1,500 characters or just a few characters short.
I myself am not really bound to character limits. I used to post images here on Hubzilla where I have over 16.7 million characters for the post, including all alt-texts. Now I post them on (streams) where I have over 24 million characters. I could theoretically write alt-texts as long as I want to, seeing as, unlike on Mastodon, they aren't separate text fields; instead, they're being woven into the image-embedding markup code in the post text.
Still, I stick to a maximum of 1,500 characters for alt-text to keep Mastodon from truncating it. If you post images into the Fediverse, the main audience for your alt-text is on Mastodon, and most of them don't understand that there's something, anything, out there in the Fediverse that does not work exactly like Mastodon. And 1,500 characters can be tight already.
But if I have to stay within Misskey's limits, I can hardly post images anymore. At least not with appropriate descriptions and explanations.
Since late 2024, I have been working on-and-off on a series of fairly simple avatar portraits or rather their image descriptions. The idea is for the long description to consist of a preamble that starts with a general summary, followed by explanations, then followed by visual descriptions of what all images in the post have in common. Next come the individual descriptions of each image. Each post shall have three or four images with three or four portraits each, all in the same pose, all with only minor differences in outfits, all with a neutral, bright white background.
In addition, of course, each image shall have an alt-text, and none of the alt-texts shall depend on each other.
Now, the problem is that I have to describe three or four individual portraits in each alt-text. I'm actually struggling to squeeze such a description plus the note that announces the long description into 1,500 characters, especially if I want to fulfill Veronica Lewis a.k.a. Veronica With Four Eyes' requirements for outfit descriptions to a tee in the alt-text as well (https://veroniiiica.com/how-to-write-alt-text-for-casual-outfits/, https://veroniiiica.com/writing-image-descriptions-for-red-carpet-outfits/; see also https://veroniiiica.com/how-to-write-alt-text-image-descriptions-visually-impaired/ and https://veroniiiica.com/how-to-create-visual-descriptions/).
But in 512 characters so that even Misskey users won't get a severely truncated version? This is absolutely impossible. Even if I limit the long description announcement to some 100 characters, even if I didn't walk people through how to get to the long description, I'd have fewer than 140 characters on average to describe each individual outfit.
The long description won't fare any better. Currently, the preamble starts with some 14,000 characters of explanations, most of which are necessary to understand the visual descriptions. But when Misskey goes and truncates the post at the 8,000-something mark, Misskey users won't even get to any visual description because all visual descriptions would be chopped off.
What makes matters worse is that the preamble grows the longer, the easier to understand I make it and the less I leave people with unexplained technical or jargon terms which you shouldn't use in image descriptions at all anyway. So the next time I go through it and rewrite it to make it easier to understand, I'll also make it even longer than it already is.
But what if I simply cut all the explanations? For one, I'd leave people to their own devices to understand extremely obscure niche content. They won't. My explanations aren't 14,000 characters long because I've artificially inflated them, but because there is so much to know before you understand the post and the images and the descriptions.
Besides, the visual descriptions alone won't fit into 8,192 characters either. What I currently have is over 5,000 characters of common visual description for all portraits in all images plus about 2,500 characters of individual visual description for the three portraits in the first image. That's over 7,500 characters altogether already. And I still have to describe nine portraits in another three images. The post will end up with some 15,000 characters of visual descriptions unless they grow longer when I simplify them again.
I guess users of Misskey or any Forkey will still have to put up with truncated alt-texts and truncated long descriptions in the future. But my future image posts will contain a paragraph at the beginning that explains that the post and/or the alt-text may be truncated on Misskey and the Forkeys, and that both are uncut at the source. Still, this means that *key users will have to put up with the extra hassle of opening my original post at a source with a quite cumbersome UI. And I've got my doubts that this UI is really accessible.
Unfortunately, this also means that *key users won't get any hashtags along with these posts. But then again, the handling of Identi.ca-style/Friendica-style hashtags with the number sign outside the link is broken on all *keys and will remain so for the foreseeable future.
#Long #LongPost #CWLong #CWLongPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #Misskey #Forkey #Forkeys #Calckey #Firefish #Sharkey #CherryPick #Iceshrimp #Iceshrimp-JS #AltText #AltTextMeta #CWAltTextMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta #CW #CWs #CWMeta #ContentWarning #ContentWarnings #ContentWarningMeta #CharacterLimit #CharacterLimits #CharacterLimitMeta #CWCharacterLimitMeta #A11y #Accessibility -
CW: Misskey only allows for 512 characters of alt-text which is bad for my image posts; CW: long (over 8,600 characters), Fediverse meta, Fediverse-beyond-Mastodon meta, alt-text meta, character limit meta, content warning meta
Just the other day, I found something out. Something very inconvenient about Misskey and maybe also the Forkeys.
It should be commonly known that Misskey has a local limit of 3,000 characters for posts (which it refers to as "notes"). What is not so well-known is that Misskey has a limit of about 8,000 characters, probably 8,192 or so, for inbound messages, ironically fewer than this post is long. Also, it has a limit of 512 characters for alt-text, both locally and in-bound.
Mastodon has a character limit for in-bound content, too, at least for Note-type objects (not for Article-type objects because it refuses to render them fully and links to the original instead). To my best knowledge, it rejects messages with over 100,000 characters. As for its 1,500-charater limit for alt-text, it enforces that by truncating alt-text that's longer.
Misskey, in contrast, truncates everything that exceeds its limits while still letting it in. If your post is longer than the inbound limit of ca. 8,000, all excess characters are chopped off and thrown away. If your alt-text is longer than 512 characters, all excess characters are chopped off and thrown away.
I don't know which Forkey behaves how in this regard, seeing as all Forkeys I know about have a configurable local post character limit that can be adjusted to well over 8,000. But even if the inbound limit is configurable, too, I don't think any *key admin cranks it over 60,000 or over 70,000 or over 100,000. It's simply unimaginable that someone, anyone, could ever post that much at once if your idea of the Fediverse is pure microblogging.
Also, I don't know what *key users do when they come across a truncated post or what blind or visually-impaired *key users do when they come across a truncated alt-text. Do they even suspect that it's a truncated copy of something that's longer at its source and then go check the source? Either way, it's very inconvenient.
It's especially inconvenient for me. My longest posts by a gigantic margin are image posts with original images. They always have a long image description block in the post itself that tends to be tens of thousands of characters long. It contains highly detailed visual descriptions of all images in the post. It contains all explanations necessary to understand the post, the images and the descriptions. It contains verbatim transcripts of all bits of text within the borders of the image that I can read, no matter whether or not my audience can.
In addition, each image has a shorter description in the alt-text, along with a bit that announces the long description, including where to find it. I even used to explain how to get to that description for Mastodon users for whom the summary and content warning hides the post text, but not the images, depending on which Mastodon version and frontend they use. This alone took up several hundred characters in the alt-text. All in all, I got to a point in which my alt-texts always ended up either at precisely 1,500 characters or just a few characters short.
I myself am not really bound to character limits. I used to post images here on Hubzilla where I have over 16.7 million characters for the post, including all alt-texts. Now I post them on (streams) where I have over 24 million characters. I could theoretically write alt-texts as long as I want to, seeing as, unlike on Mastodon, they aren't separate text fields; instead, they're being woven into the image-embedding markup code in the post text.
Still, I stick to a maximum of 1,500 characters for alt-text to keep Mastodon from truncating it. If you post images into the Fediverse, the main audience for your alt-text is on Mastodon, and most of them don't understand that there's something, anything, out there in the Fediverse that does not work exactly like Mastodon. And 1,500 characters can be tight already.
But if I have to stay within Misskey's limits, I can hardly post images anymore. At least not with appropriate descriptions and explanations.
Since late 2024, I have been working on-and-off on a series of fairly simple avatar portraits or rather their image descriptions. The idea is for the long description to consist of a preamble that starts with a general summary, followed by explanations, then followed by visual descriptions of what all images in the post have in common. Next come the individual descriptions of each image. Each post shall have three or four images with three or four portraits each, all in the same pose, all with only minor differences in outfits, all with a neutral, bright white background.
In addition, of course, each image shall have an alt-text, and none of the alt-texts shall depend on each other.
Now, the problem is that I have to describe three or four individual portraits in each alt-text. I'm actually struggling to squeeze such a description plus the note that announces the long description into 1,500 characters, especially if I want to fulfill Veronica Lewis a.k.a. Veronica With Four Eyes' requirements for outfit descriptions to a tee in the alt-text as well (https://veroniiiica.com/how-to-write-alt-text-for-casual-outfits/, https://veroniiiica.com/writing-image-descriptions-for-red-carpet-outfits/; see also https://veroniiiica.com/how-to-write-alt-text-image-descriptions-visually-impaired/ and https://veroniiiica.com/how-to-create-visual-descriptions/).
But in 512 characters so that even Misskey users won't get a severely truncated version? This is absolutely impossible. Even if I limit the long description announcement to some 100 characters, even if I didn't walk people through how to get to the long description, I'd have fewer than 140 characters on average to describe each individual outfit.
The long description won't fare any better. Currently, the preamble starts with some 14,000 characters of explanations, most of which are necessary to understand the visual descriptions. But when Misskey goes and truncates the post at the 8,000-something mark, Misskey users won't even get to any visual description because all visual descriptions would be chopped off.
What makes matters worse is that the preamble grows the longer, the easier to understand I make it and the less I leave people with unexplained technical or jargon terms which you shouldn't use in image descriptions at all anyway. So the next time I go through it and rewrite it to make it easier to understand, I'll also make it even longer than it already is.
But what if I simply cut all the explanations? For one, I'd leave people to their own devices to understand extremely obscure niche content. They won't. My explanations aren't 14,000 characters long because I've artificially inflated them, but because there is so much to know before you understand the post and the images and the descriptions.
Besides, the visual descriptions alone won't fit into 8,192 characters either. What I currently have is over 5,000 characters of common visual description for all portraits in all images plus about 2,500 characters of individual visual description for the three portraits in the first image. That's over 7,500 characters altogether already. And I still have to describe nine portraits in another three images. The post will end up with some 15,000 characters of visual descriptions unless they grow longer when I simplify them again.
I guess users of Misskey or any Forkey will still have to put up with truncated alt-texts and truncated long descriptions in the future. But my future image posts will contain a paragraph at the beginning that explains that the post and/or the alt-text may be truncated on Misskey and the Forkeys, and that both are uncut at the source. Still, this means that *key users will have to put up with the extra hassle of opening my original post at a source with a quite cumbersome UI. And I've got my doubts that this UI is really accessible.
Unfortunately, this also means that *key users won't get any hashtags along with these posts. But then again, the handling of Identi.ca-style/Friendica-style hashtags with the number sign outside the link is broken on all *keys and will remain so for the foreseeable future.
#Long #LongPost #CWLong #CWLongPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #Misskey #Forkey #Forkeys #Calckey #Firefish #Sharkey #CherryPick #Iceshrimp #Iceshrimp-JS #AltText #AltTextMeta #CWAltTextMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta #CW #CWs #CWMeta #ContentWarning #ContentWarnings #ContentWarningMeta #CharacterLimit #CharacterLimits #CharacterLimitMeta #CWCharacterLimitMeta #A11y #Accessibility -
CW: Misskey only allows for 512 characters of alt-text which is bad for my image posts; CW: long (over 8,600 characters), Fediverse meta, Fediverse-beyond-Mastodon meta, alt-text meta, character limit meta, content warning meta
Just the other day, I found something out. Something very inconvenient about Misskey and maybe also the Forkeys.
It should be commonly known that Misskey has a local limit of 3,000 characters for posts (which it refers to as "notes"). What is not so well-known is that Misskey has a limit of about 8,000 characters, probably 8,192 or so, for inbound messages, ironically fewer than this post is long. Also, it has a limit of 512 characters for alt-text, both locally and in-bound.
Mastodon has a character limit for in-bound content, too, at least for Note-type objects (not for Article-type objects because it refuses to render them fully and links to the original instead). To my best knowledge, it rejects messages with over 100,000 characters. As for its 1,500-charater limit for alt-text, it enforces that by truncating alt-text that's longer.
Misskey, in contrast, truncates everything that exceeds its limits while still letting it in. If your post is longer than the inbound limit of ca. 8,000, all excess characters are chopped off and thrown away. If your alt-text is longer than 512 characters, all excess characters are chopped off and thrown away.
I don't know which Forkey behaves how in this regard, seeing as all Forkeys I know about have a configurable local post character limit that can be adjusted to well over 8,000. But even if the inbound limit is configurable, too, I don't think any *key admin cranks it over 60,000 or over 70,000 or over 100,000. It's simply unimaginable that someone, anyone, could ever post that much at once if your idea of the Fediverse is pure microblogging.
Also, I don't know what *key users do when they come across a truncated post or what blind or visually-impaired *key users do when they come across a truncated alt-text. Do they even suspect that it's a truncated copy of something that's longer at its source and then go check the source? Either way, it's very inconvenient.
It's especially inconvenient for me. My longest posts by a gigantic margin are image posts with original images. They always have a long image description block in the post itself that tends to be tens of thousands of characters long. It contains highly detailed visual descriptions of all images in the post. It contains all explanations necessary to understand the post, the images and the descriptions. It contains verbatim transcripts of all bits of text within the borders of the image that I can read, no matter whether or not my audience can.
In addition, each image has a shorter description in the alt-text, along with a bit that announces the long description, including where to find it. I even used to explain how to get to that description for Mastodon users for whom the summary and content warning hides the post text, but not the images, depending on which Mastodon version and frontend they use. This alone took up several hundred characters in the alt-text. All in all, I got to a point in which my alt-texts always ended up either at precisely 1,500 characters or just a few characters short.
I myself am not really bound to character limits. I used to post images here on Hubzilla where I have over 16.7 million characters for the post, including all alt-texts. Now I post them on (streams) where I have over 24 million characters. I could theoretically write alt-texts as long as I want to, seeing as, unlike on Mastodon, they aren't separate text fields; instead, they're being woven into the image-embedding markup code in the post text.
Still, I stick to a maximum of 1,500 characters for alt-text to keep Mastodon from truncating it. If you post images into the Fediverse, the main audience for your alt-text is on Mastodon, and most of them don't understand that there's something, anything, out there in the Fediverse that does not work exactly like Mastodon. And 1,500 characters can be tight already.
But if I have to stay within Misskey's limits, I can hardly post images anymore. At least not with appropriate descriptions and explanations.
Since late 2024, I have been working on-and-off on a series of fairly simple avatar portraits or rather their image descriptions. The idea is for the long description to consist of a preamble that starts with a general summary, followed by explanations, then followed by visual descriptions of what all images in the post have in common. Next come the individual descriptions of each image. Each post shall have three or four images with three or four portraits each, all in the same pose, all with only minor differences in outfits, all with a neutral, bright white background.
In addition, of course, each image shall have an alt-text, and none of the alt-texts shall depend on each other.
Now, the problem is that I have to describe three or four individual portraits in each alt-text. I'm actually struggling to squeeze such a description plus the note that announces the long description into 1,500 characters, especially if I want to fulfill Veronica Lewis a.k.a. Veronica With Four Eyes' requirements for outfit descriptions to a tee in the alt-text as well (https://veroniiiica.com/how-to-write-alt-text-for-casual-outfits/, https://veroniiiica.com/writing-image-descriptions-for-red-carpet-outfits/; see also https://veroniiiica.com/how-to-write-alt-text-image-descriptions-visually-impaired/ and https://veroniiiica.com/how-to-create-visual-descriptions/).
But in 512 characters so that even Misskey users won't get a severely truncated version? This is absolutely impossible. Even if I limit the long description announcement to some 100 characters, even if I didn't walk people through how to get to the long description, I'd have fewer than 140 characters on average to describe each individual outfit.
The long description won't fare any better. Currently, the preamble starts with some 14,000 characters of explanations, most of which are necessary to understand the visual descriptions. But when Misskey goes and truncates the post at the 8,000-something mark, Misskey users won't even get to any visual description because all visual descriptions would be chopped off.
What makes matters worse is that the preamble grows the longer, the easier to understand I make it and the less I leave people with unexplained technical or jargon terms which you shouldn't use in image descriptions at all anyway. So the next time I go through it and rewrite it to make it easier to understand, I'll also make it even longer than it already is.
But what if I simply cut all the explanations? For one, I'd leave people to their own devices to understand extremely obscure niche content. They won't. My explanations aren't 14,000 characters long because I've artificially inflated them, but because there is so much to know before you understand the post and the images and the descriptions.
Besides, the visual descriptions alone won't fit into 8,192 characters either. What I currently have is over 5,000 characters of common visual description for all portraits in all images plus about 2,500 characters of individual visual description for the three portraits in the first image. That's over 7,500 characters altogether already. And I still have to describe nine portraits in another three images. The post will end up with some 15,000 characters of visual descriptions unless they grow longer when I simplify them again.
I guess users of Misskey or any Forkey will still have to put up with truncated alt-texts and truncated long descriptions in the future. But my future image posts will contain a paragraph at the beginning that explains that the post and/or the alt-text may be truncated on Misskey and the Forkeys, and that both are uncut at the source. Still, this means that *key users will have to put up with the extra hassle of opening my original post at a source with a quite cumbersome UI. And I've got my doubts that this UI is really accessible.
Unfortunately, this also means that *key users won't get any hashtags along with these posts. But then again, the handling of Identi.ca-style/Friendica-style hashtags with the number sign outside the link is broken on all *keys and will remain so for the foreseeable future.
#Long #LongPost #CWLong #CWLongPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #Misskey #Forkey #Forkeys #Calckey #Firefish #Sharkey #CherryPick #Iceshrimp #Iceshrimp-JS #AltText #AltTextMeta #CWAltTextMeta #ImageDescription #ImageDescriptions #ImageDescriptionMeta #CWImageDescriptionMeta #CW #CWs #CWMeta #ContentWarning #ContentWarnings #ContentWarningMeta #CharacterLimit #CharacterLimits #CharacterLimitMeta #CWCharacterLimitMeta #A11y #Accessibility -
Hallo liebe Fediverse-Admins,
ich bin auf der Suche nach einer Botpolicy für Anubis in Kombination mit Iceshrimp.
Meine aktuelle Policy ist folgende: paste.obco.pro/upload/dRGDKe
Jedoch scheint damit das Iceshrimp UI nicht richtig zu laden. Also der Feed und die Benachrichtigungen laden nicht richtig.
Hier ist der Browserlog: paste.obco.pro/upload/er0w4b
Wenn jemand eine Idee oder eine Policy hat, welche mit Iceshrimp funktioniert, wäre ich sehr dankbar.
:boost_animated: :blahaj: :rosahaj_thinking:
#fedimin #fedimins #iceshrimp #iceshrimpjs #anubis -
Hallo liebe Fediverse-Admins,
ich bin auf der Suche nach einer Botpolicy für Anubis in Kombination mit Iceshrimp.
Meine aktuelle Policy ist folgende: paste.obco.pro/upload/dRGDKe
Jedoch scheint damit das Iceshrimp UI nicht richtig zu laden. Also der Feed und die Benachrichtigungen laden nicht richtig.
Hier ist der Browserlog: paste.obco.pro/upload/er0w4b
Wenn jemand eine Idee oder eine Policy hat, welche mit Iceshrimp funktioniert, wäre ich sehr dankbar.
:boost_animated: :blahaj: :rosahaj_thinking:
#fedimin #fedimins #iceshrimp #iceshrimpjs #anubis -
Hallo liebe Fediverse-Admins,
ich bin auf der Suche nach einer Botpolicy für Anubis in Kombination mit Iceshrimp.
Meine aktuelle Policy ist folgende: paste.obco.pro/upload/dRGDKe
Jedoch scheint damit das Iceshrimp UI nicht richtig zu laden. Also der Feed und die Benachrichtigungen laden nicht richtig.
Hier ist der Browserlog: paste.obco.pro/upload/er0w4b
Wenn jemand eine Idee oder eine Policy hat, welche mit Iceshrimp funktioniert, wäre ich sehr dankbar.
:boost_animated: :blahaj: :rosahaj_thinking:
#fedimin #fedimins #iceshrimp #iceshrimpjs #anubis -
Hallo liebe Fediverse-Admins,
ich bin auf der Suche nach einer Botpolicy für Anubis in Kombination mit Iceshrimp.
Meine aktuelle Policy ist folgende: paste.obco.pro/upload/dRGDKe
Jedoch scheint damit das Iceshrimp UI nicht richtig zu laden. Also der Feed und die Benachrichtigungen laden nicht richtig.
Hier ist der Browserlog: paste.obco.pro/upload/er0w4b
Wenn jemand eine Idee oder eine Policy hat, welche mit Iceshrimp funktioniert, wäre ich sehr dankbar.
:boost_animated: :blahaj: :rosahaj_thinking:
#fedimin #fedimins #iceshrimp #iceshrimpjs #anubis -
Hallo liebe Fediverse-Admins,
ich bin auf der Suche nach einer Botpolicy für Anubis in Kombination mit Iceshrimp.
Meine aktuelle Policy ist folgende: paste.obco.pro/upload/dRGDKe
Jedoch scheint damit das Iceshrimp UI nicht richtig zu laden. Also der Feed und die Benachrichtigungen laden nicht richtig.
Hier ist der Browserlog: paste.obco.pro/upload/er0w4b
Wenn jemand eine Idee oder eine Policy hat, welche mit Iceshrimp funktioniert, wäre ich sehr dankbar.
:boost_animated: :blahaj: :rosahaj_thinking:
#fedimin #fedimins #iceshrimp #iceshrimpjs #anubis -
@「 Jürgen 」:fedi_mastodon: @crossgolf_rebel - kostenlose Kwalitätsposts @Don di Dislessia Soweit ich weiß, war es so (man möge mich wiederum korrigieren; kursiv sind die "Forkeys höheren Grades", die nicht von Misskey geforkt wurden):
- Misskey
Der Ursprung in Japan. - Calckey
Soft-Fork von Misskey mit einigen Extrafeatures.
Entwickler hatte irgendwann keine Zeit/keinen Bock mehr. - Firefish
Fortführung von Calckey unter neuem Management. Wurde aufgrund einer massiven Werbeaktion eines begeisterten Nutzers mit viel Reichweite so populär, daß der Name "Calckey" irgendwann einfach doof war und das ganze Ding eine neue Identität bekam.
Entwickler verschwand irgendwann sang- und klanglos von der Bildfläche. Nach einem halben Jahr stellte sich raus: Entwickler hatte wegen Abschlußarbeit usw. keine Zeit mehr, nicht mal, sich zu verabschieden.
Wurde unter neuer Führung mit neuem Repository und neuer Leuchtturminstanz weitergeführt, aber ohne neue Website. Wurde wieder eingestellt, weil es für eine einzige Entwicklerin viel zuviel war, die ganzen alten Co-Entwickler von Firefish alle zu Iceshrimp gewechselt waren und keine neuen Mitentwickler angeheuert werden konnten. - Iceshrimp
Fork von Firefish von ehemaligen Firefish-Entwicklern. Rebased nach Misskey, weil es auf Firefish nicht weiterging. Erklärtes Ziel war, Stabilität über Calckeys Featuritis zu stellen.
Weil Misskeys Codebase an sich einiges an grundsätzlichen Macken hatte, wurde beschlossen, es ist einfacher, das ganze Zeugs von Grund auf neu zu schreiben, als zu versuchen, das alles auszubügeln. Und bei der Gelegenheit wollte man von JavaScript (TypeScript und Vue.js) weg. Also hat man angefangen, das ganze Ding in C# neu zu schreiben als Iceshrimp.NET. Ziel ist featuremäßige Deckungsgleichheit mit dem bisherigen Iceshrimp und gleichzeitig Anpassung an Mastodon. Iceshrimp.NET ist noch sehr unfertig.
Bei der Gelegenheit wurde das alte Iceshrimp in Iceshrimp-JS umbenannt und in den Wartungsmodus versetzt = gibt keine Weiterentwicklung und keine neuen Features mehr, nur Sicherheitspatches und evtl. Bugfixes. - Catodon
Soft-Fork von Firefish. Ziel war, ein Forkey zu haben, das sich in der Bedienung wie Mastodon anfühlt. Quasi Mastodon-Fassade mit *key-Struktur dahinter. Also ein Forkey für den möglichst leichten Umstieg von Mastodon nach *key.
Wurde, als Firefish endgültig eingestellt wurde und damit die Basis tot war, nach Iceshrimp rebased.
Nachdem Iceshrimp zu Iceshrimp-JS wurde und in den Wartungsmodus ging, hatte Catodon ein Problem: Man kann nicht einfach von etwas, das in TypeScript und Vue.js geschrieben ist, nach etwas rebasen, das in C# geschrieben ist (Iceshrimp.NET). Und zur Weiterentwicklung wäre ein Rebase nötig gewesen, weil man jetzt wieder eine fast tote Basis hatte. Aber nach Misskey wollte man nicht (siehe Hajkey weiter unten), und Sharkey war einerseits zu mächtig und andererseits zu kapriziös.
Inzwischen gehen Catodon die Instanzen ein. Die meisten dürften geschlossen worden sein. - Sharkey
Soft-Fork von Misskey mit dem Ziel, so ziemlich alles, was je für Forkeys an Features entwickelt worden war, in einen einzigen Forkey zu packen und noch mehr obendrauf.
Quasi das Yang zu Iceshrimps Yin. Machte lange den Eindruck, als gingen Features über Zuverlässigkeit. Sharkeys Implementation der Mastodon Client API war schon regelrecht legendär schlecht, und alle hofften auf den Retter, der daherkommt und die Implementation von Grund auf neu schreibt, weil sie als unrettbar im Eimer galt. Inzwischen soll Sharkey seine Zuverlässigkeit im Griff haben, zumindest insofern, wie auch Misskey zuverlässig ist.
Die Entwickler haben eine zweifelhafte Reputation. Unter anderem haben sie Crowdfunding-Gelder für einen Sharkey-Server gesammelt und von dem Geld einen Minecraft-Server aufgezogen, aber keinen Sharkey-Server. Mitunter wird deshalb zum Boykott von Sharkey aufgerufen. - CherryPick
Südkoreanischer Soft-Fork von Misskey. Tatsächlich älter als Sharkey, mindestens von 2021, aber später wurde wohl einiges von Sharkey nach CherryPick portiert. Das Ziel war, einen stabilen und zuverlässigen Forkey zu haben ohne die Macken von Sharkey und sogar ohne die Macken von Misskey, der aber gleichzeitig gut Features hat. Das ist wohl sogar weitestgehend gelungen.
Vom ästhetischen Stil her so ähnlich, wie Misskey früher mal war, also sehr auf den japanischen bzw. südkoreanischen Geschmack ausgelegt: bunt, grell, poppig, genki, kawaii. Alleine das zeigt, daß CherryPick von Misskey geforkt wurde, bevor Misskey gänzlich an den westlichen Geschmack angepaßt wurde.
Unklar, ob es Entwickler hat, die Englisch verstehen bzw. schreiben können; falls nicht, dann als Soft-Fork-Basis ungeeignet. Wäre ohne die Kommunikationsbarriere vermutlich der ultimative *key.
Mein letzter Stand: Bis auf eine Instanz im Großraum Washington, D.C. gibt es CherryPick-Instanzen nur in den Großräumen Tokyo und Seoul. War deshalb lange Zeit im westlichen Fediverse fast unbekannt. - FoundKey
Wohl der erste Forkey, der in PascalCase geschrieben wurde.
Meines Wissens direkter Soft-Fork von Misskey. Wird tatsächlich (oh Wunder) kleckerweise gepflegt, aber nicht für Instanzen mit mehr als 20 Nutzern empfohlen.
Ein paar andere *keys:- Hajkey
Soft-Fork von Calckey, der meines Wissens erst nach Firefish und dann nach Misskey rebased wurde. Wurde exklusiv nur für eine einzige Instanz entwickelt: transfem.social. Inzwischen eingestellt, weil wohl der Aufwand, so einen Wolpertinger nur für eine Instanz zu pflegen (man rebaset nicht mal eben von Firefish nach Misskey), zu groß war. transfem.social wurde entsprechend auf Sharkey umgestellt und ist da jetzt eine der größten Instanzen. - Neko
Soft-Fork von Misskey mit dem Ziel, es tauglich für Docker zu machen. Laut Repository nur für eine einzige Instanz gebaut. Ist nie released worden, was aber Wurscht ist, wenn es eh nur einen Admin als Zielgruppe hatte.git fetchkönnen Releases nämlich piepegal sein. - Meisskey
Japanischer Soft-Fork von Misskey von 2019 (!), der aber lange Zeit der Weiterentwicklung der Basis hinterherhinkte. Wird tatsächlich immer noch weiterentwickelt. - Leisskey
Wiederum japanischer Soft-Fork von Meisskey. Ist seit mindestens 2021 in Entwicklung, seit Februar 2023 aber eine ewige Beta, weil es seit damals keinen Release mehr gegeben hat. Dürfte zu den letzten noch in Entwicklung befindlichen "Forkeys 2. Grades" gehören. - Tanukey
Weiterer japanischer Soft-Fork von Misskey und daher so obskur im Westen, daß man schon die Websuche anstrengen muß, um das Repository zu finden (ist wohl im letzten Oktober von GitHub nach GitLab umgezogen). Noch ein Beispiel für *keys, bei denen selbst das Wissen über ihre Existenz Ostasien kaum je verlassen hat. - Backspacekey
Noch ein ambitionierter, aber eingeschlafener westlicher Misskey-Soft-Fork.
Siehe übrigens auch die Delightful Fediverse Experience: hier und hier.
Verglichen damit ist die Familie von Mistpark bis Forte schon wieder übersichtlich, vor allem, wenn die anderen mehreren Dutzend toten Forkeys mit dazugeholt werden. Immerhin stammt von Mistpark bis Forte über etwa 15 Jahre alles vom selben Schöpfer, der einfach nur sein eigenes Zeug geforkt hat.
#Long #LongPost #CWLong #CWLongPost #LangerPost #CWLangerPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #Misskey #Forkey #Forkeys #Calckey #Firefish #Iceshrimp #IceshrimpJS #Iceshrimp.NET #Catodon #Sharkey #CherryPick #FoundKey #Hajkey #Neko #Meisskey #Leisskey #Tanukey #Backspacekey - Misskey
-
@「 Jürgen 」:fedi_mastodon: @crossgolf_rebel - kostenlose Kwalitätsposts @Don di Dislessia Soweit ich weiß, war es so (man möge mich wiederum korrigieren; kursiv sind die "Forkeys höheren Grades", die nicht von Misskey geforkt wurden):
- Misskey
Der Ursprung in Japan. - Calckey
Soft-Fork von Misskey mit einigen Extrafeatures.
Entwickler hatte irgendwann keine Zeit/keinen Bock mehr. - Firefish
Fortführung von Calckey unter neuem Management. Wurde aufgrund einer massiven Werbeaktion eines begeisterten Nutzers mit viel Reichweite so populär, daß der Name "Calckey" irgendwann einfach doof war und das ganze Ding eine neue Identität bekam.
Entwickler verschwand irgendwann sang- und klanglos von der Bildfläche. Nach einem halben Jahr stellte sich raus: Entwickler hatte wegen Abschlußarbeit usw. keine Zeit mehr, nicht mal, sich zu verabschieden.
Wurde unter neuer Führung mit neuem Repository und neuer Leuchtturminstanz weitergeführt, aber ohne neue Website. Wurde wieder eingestellt, weil es für eine einzige Entwicklerin viel zuviel war, die ganzen alten Co-Entwickler von Firefish alle zu Iceshrimp gewechselt waren und keine neuen Mitentwickler angeheuert werden konnten. - Iceshrimp
Fork von Firefish von ehemaligen Firefish-Entwicklern. Rebased nach Misskey, weil es auf Firefish nicht weiterging. Erklärtes Ziel war, Stabilität über Calckeys Featuritis zu stellen.
Weil Misskeys Codebase an sich einiges an grundsätzlichen Macken hatte, wurde beschlossen, es ist einfacher, das ganze Zeugs von Grund auf neu zu schreiben, als zu versuchen, das alles auszubügeln. Und bei der Gelegenheit wollte man von JavaScript (TypeScript und Vue.js) weg. Also hat man angefangen, das ganze Ding in C# neu zu schreiben als Iceshrimp.NET. Ziel ist featuremäßige Deckungsgleichheit mit dem bisherigen Iceshrimp und gleichzeitig Anpassung an Mastodon. Iceshrimp.NET ist noch sehr unfertig.
Bei der Gelegenheit wurde das alte Iceshrimp in Iceshrimp-JS umbenannt und in den Wartungsmodus versetzt = gibt keine Weiterentwicklung und keine neuen Features mehr, nur Sicherheitspatches und evtl. Bugfixes. - Catodon
Soft-Fork von Firefish. Ziel war, ein Forkey zu haben, das sich in der Bedienung wie Mastodon anfühlt. Quasi Mastodon-Fassade mit *key-Struktur dahinter. Also ein Forkey für den möglichst leichten Umstieg von Mastodon nach *key.
Wurde, als Firefish endgültig eingestellt wurde und damit die Basis tot war, nach Iceshrimp rebased.
Nachdem Iceshrimp zu Iceshrimp-JS wurde und in den Wartungsmodus ging, hatte Catodon ein Problem: Man kann nicht einfach von etwas, das in TypeScript und Vue.js geschrieben ist, nach etwas rebasen, das in C# geschrieben ist (Iceshrimp.NET). Und zur Weiterentwicklung wäre ein Rebase nötig gewesen, weil man jetzt wieder eine fast tote Basis hatte. Aber nach Misskey wollte man nicht (siehe Hajkey weiter unten), und Sharkey war einerseits zu mächtig und andererseits zu kapriziös.
Inzwischen gehen Catodon die Instanzen ein. Die meisten dürften geschlossen worden sein. - Sharkey
Soft-Fork von Misskey mit dem Ziel, so ziemlich alles, was je für Forkeys an Features entwickelt worden war, in einen einzigen Forkey zu packen und noch mehr obendrauf.
Quasi das Yang zu Iceshrimps Yin. Machte lange den Eindruck, als gingen Features über Zuverlässigkeit. Sharkeys Implementation der Mastodon Client API war schon regelrecht legendär schlecht, und alle hofften auf den Retter, der daherkommt und die Implementation von Grund auf neu schreibt, weil sie als unrettbar im Eimer galt. Inzwischen soll Sharkey seine Zuverlässigkeit im Griff haben, zumindest insofern, wie auch Misskey zuverlässig ist.
Die Entwickler haben eine zweifelhafte Reputation. Unter anderem haben sie Crowdfunding-Gelder für einen Sharkey-Server gesammelt und von dem Geld einen Minecraft-Server aufgezogen, aber keinen Sharkey-Server. Mitunter wird deshalb zum Boykott von Sharkey aufgerufen. - CherryPick
Südkoreanischer Soft-Fork von Misskey. Tatsächlich älter als Sharkey, mindestens von 2021, aber später wurde wohl einiges von Sharkey nach CherryPick portiert. Das Ziel war, einen stabilen und zuverlässigen Forkey zu haben ohne die Macken von Sharkey und sogar ohne die Macken von Misskey, der aber gleichzeitig gut Features hat. Das ist wohl sogar weitestgehend gelungen.
Vom ästhetischen Stil her so ähnlich, wie Misskey früher mal war, also sehr auf den japanischen bzw. südkoreanischen Geschmack ausgelegt: bunt, grell, poppig, genki, kawaii. Alleine das zeigt, daß CherryPick von Misskey geforkt wurde, bevor Misskey gänzlich an den westlichen Geschmack angepaßt wurde.
Unklar, ob es Entwickler hat, die Englisch verstehen bzw. schreiben können; falls nicht, dann als Soft-Fork-Basis ungeeignet. Wäre ohne die Kommunikationsbarriere vermutlich der ultimative *key.
Mein letzter Stand: Bis auf eine Instanz im Großraum Washington, D.C. gibt es CherryPick-Instanzen nur in den Großräumen Tokyo und Seoul. War deshalb lange Zeit im westlichen Fediverse fast unbekannt. - FoundKey
Wohl der erste Forkey, der in PascalCase geschrieben wurde.
Meines Wissens direkter Soft-Fork von Misskey. Wird tatsächlich (oh Wunder) kleckerweise gepflegt, aber nicht für Instanzen mit mehr als 20 Nutzern empfohlen.
Ein paar andere *keys:- Hajkey
Soft-Fork von Calckey, der meines Wissens erst nach Firefish und dann nach Misskey rebased wurde. Wurde exklusiv nur für eine einzige Instanz entwickelt: transfem.social. Inzwischen eingestellt, weil wohl der Aufwand, so einen Wolpertinger nur für eine Instanz zu pflegen (man rebaset nicht mal eben von Firefish nach Misskey), zu groß war. transfem.social wurde entsprechend auf Sharkey umgestellt und ist da jetzt eine der größten Instanzen. - Neko
Soft-Fork von Misskey mit dem Ziel, es tauglich für Docker zu machen. Laut Repository nur für eine einzige Instanz gebaut. Ist nie released worden, was aber Wurscht ist, wenn es eh nur einen Admin als Zielgruppe hatte.git fetchkönnen Releases nämlich piepegal sein. - Meisskey
Japanischer Soft-Fork von Misskey von 2019 (!), der aber lange Zeit der Weiterentwicklung der Basis hinterherhinkte. Wird tatsächlich immer noch weiterentwickelt. - Leisskey
Wiederum japanischer Soft-Fork von Meisskey. Ist seit mindestens 2021 in Entwicklung, seit Februar 2023 aber eine ewige Beta, weil es seit damals keinen Release mehr gegeben hat. Dürfte zu den letzten noch in Entwicklung befindlichen "Forkeys 2. Grades" gehören. - Tanukey
Weiterer japanischer Soft-Fork von Misskey und daher so obskur im Westen, daß man schon die Websuche anstrengen muß, um das Repository zu finden (ist wohl im letzten Oktober von GitHub nach GitLab umgezogen). Noch ein Beispiel für *keys, bei denen selbst das Wissen über ihre Existenz Ostasien kaum je verlassen hat. - Backspacekey
Noch ein ambitionierter, aber eingeschlafener westlicher Misskey-Soft-Fork.
Siehe übrigens auch die Delightful Fediverse Experience: hier und hier.
Verglichen damit ist die Familie von Mistpark bis Forte schon wieder übersichtlich, vor allem, wenn die anderen mehreren Dutzend toten Forkeys mit dazugeholt werden. Immerhin stammt von Mistpark bis Forte über etwa 15 Jahre alles vom selben Schöpfer, der einfach nur sein eigenes Zeug geforkt hat.
#Long #LongPost #CWLong #CWLongPost #LangerPost #CWLangerPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #Misskey #Forkey #Forkeys #Calckey #Firefish #Iceshrimp #IceshrimpJS #Iceshrimp.NET #Catodon #Sharkey #CherryPick #FoundKey #Hajkey #Neko #Meisskey #Leisskey #Tanukey #Backspacekey - Misskey
-
@「 Jürgen 」:fedi_mastodon: @crossgolf_rebel - kostenlose Kwalitätsposts @Don di Dislessia Soweit ich weiß, war es so (man möge mich wiederum korrigieren; kursiv sind die "Forkeys höheren Grades", die nicht von Misskey geforkt wurden):
- Misskey
Der Ursprung in Japan. - Calckey
Soft-Fork von Misskey mit einigen Extrafeatures.
Entwickler hatte irgendwann keine Zeit/keinen Bock mehr. - Firefish
Fortführung von Calckey unter neuem Management. Wurde aufgrund einer massiven Werbeaktion eines begeisterten Nutzers mit viel Reichweite so populär, daß der Name "Calckey" irgendwann einfach doof war und das ganze Ding eine neue Identität bekam.
Entwickler verschwand irgendwann sang- und klanglos von der Bildfläche. Nach einem halben Jahr stellte sich raus: Entwickler hatte wegen Abschlußarbeit usw. keine Zeit mehr, nicht mal, sich zu verabschieden.
Wurde unter neuer Führung mit neuem Repository und neuer Leuchtturminstanz weitergeführt, aber ohne neue Website. Wurde wieder eingestellt, weil es für eine einzige Entwicklerin viel zuviel war, die ganzen alten Co-Entwickler von Firefish alle zu Iceshrimp gewechselt waren und keine neuen Mitentwickler angeheuert werden konnten. - Iceshrimp
Fork von Firefish von ehemaligen Firefish-Entwicklern. Rebased nach Misskey, weil es auf Firefish nicht weiterging. Erklärtes Ziel war, Stabilität über Calckeys Featuritis zu stellen.
Weil Misskeys Codebase an sich einiges an grundsätzlichen Macken hatte, wurde beschlossen, es ist einfacher, das ganze Zeugs von Grund auf neu zu schreiben, als zu versuchen, das alles auszubügeln. Und bei der Gelegenheit wollte man von JavaScript (TypeScript und Vue.js) weg. Also hat man angefangen, das ganze Ding in C# neu zu schreiben als Iceshrimp.NET. Ziel ist featuremäßige Deckungsgleichheit mit dem bisherigen Iceshrimp und gleichzeitig Anpassung an Mastodon. Iceshrimp.NET ist noch sehr unfertig.
Bei der Gelegenheit wurde das alte Iceshrimp in Iceshrimp-JS umbenannt und in den Wartungsmodus versetzt = gibt keine Weiterentwicklung und keine neuen Features mehr, nur Sicherheitspatches und evtl. Bugfixes. - Catodon
Soft-Fork von Firefish. Ziel war, ein Forkey zu haben, das sich in der Bedienung wie Mastodon anfühlt. Quasi Mastodon-Fassade mit *key-Struktur dahinter. Also ein Forkey für den möglichst leichten Umstieg von Mastodon nach *key.
Wurde, als Firefish endgültig eingestellt wurde und damit die Basis tot war, nach Iceshrimp rebased.
Nachdem Iceshrimp zu Iceshrimp-JS wurde und in den Wartungsmodus ging, hatte Catodon ein Problem: Man kann nicht einfach von etwas, das in TypeScript und Vue.js geschrieben ist, nach etwas rebasen, das in C# geschrieben ist (Iceshrimp.NET). Und zur Weiterentwicklung wäre ein Rebase nötig gewesen, weil man jetzt wieder eine fast tote Basis hatte. Aber nach Misskey wollte man nicht (siehe Hajkey weiter unten), und Sharkey war einerseits zu mächtig und andererseits zu kapriziös.
Inzwischen gehen Catodon die Instanzen ein. Die meisten dürften geschlossen worden sein. - Sharkey
Soft-Fork von Misskey mit dem Ziel, so ziemlich alles, was je für Forkeys an Features entwickelt worden war, in einen einzigen Forkey zu packen und noch mehr obendrauf.
Quasi das Yang zu Iceshrimps Yin. Machte lange den Eindruck, als gingen Features über Zuverlässigkeit. Sharkeys Implementation der Mastodon Client API war schon regelrecht legendär schlecht, und alle hofften auf den Retter, der daherkommt und die Implementation von Grund auf neu schreibt, weil sie als unrettbar im Eimer galt. Inzwischen soll Sharkey seine Zuverlässigkeit im Griff haben, zumindest insofern, wie auch Misskey zuverlässig ist.
Die Entwickler haben eine zweifelhafte Reputation. Unter anderem haben sie Crowdfunding-Gelder für einen Sharkey-Server gesammelt und von dem Geld einen Minecraft-Server aufgezogen, aber keinen Sharkey-Server. Mitunter wird deshalb zum Boykott von Sharkey aufgerufen. - CherryPick
Südkoreanischer Soft-Fork von Misskey. Tatsächlich älter als Sharkey, mindestens von 2021, aber später wurde wohl einiges von Sharkey nach CherryPick portiert. Das Ziel war, einen stabilen und zuverlässigen Forkey zu haben ohne die Macken von Sharkey und sogar ohne die Macken von Misskey, der aber gleichzeitig gut Features hat. Das ist wohl sogar weitestgehend gelungen.
Vom ästhetischen Stil her so ähnlich, wie Misskey früher mal war, also sehr auf den japanischen bzw. südkoreanischen Geschmack ausgelegt: bunt, grell, poppig, genki, kawaii. Alleine das zeigt, daß CherryPick von Misskey geforkt wurde, bevor Misskey gänzlich an den westlichen Geschmack angepaßt wurde.
Unklar, ob es Entwickler hat, die Englisch verstehen bzw. schreiben können; falls nicht, dann als Soft-Fork-Basis ungeeignet. Wäre ohne die Kommunikationsbarriere vermutlich der ultimative *key.
Mein letzter Stand: Bis auf eine Instanz im Großraum Washington, D.C. gibt es CherryPick-Instanzen nur in den Großräumen Tokyo und Seoul. War deshalb lange Zeit im westlichen Fediverse fast unbekannt. - FoundKey
Wohl der erste Forkey, der in PascalCase geschrieben wurde.
Meines Wissens direkter Soft-Fork von Misskey. Wird tatsächlich (oh Wunder) kleckerweise gepflegt, aber nicht für Instanzen mit mehr als 20 Nutzern empfohlen.
Ein paar andere *keys:- Hajkey
Soft-Fork von Calckey, der meines Wissens erst nach Firefish und dann nach Misskey rebased wurde. Wurde exklusiv nur für eine einzige Instanz entwickelt: transfem.social. Inzwischen eingestellt, weil wohl der Aufwand, so einen Wolpertinger nur für eine Instanz zu pflegen (man rebaset nicht mal eben von Firefish nach Misskey), zu groß war. transfem.social wurde entsprechend auf Sharkey umgestellt und ist da jetzt eine der größten Instanzen. - Neko
Soft-Fork von Misskey mit dem Ziel, es tauglich für Docker zu machen. Laut Repository nur für eine einzige Instanz gebaut. Ist nie released worden, was aber Wurscht ist, wenn es eh nur einen Admin als Zielgruppe hatte.git fetchkönnen Releases nämlich piepegal sein. - Meisskey
Japanischer Soft-Fork von Misskey von 2019 (!), der aber lange Zeit der Weiterentwicklung der Basis hinterherhinkte. Wird tatsächlich immer noch weiterentwickelt. - Leisskey
Wiederum japanischer Soft-Fork von Meisskey. Ist seit mindestens 2021 in Entwicklung, seit Februar 2023 aber eine ewige Beta, weil es seit damals keinen Release mehr gegeben hat. Dürfte zu den letzten noch in Entwicklung befindlichen "Forkeys 2. Grades" gehören. - Tanukey
Weiterer japanischer Soft-Fork von Misskey und daher so obskur im Westen, daß man schon die Websuche anstrengen muß, um das Repository zu finden (ist wohl im letzten Oktober von GitHub nach GitLab umgezogen). Noch ein Beispiel für *keys, bei denen selbst das Wissen über ihre Existenz Ostasien kaum je verlassen hat. - Backspacekey
Noch ein ambitionierter, aber eingeschlafener westlicher Misskey-Soft-Fork.
Siehe übrigens auch die Delightful Fediverse Experience: hier und hier.
Verglichen damit ist die Familie von Mistpark bis Forte schon wieder übersichtlich, vor allem, wenn die anderen mehreren Dutzend toten Forkeys mit dazugeholt werden. Immerhin stammt von Mistpark bis Forte über etwa 15 Jahre alles vom selben Schöpfer, der einfach nur sein eigenes Zeug geforkt hat.
#Long #LongPost #CWLong #CWLongPost #LangerPost #CWLangerPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #Misskey #Forkey #Forkeys #Calckey #Firefish #Iceshrimp #IceshrimpJS #Iceshrimp.NET #Catodon #Sharkey #CherryPick #FoundKey #Hajkey #Neko #Meisskey #Leisskey #Tanukey #Backspacekey - Misskey
-
@「 Jürgen 」:fedi_mastodon: @crossgolf_rebel - kostenlose Kwalitätsposts @Don di Dislessia Soweit ich weiß, war es so (man möge mich wiederum korrigieren; kursiv sind die "Forkeys höheren Grades", die nicht von Misskey geforkt wurden):
- Misskey
Der Ursprung in Japan. - Calckey
Soft-Fork von Misskey mit einigen Extrafeatures.
Entwickler hatte irgendwann keine Zeit/keinen Bock mehr. - Firefish
Fortführung von Calckey unter neuem Management. Wurde aufgrund einer massiven Werbeaktion eines begeisterten Nutzers mit viel Reichweite so populär, daß der Name "Calckey" irgendwann einfach doof war und das ganze Ding eine neue Identität bekam.
Entwickler verschwand irgendwann sang- und klanglos von der Bildfläche. Nach einem halben Jahr stellte sich raus: Entwickler hatte wegen Abschlußarbeit usw. keine Zeit mehr, nicht mal, sich zu verabschieden.
Wurde unter neuer Führung mit neuem Repository und neuer Leuchtturminstanz weitergeführt, aber ohne neue Website. Wurde wieder eingestellt, weil es für eine einzige Entwicklerin viel zuviel war, die ganzen alten Co-Entwickler von Firefish alle zu Iceshrimp gewechselt waren und keine neuen Mitentwickler angeheuert werden konnten. - Iceshrimp
Fork von Firefish von ehemaligen Firefish-Entwicklern. Rebased nach Misskey, weil es auf Firefish nicht weiterging. Erklärtes Ziel war, Stabilität über Calckeys Featuritis zu stellen.
Weil Misskeys Codebase an sich einiges an grundsätzlichen Macken hatte, wurde beschlossen, es ist einfacher, das ganze Zeugs von Grund auf neu zu schreiben, als zu versuchen, das alles auszubügeln. Und bei der Gelegenheit wollte man von JavaScript (TypeScript und Vue.js) weg. Also hat man angefangen, das ganze Ding in C# neu zu schreiben als Iceshrimp.NET. Ziel ist featuremäßige Deckungsgleichheit mit dem bisherigen Iceshrimp und gleichzeitig Anpassung an Mastodon. Iceshrimp.NET ist noch sehr unfertig.
Bei der Gelegenheit wurde das alte Iceshrimp in Iceshrimp-JS umbenannt und in den Wartungsmodus versetzt = gibt keine Weiterentwicklung und keine neuen Features mehr, nur Sicherheitspatches und evtl. Bugfixes. - Catodon
Soft-Fork von Firefish. Ziel war, ein Forkey zu haben, das sich in der Bedienung wie Mastodon anfühlt. Quasi Mastodon-Fassade mit *key-Struktur dahinter. Also ein Forkey für den möglichst leichten Umstieg von Mastodon nach *key.
Wurde, als Firefish endgültig eingestellt wurde und damit die Basis tot war, nach Iceshrimp rebased.
Nachdem Iceshrimp zu Iceshrimp-JS wurde und in den Wartungsmodus ging, hatte Catodon ein Problem: Man kann nicht einfach von etwas, das in TypeScript und Vue.js geschrieben ist, nach etwas rebasen, das in C# geschrieben ist (Iceshrimp.NET). Und zur Weiterentwicklung wäre ein Rebase nötig gewesen, weil man jetzt wieder eine fast tote Basis hatte. Aber nach Misskey wollte man nicht (siehe Hajkey weiter unten), und Sharkey war einerseits zu mächtig und andererseits zu kapriziös.
Inzwischen gehen Catodon die Instanzen ein. Die meisten dürften geschlossen worden sein. - Sharkey
Soft-Fork von Misskey mit dem Ziel, so ziemlich alles, was je für Forkeys an Features entwickelt worden war, in einen einzigen Forkey zu packen und noch mehr obendrauf.
Quasi das Yang zu Iceshrimps Yin. Machte lange den Eindruck, als gingen Features über Zuverlässigkeit. Sharkeys Implementation der Mastodon Client API war schon regelrecht legendär schlecht, und alle hofften auf den Retter, der daherkommt und die Implementation von Grund auf neu schreibt, weil sie als unrettbar im Eimer galt. Inzwischen soll Sharkey seine Zuverlässigkeit im Griff haben, zumindest insofern, wie auch Misskey zuverlässig ist.
Die Entwickler haben eine zweifelhafte Reputation. Unter anderem haben sie Crowdfunding-Gelder für einen Sharkey-Server gesammelt und von dem Geld einen Minecraft-Server aufgezogen, aber keinen Sharkey-Server. Mitunter wird deshalb zum Boykott von Sharkey aufgerufen. - CherryPick
Südkoreanischer Soft-Fork von Misskey. Tatsächlich älter als Sharkey, mindestens von 2021, aber später wurde wohl einiges von Sharkey nach CherryPick portiert. Das Ziel war, einen stabilen und zuverlässigen Forkey zu haben ohne die Macken von Sharkey und sogar ohne die Macken von Misskey, der aber gleichzeitig gut Features hat. Das ist wohl sogar weitestgehend gelungen.
Vom ästhetischen Stil her so ähnlich, wie Misskey früher mal war, also sehr auf den japanischen bzw. südkoreanischen Geschmack ausgelegt: bunt, grell, poppig, genki, kawaii. Alleine das zeigt, daß CherryPick von Misskey geforkt wurde, bevor Misskey gänzlich an den westlichen Geschmack angepaßt wurde.
Unklar, ob es Entwickler hat, die Englisch verstehen bzw. schreiben können; falls nicht, dann als Soft-Fork-Basis ungeeignet. Wäre ohne die Kommunikationsbarriere vermutlich der ultimative *key.
Mein letzter Stand: Bis auf eine Instanz im Großraum Washington, D.C. gibt es CherryPick-Instanzen nur in den Großräumen Tokyo und Seoul. War deshalb lange Zeit im westlichen Fediverse fast unbekannt. - FoundKey
Wohl der erste Forkey, der in PascalCase geschrieben wurde.
Meines Wissens direkter Soft-Fork von Misskey. Wird tatsächlich (oh Wunder) kleckerweise gepflegt, aber nicht für Instanzen mit mehr als 20 Nutzern empfohlen.
Ein paar andere *keys:- Hajkey
Soft-Fork von Calckey, der meines Wissens erst nach Firefish und dann nach Misskey rebased wurde. Wurde exklusiv nur für eine einzige Instanz entwickelt: transfem.social. Inzwischen eingestellt, weil wohl der Aufwand, so einen Wolpertinger nur für eine Instanz zu pflegen (man rebaset nicht mal eben von Firefish nach Misskey), zu groß war. transfem.social wurde entsprechend auf Sharkey umgestellt und ist da jetzt eine der größten Instanzen. - Neko
Soft-Fork von Misskey mit dem Ziel, es tauglich für Docker zu machen. Laut Repository nur für eine einzige Instanz gebaut. Ist nie released worden, was aber Wurscht ist, wenn es eh nur einen Admin als Zielgruppe hatte.git fetchkönnen Releases nämlich piepegal sein. - Meisskey
Japanischer Soft-Fork von Misskey von 2019 (!), der aber lange Zeit der Weiterentwicklung der Basis hinterherhinkte. Wird tatsächlich immer noch weiterentwickelt. - Leisskey
Wiederum japanischer Soft-Fork von Meisskey. Ist seit mindestens 2021 in Entwicklung, seit Februar 2023 aber eine ewige Beta, weil es seit damals keinen Release mehr gegeben hat. Dürfte zu den letzten noch in Entwicklung befindlichen "Forkeys 2. Grades" gehören. - Tanukey
Weiterer japanischer Soft-Fork von Misskey und daher so obskur im Westen, daß man schon die Websuche anstrengen muß, um das Repository zu finden (ist wohl im letzten Oktober von GitHub nach GitLab umgezogen). Noch ein Beispiel für *keys, bei denen selbst das Wissen über ihre Existenz Ostasien kaum je verlassen hat. - Backspacekey
Noch ein ambitionierter, aber eingeschlafener westlicher Misskey-Soft-Fork.
Siehe übrigens auch die Delightful Fediverse Experience: hier und hier.
Verglichen damit ist die Familie von Mistpark bis Forte schon wieder übersichtlich, vor allem, wenn die anderen mehreren Dutzend toten Forkeys mit dazugeholt werden. Immerhin stammt von Mistpark bis Forte über etwa 15 Jahre alles vom selben Schöpfer, der einfach nur sein eigenes Zeug geforkt hat.
#Long #LongPost #CWLong #CWLongPost #LangerPost #CWLangerPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #Misskey #Forkey #Forkeys #Calckey #Firefish #Iceshrimp #IceshrimpJS #Iceshrimp.NET #Catodon #Sharkey #CherryPick #FoundKey #Hajkey #Neko #Meisskey #Leisskey #Tanukey #Backspacekey - Misskey
-
@「 Jürgen 」:fedi_mastodon: @crossgolf_rebel - kostenlose Kwalitätsposts @Don di Dislessia Soweit ich weiß, war es so (man möge mich wiederum korrigieren; kursiv sind die "Forkeys höheren Grades", die nicht von Misskey geforkt wurden):
- Misskey
Der Ursprung in Japan. - Calckey
Soft-Fork von Misskey mit einigen Extrafeatures.
Entwickler hatte irgendwann keine Zeit/keinen Bock mehr. - Firefish
Fortführung von Calckey unter neuem Management. Wurde aufgrund einer massiven Werbeaktion eines begeisterten Nutzers mit viel Reichweite so populär, daß der Name "Calckey" irgendwann einfach doof war und das ganze Ding eine neue Identität bekam.
Entwickler verschwand irgendwann sang- und klanglos von der Bildfläche. Nach einem halben Jahr stellte sich raus: Entwickler hatte wegen Abschlußarbeit usw. keine Zeit mehr, nicht mal, sich zu verabschieden.
Wurde unter neuer Führung mit neuem Repository und neuer Leuchtturminstanz weitergeführt, aber ohne neue Website. Wurde wieder eingestellt, weil es für eine einzige Entwicklerin viel zuviel war, die ganzen alten Co-Entwickler von Firefish alle zu Iceshrimp gewechselt waren und keine neuen Mitentwickler angeheuert werden konnten. - Iceshrimp
Fork von Firefish von ehemaligen Firefish-Entwicklern. Rebased nach Misskey, weil es auf Firefish nicht weiterging. Erklärtes Ziel war, Stabilität über Calckeys Featuritis zu stellen.
Weil Misskeys Codebase an sich einiges an grundsätzlichen Macken hatte, wurde beschlossen, es ist einfacher, das ganze Zeugs von Grund auf neu zu schreiben, als zu versuchen, das alles auszubügeln. Und bei der Gelegenheit wollte man von JavaScript (TypeScript und Vue.js) weg. Also hat man angefangen, das ganze Ding in C# neu zu schreiben als Iceshrimp.NET. Ziel ist featuremäßige Deckungsgleichheit mit dem bisherigen Iceshrimp und gleichzeitig Anpassung an Mastodon. Iceshrimp.NET ist noch sehr unfertig.
Bei der Gelegenheit wurde das alte Iceshrimp in Iceshrimp-JS umbenannt und in den Wartungsmodus versetzt = gibt keine Weiterentwicklung und keine neuen Features mehr, nur Sicherheitspatches und evtl. Bugfixes. - Catodon
Soft-Fork von Firefish. Ziel war, ein Forkey zu haben, das sich in der Bedienung wie Mastodon anfühlt. Quasi Mastodon-Fassade mit *key-Struktur dahinter. Also ein Forkey für den möglichst leichten Umstieg von Mastodon nach *key.
Wurde, als Firefish endgültig eingestellt wurde und damit die Basis tot war, nach Iceshrimp rebased.
Nachdem Iceshrimp zu Iceshrimp-JS wurde und in den Wartungsmodus ging, hatte Catodon ein Problem: Man kann nicht einfach von etwas, das in TypeScript und Vue.js geschrieben ist, nach etwas rebasen, das in C# geschrieben ist (Iceshrimp.NET). Und zur Weiterentwicklung wäre ein Rebase nötig gewesen, weil man jetzt wieder eine fast tote Basis hatte. Aber nach Misskey wollte man nicht (siehe Hajkey weiter unten), und Sharkey war einerseits zu mächtig und andererseits zu kapriziös.
Inzwischen gehen Catodon die Instanzen ein. Die meisten dürften geschlossen worden sein. - Sharkey
Soft-Fork von Misskey mit dem Ziel, so ziemlich alles, was je für Forkeys an Features entwickelt worden war, in einen einzigen Forkey zu packen und noch mehr obendrauf.
Quasi das Yang zu Iceshrimps Yin. Machte lange den Eindruck, als gingen Features über Zuverlässigkeit. Sharkeys Implementation der Mastodon Client API war schon regelrecht legendär schlecht, und alle hofften auf den Retter, der daherkommt und die Implementation von Grund auf neu schreibt, weil sie als unrettbar im Eimer galt. Inzwischen soll Sharkey seine Zuverlässigkeit im Griff haben, zumindest insofern, wie auch Misskey zuverlässig ist.
Die Entwickler haben eine zweifelhafte Reputation. Unter anderem haben sie Crowdfunding-Gelder für einen Sharkey-Server gesammelt und von dem Geld einen Minecraft-Server aufgezogen, aber keinen Sharkey-Server. Mitunter wird deshalb zum Boykott von Sharkey aufgerufen. - CherryPick
Südkoreanischer Soft-Fork von Misskey. Tatsächlich älter als Sharkey, mindestens von 2021, aber später wurde wohl einiges von Sharkey nach CherryPick portiert. Das Ziel war, einen stabilen und zuverlässigen Forkey zu haben ohne die Macken von Sharkey und sogar ohne die Macken von Misskey, der aber gleichzeitig gut Features hat. Das ist wohl sogar weitestgehend gelungen.
Vom ästhetischen Stil her so ähnlich, wie Misskey früher mal war, also sehr auf den japanischen bzw. südkoreanischen Geschmack ausgelegt: bunt, grell, poppig, genki, kawaii. Alleine das zeigt, daß CherryPick von Misskey geforkt wurde, bevor Misskey gänzlich an den westlichen Geschmack angepaßt wurde.
Unklar, ob es Entwickler hat, die Englisch verstehen bzw. schreiben können; falls nicht, dann als Soft-Fork-Basis ungeeignet. Wäre ohne die Kommunikationsbarriere vermutlich der ultimative *key.
Mein letzter Stand: Bis auf eine Instanz im Großraum Washington, D.C. gibt es CherryPick-Instanzen nur in den Großräumen Tokyo und Seoul. War deshalb lange Zeit im westlichen Fediverse fast unbekannt. - FoundKey
Wohl der erste Forkey, der in PascalCase geschrieben wurde.
Meines Wissens direkter Soft-Fork von Misskey. Wird tatsächlich (oh Wunder) kleckerweise gepflegt, aber nicht für Instanzen mit mehr als 20 Nutzern empfohlen.
Ein paar andere *keys:- Hajkey
Soft-Fork von Calckey, der meines Wissens erst nach Firefish und dann nach Misskey rebased wurde. Wurde exklusiv nur für eine einzige Instanz entwickelt: transfem.social. Inzwischen eingestellt, weil wohl der Aufwand, so einen Wolpertinger nur für eine Instanz zu pflegen (man rebaset nicht mal eben von Firefish nach Misskey), zu groß war. transfem.social wurde entsprechend auf Sharkey umgestellt und ist da jetzt eine der größten Instanzen. - Neko
Soft-Fork von Misskey mit dem Ziel, es tauglich für Docker zu machen. Laut Repository nur für eine einzige Instanz gebaut. Ist nie released worden, was aber Wurscht ist, wenn es eh nur einen Admin als Zielgruppe hatte.git fetchkönnen Releases nämlich piepegal sein. - Meisskey
Japanischer Soft-Fork von Misskey von 2019 (!), der aber lange Zeit der Weiterentwicklung der Basis hinterherhinkte. Wird tatsächlich immer noch weiterentwickelt. - Leisskey
Wiederum japanischer Soft-Fork von Meisskey. Ist seit mindestens 2021 in Entwicklung, seit Februar 2023 aber eine ewige Beta, weil es seit damals keinen Release mehr gegeben hat. Dürfte zu den letzten noch in Entwicklung befindlichen "Forkeys 2. Grades" gehören. - Tanukey
Weiterer japanischer Soft-Fork von Misskey und daher so obskur im Westen, daß man schon die Websuche anstrengen muß, um das Repository zu finden (ist wohl im letzten Oktober von GitHub nach GitLab umgezogen). Noch ein Beispiel für *keys, bei denen selbst das Wissen über ihre Existenz Ostasien kaum je verlassen hat. - Backspacekey
Noch ein ambitionierter, aber eingeschlafener westlicher Misskey-Soft-Fork.
Siehe übrigens auch die Delightful Fediverse Experience: hier und hier.
Verglichen damit ist die Familie von Mistpark bis Forte schon wieder übersichtlich, vor allem, wenn die anderen mehreren Dutzend toten Forkeys mit dazugeholt werden. Immerhin stammt von Mistpark bis Forte über etwa 15 Jahre alles vom selben Schöpfer, der einfach nur sein eigenes Zeug geforkt hat.
#Long #LongPost #CWLong #CWLongPost #LangerPost #CWLangerPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #Misskey #Forkey #Forkeys #Calckey #Firefish #Iceshrimp #IceshrimpJS #Iceshrimp.NET #Catodon #Sharkey #CherryPick #FoundKey #Hajkey #Neko #Meisskey #Leisskey #Tanukey #Backspacekey - Misskey
-
CW: Why Sharkey fails to render hashtags from Friendica, Hubzilla & Co. properly, and how long this bug has been known already; CW: long (over 6,600 characters), Fediverse meta, Fediverse-beyond-Mastodon meta
So there's that nasty bug on Sharkey that mangles hashtags in messages from Hubzilla and probably also Friendica, (streams) and Forte. They always look like this:
#[Hashtag](https://hub.netzgemeinde.eu/search?tag=Hashtag)
Basically, Sharkey receives fully standard Rich Text from Hubzilla. It manages to convert this Rich Text into its own Misskey-Flavored Markdown. But then its Markdown parser does not parse it and leaves the Markdown code visible to everyone. It simply doesn't expect there to be a hashtag character in front of an embedded link because, seriously, who'd ever do that and why?!
Friendica would. In fact, Friendica does. It puts the hashtag character in front of the tag, as in outside the tag, as opposed to at the beginning of the tag. It has been doing that since its beginnings in 2010 because it was designed from the get-go to also federate with StatusNet from 2008. And StatusNet does hashtags the same way on its few remaining servers. In fact, so did Identi.ca from 2008, from which StatusNet emerged.
Hubzilla, (streams) and Forte do it, too, because they have inherited it from Friendica.
On StatusNet, Friendica, Hubzilla, (streams) and Forte, a hashtag in a message looks like this:
#Hashtag
Notice how the hashtag character has the same colour as the rest of the post text. And not the same colour as the rest of the hashtag. This means that the hashtag character is not part of the link. (To Mastodon users who don't know this: If something in a "toot" has a different colour from the rest of the "toot", it's a link. Even if it doesn't show a URL in plain sight.)
On 𝕏, Mastodon, Pleroma, Akkoma, Misskey, the various Forkeys and a whole lot of other Fediverse software, a hashtag in a message looks like this:
#Hashtag
Notice how now the hashtag character has the same colour as the rest of the hashtag. This means that the hashtag character is part of the link.
But why did Identi.ca do hashtags differently from Twitter? Because Identi.ca did hashtags before Twitter. AFAIK, when Identi.ca was launched, it had support for hashtags right away. About one year before Twitter.
The hashtag itself had already been invented by the Twitter community. Chris Messina had already codified it in 2007. But it wasn't until 2009 that Twitter actually introduced a technological implementation to support it.
Again, Identi.ca must have had hashtags as early as 2008, and there was no way that Identi.ca creator Evan Prodromou could possibly predict what Twitter would do the following year. So he did what he thought was right and what actually made sense to him.
But nowadays, everybody "knows" that Twitter had the world's very first hashtag implementation ever because nobody, even in the Fediverse, has ever heard of Identi.ca. I mean, the majority of Fediverse users "know" that the Fediverse started with Mastodon.
You know, just like Officer James Barrett "knew" that there is no intelligent life outside Earth only a few minutes before he became Agent J of the Men In Black.
This is also why just about all Fediverse software that does hashtags the Twitter way expects everything to do hashtags the Twitter way. It does not expect hashtags to be done differently. And when a message comes in from Friendica, Hubzilla, (streams) or Forte with hashtags in it, it fails at varying degrees of ungracefully.
Hashtags with the hashtag character outside the link are older than hashtags with the hashtag character inside that they're not only completely unexpected, that they cause software to malfunction, but the same software often can't even handle that malfunction. It's a miracle that the Friendica/Hubzilla family doesn't cause Fediverse servers to crash or even server databases to go corrupt by simply sending hashtags.
Mastodon used to be an exception of sorts, but only because, before version 4.0 from October, 2022, its HTML "sanitiser" actually ripped out any and all rich text code from incoming messages and left nothing but plain text behind. And then it didn't recognise hashtags in messages from outside Mastodon as hashtags at all.
When Mastodon 4.0 came and supported some rich text, including embedded links, it went haywire, of course. But then someone from Friendica and Hubzilla went in and complained about this malfunction and explained what happened, why it happened and why it was not Friendica and Hubzilla that did things wrong. Besides, if something utterly defaces "toots", then Mastodon developers do step in to stop it. After all, Mastodon has a few more of them at hand, paid, full-time professionals even. You have to give it that.
Which takes us back to Sharkey. Sharkey is developed by a small handful of individuals in their spare time. Granted, it's a soft fork of Misskey, so a lot of development work is done by the Misskey devs and taken over by the Sharkey devs, but they still have to weave the code changes coming from Misskey in and make them work with what's different on Sharkey.
So it turned out that (Link content warning: eye contact) this bug has already been filed to the Sharkey devs in October, 2024. All that has happened since then until today was that Hazelnoot added two labels. But the bug report came with no explanations. In fact, it misattributed one of my Hubzilla posts as a Friendica post.
And in fact, it turned out that this is actually (Link content warning: Microsoft GitHub link, eye contact) a Misskey bug which has been filed in January, 2024, two years ago. The bug report is a bit more elaborate, but the reporter still knew precious little about what's going on. So I wrote a comment in which I explained the bug from a Friendica/Hubzilla POV as well as what's going on on the technical side, and why the error has to be on Misskey's side.
I hope this will finally help get the bug fixed. Unfortunately, this fix would come too late for Iceshrimp. Iceshrimp-JS is a true Forkey, but in maintenance mode, so I guess only security patches and critical bugfixes will be merged from Misskey, if anything. And Iceshrimp.NET is a complete rewrite of a pre-this-fix Misskey fork, so the Iceshrimp devs probably don't know about this issue either. If it fails ungracefully upon receiving hashtags with the hashtag character outside, it will require its own bug report.
#Long #LongPost #CWLong #CWLongPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #Hashtag #Hashtags #HashtagMeta #CWHashtagMeta #Twitter #𝕏 #Fediverse #Mastodon #Pleroma #Akkoma #Misskey #Forkey #Forkeys #Sharkey #Iceshrimp #Iceshrimp-JS #Iceshrimp.NET #Identi.ca #Laconi.ca #StatusNet #Friendica #Hubzilla #Streams #(streams) #Forte -
CW: Why Sharkey fails to render hashtags from Friendica, Hubzilla & Co. properly, and how long this bug has been known already; CW: long (over 6,600 characters), Fediverse meta, Fediverse-beyond-Mastodon meta
So there's that nasty bug on Sharkey that mangles hashtags in messages from Hubzilla and probably also Friendica, (streams) and Forte. They always look like this:
#[Hashtag](https://hub.netzgemeinde.eu/search?tag=Hashtag)
Basically, Sharkey receives fully standard Rich Text from Hubzilla. It manages to convert this Rich Text into its own Misskey-Flavored Markdown. But then its Markdown parser does not parse it and leaves the Markdown code visible to everyone. It simply doesn't expect there to be a hashtag character in front of an embedded link because, seriously, who'd ever do that and why?!
Friendica would. In fact, Friendica does. It puts the hashtag character in front of the tag, as in outside the tag, as opposed to at the beginning of the tag. It has been doing that since its beginnings in 2010 because it was designed from the get-go to also federate with StatusNet from 2008. And StatusNet does hashtags the same way on its few remaining servers. In fact, so did Identi.ca from 2008, from which StatusNet emerged.
Hubzilla, (streams) and Forte do it, too, because they have inherited it from Friendica.
On StatusNet, Friendica, Hubzilla, (streams) and Forte, a hashtag in a message looks like this:
#Hashtag
Notice how the hashtag character has the same colour as the rest of the post text. And not the same colour as the rest of the hashtag. This means that the hashtag character is not part of the link. (To Mastodon users who don't know this: If something in a "toot" has a different colour from the rest of the "toot", it's a link. Even if it doesn't show a URL in plain sight.)
On 𝕏, Mastodon, Pleroma, Akkoma, Misskey, the various Forkeys and a whole lot of other Fediverse software, a hashtag in a message looks like this:
#Hashtag
Notice how now the hashtag character has the same colour as the rest of the hashtag. This means that the hashtag character is part of the link.
But why did Identi.ca do hashtags differently from Twitter? Because Identi.ca did hashtags before Twitter. AFAIK, when Identi.ca was launched, it had support for hashtags right away. About one year before Twitter.
The hashtag itself had already been invented by the Twitter community. Chris Messina had already codified it in 2007. But it wasn't until 2009 that Twitter actually introduced a technological implementation to support it.
Again, Identi.ca must have had hashtags as early as 2008, and there was no way that Identi.ca creator Evan Prodromou could possibly predict what Twitter would do the following year. So he did what he thought was right and what actually made sense to him.
But nowadays, everybody "knows" that Twitter had the world's very first hashtag implementation ever because nobody, even in the Fediverse, has ever heard of Identi.ca. I mean, the majority of Fediverse users "know" that the Fediverse started with Mastodon.
You know, just like Officer James Barrett "knew" that there is no intelligent life outside Earth only a few minutes before he became Agent J of the Men In Black.
This is also why just about all Fediverse software that does hashtags the Twitter way expects everything to do hashtags the Twitter way. It does not expect hashtags to be done differently. And when a message comes in from Friendica, Hubzilla, (streams) or Forte with hashtags in it, it fails at varying degrees of ungracefully.
Hashtags with the hashtag character outside the link are older than hashtags with the hashtag character inside that they're not only completely unexpected, that they cause software to malfunction, but the same software often can't even handle that malfunction. It's a miracle that the Friendica/Hubzilla family doesn't cause Fediverse servers to crash or even server databases to go corrupt by simply sending hashtags.
Mastodon used to be an exception of sorts, but only because, before version 4.0 from October, 2022, its HTML "sanitiser" actually ripped out any and all rich text code from incoming messages and left nothing but plain text behind. And then it didn't recognise hashtags in messages from outside Mastodon as hashtags at all.
When Mastodon 4.0 came and supported some rich text, including embedded links, it went haywire, of course. But then someone from Friendica and Hubzilla went in and complained about this malfunction and explained what happened, why it happened and why it was not Friendica and Hubzilla that did things wrong. Besides, if something utterly defaces "toots", then Mastodon developers do step in to stop it. After all, Mastodon has a few more of them at hand, paid, full-time professionals even. You have to give it that.
Which takes us back to Sharkey. Sharkey is developed by a small handful of individuals in their spare time. Granted, it's a soft fork of Misskey, so a lot of development work is done by the Misskey devs and taken over by the Sharkey devs, but they still have to weave the code changes coming from Misskey in and make them work with what's different on Sharkey.
So it turned out that (Link content warning: eye contact) this bug has already been filed to the Sharkey devs in October, 2024. All that has happened since then until today was that Hazelnoot added two labels. But the bug report came with no explanations. In fact, it misattributed one of my Hubzilla posts as a Friendica post.
And in fact, it turned out that this is actually (Link content warning: Microsoft GitHub link, eye contact) a Misskey bug which has been filed in January, 2024, two years ago. The bug report is a bit more elaborate, but the reporter still knew precious little about what's going on. So I wrote a comment in which I explained the bug from a Friendica/Hubzilla POV as well as what's going on on the technical side, and why the error has to be on Misskey's side.
I hope this will finally help get the bug fixed. Unfortunately, this fix would come too late for Iceshrimp. Iceshrimp-JS is a true Forkey, but in maintenance mode, so I guess only security patches and critical bugfixes will be merged from Misskey, if anything. And Iceshrimp.NET is a complete rewrite of a pre-this-fix Misskey fork, so the Iceshrimp devs probably don't know about this issue either. If it fails ungracefully upon receiving hashtags with the hashtag character outside, it will require its own bug report.
#Long #LongPost #CWLong #CWLongPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #Hashtag #Hashtags #HashtagMeta #CWHashtagMeta #Twitter #𝕏 #Fediverse #Mastodon #Pleroma #Akkoma #Misskey #Forkey #Forkeys #Sharkey #Iceshrimp #Iceshrimp-JS #Iceshrimp.NET #Identi.ca #Laconi.ca #StatusNet #Friendica #Hubzilla #Streams #(streams) #Forte -
CW: Why Sharkey fails to render hashtags from Friendica, Hubzilla & Co. properly, and how long this bug has been known already; CW: long (over 6,600 characters), Fediverse meta, Fediverse-beyond-Mastodon meta
So there's that nasty bug on Sharkey that mangles hashtags in messages from Hubzilla and probably also Friendica, (streams) and Forte. They always look like this:
#[Hashtag](https://hub.netzgemeinde.eu/search?tag=Hashtag)
Basically, Sharkey receives fully standard Rich Text from Hubzilla. It manages to convert this Rich Text into its own Misskey-Flavored Markdown. But then its Markdown parser does not parse it and leaves the Markdown code visible to everyone. It simply doesn't expect there to be a hashtag character in front of an embedded link because, seriously, who'd ever do that and why?!
Friendica would. In fact, Friendica does. It puts the hashtag character in front of the tag, as in outside the tag, as opposed to at the beginning of the tag. It has been doing that since its beginnings in 2010 because it was designed from the get-go to also federate with StatusNet from 2008. And StatusNet does hashtags the same way on its few remaining servers. In fact, so did Identi.ca from 2008, from which StatusNet emerged.
Hubzilla, (streams) and Forte do it, too, because they have inherited it from Friendica.
On StatusNet, Friendica, Hubzilla, (streams) and Forte, a hashtag in a message looks like this:
#Hashtag
Notice how the hashtag character has the same colour as the rest of the post text. And not the same colour as the rest of the hashtag. This means that the hashtag character is not part of the link. (To Mastodon users who don't know this: If something in a "toot" has a different colour from the rest of the "toot", it's a link. Even if it doesn't show a URL in plain sight.)
On 𝕏, Mastodon, Pleroma, Akkoma, Misskey, the various Forkeys and a whole lot of other Fediverse software, a hashtag in a message looks like this:
#Hashtag
Notice how now the hashtag character has the same colour as the rest of the hashtag. This means that the hashtag character is part of the link.
But why did Identi.ca do hashtags differently from Twitter? Because Identi.ca did hashtags before Twitter. AFAIK, when Identi.ca was launched, it had support for hashtags right away. About one year before Twitter.
The hashtag itself had already been invented by the Twitter community. Chris Messina had already codified it in 2007. But it wasn't until 2009 that Twitter actually introduced a technological implementation to support it.
Again, Identi.ca must have had hashtags as early as 2008, and there was no way that Identi.ca creator Evan Prodromou could possibly predict what Twitter would do the following year. So he did what he thought was right and what actually made sense to him.
But nowadays, everybody "knows" that Twitter had the world's very first hashtag implementation ever because nobody, even in the Fediverse, has ever heard of Identi.ca. I mean, the majority of Fediverse users "know" that the Fediverse started with Mastodon.
You know, just like Officer James Barrett "knew" that there is no intelligent life outside Earth only a few minutes before he became Agent J of the Men In Black.
This is also why just about all Fediverse software that does hashtags the Twitter way expects everything to do hashtags the Twitter way. It does not expect hashtags to be done differently. And when a message comes in from Friendica, Hubzilla, (streams) or Forte with hashtags in it, it fails at varying degrees of ungracefully.
Hashtags with the hashtag character outside the link are older than hashtags with the hashtag character inside that they're not only completely unexpected, that they cause software to malfunction, but the same software often can't even handle that malfunction. It's a miracle that the Friendica/Hubzilla family doesn't cause Fediverse servers to crash or even server databases to go corrupt by simply sending hashtags.
Mastodon used to be an exception of sorts, but only because, before version 4.0 from October, 2022, its HTML "sanitiser" actually ripped out any and all rich text code from incoming messages and left nothing but plain text behind. And then it didn't recognise hashtags in messages from outside Mastodon as hashtags at all.
When Mastodon 4.0 came and supported some rich text, including embedded links, it went haywire, of course. But then someone from Friendica and Hubzilla went in and complained about this malfunction and explained what happened, why it happened and why it was not Friendica and Hubzilla that did things wrong. Besides, if something utterly defaces "toots", then Mastodon developers do step in to stop it. After all, Mastodon has a few more of them at hand, paid, full-time professionals even. You have to give it that.
Which takes us back to Sharkey. Sharkey is developed by a small handful of individuals in their spare time. Granted, it's a soft fork of Misskey, so a lot of development work is done by the Misskey devs and taken over by the Sharkey devs, but they still have to weave the code changes coming from Misskey in and make them work with what's different on Sharkey.
So it turned out that (Link content warning: eye contact) this bug has already been filed to the Sharkey devs in October, 2024. All that has happened since then until today was that Hazelnoot added two labels. But the bug report came with no explanations. In fact, it misattributed one of my Hubzilla posts as a Friendica post.
And in fact, it turned out that this is actually (Link content warning: Microsoft GitHub link, eye contact) a Misskey bug which has been filed in January, 2024, two years ago. The bug report is a bit more elaborate, but the reporter still knew precious little about what's going on. So I wrote a comment in which I explained the bug from a Friendica/Hubzilla POV as well as what's going on on the technical side, and why the error has to be on Misskey's side.
I hope this will finally help get the bug fixed. Unfortunately, this fix would come too late for Iceshrimp. Iceshrimp-JS is a true Forkey, but in maintenance mode, so I guess only security patches and critical bugfixes will be merged from Misskey, if anything. And Iceshrimp.NET is a complete rewrite of a pre-this-fix Misskey fork, so the Iceshrimp devs probably don't know about this issue either. If it fails ungracefully upon receiving hashtags with the hashtag character outside, it will require its own bug report.
#Long #LongPost #CWLong #CWLongPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #Hashtag #Hashtags #HashtagMeta #CWHashtagMeta #Twitter #𝕏 #Fediverse #Mastodon #Pleroma #Akkoma #Misskey #Forkey #Forkeys #Sharkey #Iceshrimp #Iceshrimp-JS #Iceshrimp.NET #Identi.ca #Laconi.ca #StatusNet #Friendica #Hubzilla #Streams #(streams) #Forte -
CW: Why Sharkey fails to render hashtags from Friendica, Hubzilla & Co. properly, and how long this bug has been known already; CW: long (over 6,600 characters), Fediverse meta, Fediverse-beyond-Mastodon meta
So there's that nasty bug on Sharkey that mangles hashtags in messages from Hubzilla and probably also Friendica, (streams) and Forte. They always look like this:
#[Hashtag](https://hub.netzgemeinde.eu/search?tag=Hashtag)
Basically, Sharkey receives fully standard Rich Text from Hubzilla. It manages to convert this Rich Text into its own Misskey-Flavored Markdown. But then its Markdown parser does not parse it and leaves the Markdown code visible to everyone. It simply doesn't expect there to be a hashtag character in front of an embedded link because, seriously, who'd ever do that and why?!
Friendica would. In fact, Friendica does. It puts the hashtag character in front of the tag, as in outside the tag, as opposed to at the beginning of the tag. It has been doing that since its beginnings in 2010 because it was designed from the get-go to also federate with StatusNet from 2008. And StatusNet does hashtags the same way on its few remaining servers. In fact, so did Identi.ca from 2008, from which StatusNet emerged.
Hubzilla, (streams) and Forte do it, too, because they have inherited it from Friendica.
On StatusNet, Friendica, Hubzilla, (streams) and Forte, a hashtag in a message looks like this:
#Hashtag
Notice how the hashtag character has the same colour as the rest of the post text. And not the same colour as the rest of the hashtag. This means that the hashtag character is not part of the link. (To Mastodon users who don't know this: If something in a "toot" has a different colour from the rest of the "toot", it's a link. Even if it doesn't show a URL in plain sight.)
On 𝕏, Mastodon, Pleroma, Akkoma, Misskey, the various Forkeys and a whole lot of other Fediverse software, a hashtag in a message looks like this:
#Hashtag
Notice how now the hashtag character has the same colour as the rest of the hashtag. This means that the hashtag character is part of the link.
But why did Identi.ca do hashtags differently from Twitter? Because Identi.ca did hashtags before Twitter. AFAIK, when Identi.ca was launched, it had support for hashtags right away. About one year before Twitter.
The hashtag itself had already been invented by the Twitter community. Chris Messina had already codified it in 2007. But it wasn't until 2009 that Twitter actually introduced a technological implementation to support it.
Again, Identi.ca must have had hashtags as early as 2008, and there was no way that Identi.ca creator Evan Prodromou could possibly predict what Twitter would do the following year. So he did what he thought was right and what actually made sense to him.
But nowadays, everybody "knows" that Twitter had the world's very first hashtag implementation ever because nobody, even in the Fediverse, has ever heard of Identi.ca. I mean, the majority of Fediverse users "know" that the Fediverse started with Mastodon.
You know, just like Officer James Barrett "knew" that there is no intelligent life outside Earth only a few minutes before he became Agent J of the Men In Black.
This is also why just about all Fediverse software that does hashtags the Twitter way expects everything to do hashtags the Twitter way. It does not expect hashtags to be done differently. And when a message comes in from Friendica, Hubzilla, (streams) or Forte with hashtags in it, it fails at varying degrees of ungracefully.
Hashtags with the hashtag character outside the link are older than hashtags with the hashtag character inside that they're not only completely unexpected, that they cause software to malfunction, but the same software often can't even handle that malfunction. It's a miracle that the Friendica/Hubzilla family doesn't cause Fediverse servers to crash or even server databases to go corrupt by simply sending hashtags.
Mastodon used to be an exception of sorts, but only because, before version 4.0 from October, 2022, its HTML "sanitiser" actually ripped out any and all rich text code from incoming messages and left nothing but plain text behind. And then it didn't recognise hashtags in messages from outside Mastodon as hashtags at all.
When Mastodon 4.0 came and supported some rich text, including embedded links, it went haywire, of course. But then someone from Friendica and Hubzilla went in and complained about this malfunction and explained what happened, why it happened and why it was not Friendica and Hubzilla that did things wrong. Besides, if something utterly defaces "toots", then Mastodon developers do step in to stop it. After all, Mastodon has a few more of them at hand, paid, full-time professionals even. You have to give it that.
Which takes us back to Sharkey. Sharkey is developed by a small handful of individuals in their spare time. Granted, it's a soft fork of Misskey, so a lot of development work is done by the Misskey devs and taken over by the Sharkey devs, but they still have to weave the code changes coming from Misskey in and make them work with what's different on Sharkey.
So it turned out that (Link content warning: eye contact) this bug has already been filed to the Sharkey devs in October, 2024. All that has happened since then until today was that Hazelnoot added two labels. But the bug report came with no explanations. In fact, it misattributed one of my Hubzilla posts as a Friendica post.
And in fact, it turned out that this is actually (Link content warning: Microsoft GitHub link, eye contact) a Misskey bug which has been filed in January, 2024, two years ago. The bug report is a bit more elaborate, but the reporter still knew precious little about what's going on. So I wrote a comment in which I explained the bug from a Friendica/Hubzilla POV as well as what's going on on the technical side, and why the error has to be on Misskey's side.
I hope this will finally help get the bug fixed. Unfortunately, this fix would come too late for Iceshrimp. Iceshrimp-JS is a true Forkey, but in maintenance mode, so I guess only security patches and critical bugfixes will be merged from Misskey, if anything. And Iceshrimp.NET is a complete rewrite of a pre-this-fix Misskey fork, so the Iceshrimp devs probably don't know about this issue either. If it fails ungracefully upon receiving hashtags with the hashtag character outside, it will require its own bug report.
#Long #LongPost #CWLong #CWLongPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #Hashtag #Hashtags #HashtagMeta #CWHashtagMeta #Twitter #𝕏 #Fediverse #Mastodon #Pleroma #Akkoma #Misskey #Forkey #Forkeys #Sharkey #Iceshrimp #Iceshrimp-JS #Iceshrimp.NET #Identi.ca #Laconi.ca #StatusNet #Friendica #Hubzilla #Streams #(streams) #Forte -
CW: Why Sharkey fails to render hashtags from Friendica, Hubzilla & Co. properly, and how long this bug has been known already; CW: long (over 6,600 characters), Fediverse meta, Fediverse-beyond-Mastodon meta
So there's that nasty bug on Sharkey that mangles hashtags in messages from Hubzilla and probably also Friendica, (streams) and Forte. They always look like this:
#[Hashtag](https://hub.netzgemeinde.eu/search?tag=Hashtag)
Basically, Sharkey receives fully standard Rich Text from Hubzilla. It manages to convert this Rich Text into its own Misskey-Flavored Markdown. But then its Markdown parser does not parse it and leaves the Markdown code visible to everyone. It simply doesn't expect there to be a hashtag character in front of an embedded link because, seriously, who'd ever do that and why?!
Friendica would. In fact, Friendica does. It puts the hashtag character in front of the tag, as in outside the tag, as opposed to at the beginning of the tag. It has been doing that since its beginnings in 2010 because it was designed from the get-go to also federate with StatusNet from 2008. And StatusNet does hashtags the same way on its few remaining servers. In fact, so did Identi.ca from 2008, from which StatusNet emerged.
Hubzilla, (streams) and Forte do it, too, because they have inherited it from Friendica.
On StatusNet, Friendica, Hubzilla, (streams) and Forte, a hashtag in a message looks like this:
#Hashtag
Notice how the hashtag character has the same colour as the rest of the post text. And not the same colour as the rest of the hashtag. This means that the hashtag character is not part of the link. (To Mastodon users who don't know this: If something in a "toot" has a different colour from the rest of the "toot", it's a link. Even if it doesn't show a URL in plain sight.)
On 𝕏, Mastodon, Pleroma, Akkoma, Misskey, the various Forkeys and a whole lot of other Fediverse software, a hashtag in a message looks like this:
#Hashtag
Notice how now the hashtag character has the same colour as the rest of the hashtag. This means that the hashtag character is part of the link.
But why did Identi.ca do hashtags differently from Twitter? Because Identi.ca did hashtags before Twitter. AFAIK, when Identi.ca was launched, it had support for hashtags right away. About one year before Twitter.
The hashtag itself had already been invented by the Twitter community. Chris Messina had already codified it in 2007. But it wasn't until 2009 that Twitter actually introduced a technological implementation to support it.
Again, Identi.ca must have had hashtags as early as 2008, and there was no way that Identi.ca creator Evan Prodromou could possibly predict what Twitter would do the following year. So he did what he thought was right and what actually made sense to him.
But nowadays, everybody "knows" that Twitter had the world's very first hashtag implementation ever because nobody, even in the Fediverse, has ever heard of Identi.ca. I mean, the majority of Fediverse users "know" that the Fediverse started with Mastodon.
You know, just like Officer James Barrett "knew" that there is no intelligent life outside Earth only a few minutes before he became Agent J of the Men In Black.
This is also why just about all Fediverse software that does hashtags the Twitter way expects everything to do hashtags the Twitter way. It does not expect hashtags to be done differently. And when a message comes in from Friendica, Hubzilla, (streams) or Forte with hashtags in it, it fails at varying degrees of ungracefully.
Hashtags with the hashtag character outside the link are older than hashtags with the hashtag character inside that they're not only completely unexpected, that they cause software to malfunction, but the same software often can't even handle that malfunction. It's a miracle that the Friendica/Hubzilla family doesn't cause Fediverse servers to crash or even server databases to go corrupt by simply sending hashtags.
Mastodon used to be an exception of sorts, but only because, before version 4.0 from October, 2022, its HTML "sanitiser" actually ripped out any and all rich text code from incoming messages and left nothing but plain text behind. And then it didn't recognise hashtags in messages from outside Mastodon as hashtags at all.
When Mastodon 4.0 came and supported some rich text, including embedded links, it went haywire, of course. But then someone from Friendica and Hubzilla went in and complained about this malfunction and explained what happened, why it happened and why it was not Friendica and Hubzilla that did things wrong. Besides, if something utterly defaces "toots", then Mastodon developers do step in to stop it. After all, Mastodon has a few more of them at hand, paid, full-time professionals even. You have to give it that.
Which takes us back to Sharkey. Sharkey is developed by a small handful of individuals in their spare time. Granted, it's a soft fork of Misskey, so a lot of development work is done by the Misskey devs and taken over by the Sharkey devs, but they still have to weave the code changes coming from Misskey in and make them work with what's different on Sharkey.
So it turned out that (Link content warning: eye contact) this bug has already been filed to the Sharkey devs in October, 2024. All that has happened since then until today was that Hazelnoot added two labels. But the bug report came with no explanations. In fact, it misattributed one of my Hubzilla posts as a Friendica post.
And in fact, it turned out that this is actually (Link content warning: Microsoft GitHub link, eye contact) a Misskey bug which has been filed in January, 2024, two years ago. The bug report is a bit more elaborate, but the reporter still knew precious little about what's going on. So I wrote a comment in which I explained the bug from a Friendica/Hubzilla POV as well as what's going on on the technical side, and why the error has to be on Misskey's side.
I hope this will finally help get the bug fixed. Unfortunately, this fix would come too late for Iceshrimp. Iceshrimp-JS is a true Forkey, but in maintenance mode, so I guess only security patches and critical bugfixes will be merged from Misskey, if anything. And Iceshrimp.NET is a complete rewrite of a pre-this-fix Misskey fork, so the Iceshrimp devs probably don't know about this issue either. If it fails ungracefully upon receiving hashtags with the hashtag character outside, it will require its own bug report.
#Long #LongPost #CWLong #CWLongPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #Hashtag #Hashtags #HashtagMeta #CWHashtagMeta #Twitter #𝕏 #Fediverse #Mastodon #Pleroma #Akkoma #Misskey #Forkey #Forkeys #Sharkey #Iceshrimp #Iceshrimp-JS #Iceshrimp.NET #Identi.ca #Laconi.ca #StatusNet #Friendica #Hubzilla #Streams #(streams) #Forte -
@Herr TurTur @crossgolf_rebel - kostenlose Kwalitätsposts @tux0r :openbsd: Die Idee hinter Sharkey war mal, in einen einzigen Forkey quasi alles auf einmal reinzuschmeißen bis hin zur Küchenspüle. Das fiel vor allem auf, als es noch mehr lebendige Forkeys zur Auswahl gab.
Und in der Zeit hatte Sharkey auch noch gewisse Macken, z. B. eine schon legendär unzuverlässige Implementation der Mastodon Client API, wo es hieß, am besten setzt sich einer hin und schreibt die neu. Wer was Zuverlässiges wollte, setzte damals auf Iceshrimp. Sharkey war für wagemutige Spielkinder, die sich an Hubzilla nicht rantrauten.
Inzwischen ist Calckey tot, Firefish ist tot, Iceshrimp-JS ist im Wartungsmodus (Iceshrimp.NET ist derweil noch meilenweit von "fertig" entfernt), Catodon ist auch schon siech, und das nur für eine Instanz gebaute Hajkey ist schon verwest.
CherryPick hat fast nur Server in den Großräumen Tokyo und Seoul und fühlt sich so auf den ostasiatischen Geschmack ausgelegt an wie vor ein paar Jahren noch Misskey selbst. Vielleicht erinnern sich ein paar noch daran, wie Misskey mal den Charme japanischer Verkaufsautomaten hatte.
Und der ganze noch lebende Rest scheint selbst mit Misskeys Entwicklung nicht mithalten zu können, also im Kern veraltet zu sein.
Derweil ist Sharkey wohl inzwischen ziemlich zuverlässig. So ist es zumindest im Westen zum einzigen noch nennenswerten Forkey avanciert.
#Long #LongPost #CWLong #CWLongPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #Misskey #Forkey #Forkeys #Sharkey #Calckey #Firefish #Iceshrimp #Iceshrimp-JS #Iceshrimp.NET #Catodon #Hajkey #CherryPick -
I really don't know what to do about this server. The development of #Firefish was abandoned last year. If I downgrade to an earlier version, I can then migrate to #IceShrimpJS and eventually #IceShrimpNet. I like this because then I could go back to using @[email protected] on my phone. I'm (reasonably) hesitant to rely on a smaller project again, though. I've got no qualms about setting up and using @[email protected] but starting from scratch would necessitate using a different subdomain.
Why do YOU think I should do? Is there a satisfactory option I've not considered?
#feditips #fedihelp #adviceneeded