#numere — Public Fediverse posts

What do you folks think about this grey highlighting of the !!CODE!! sequences in the doc comments in this #NumeRe file? Does it increase readability? Would you prefer another colour?

@jakehamilton we are grateful for any collaboration on #NumeRe or similar things.

We just add changes to #NumeRe to provide a n-dimensional tensor product with tensorprod() (yes, the name is identical to #Octave's implementation and that's for a reason) and an expansion of trace() to provide also the tensor contraction as a generalisation of the matrix trace.

@beka_valentine tbh I think that this is kind of an unfair generalization. We've always been helpful in the past, but to be fair here, we didn't get much chance to so otherwise.

However, with the advent of BS LLMs, the interest in apps like #NumeRe has dropped significantly, which means we might not get any chance in the future.

The #release of the final #NumeRe v1.1.8 is still some weeks away, but we already prepared the new #package #repository including contribution guidelines.

https://github.com/numere-org/NumeRe-Packages

To simplify the contribution process, you can install the packaging tools from the new repository, which will handle the file structure, forking, branch and commit creation as well as the pull request for you.

The next version of #NumeRe will make use of the new #package #repository on #Github. And if you want to have a personal or private repository: you can create and configure your own. We do support GitHub and #Gitlab out-of-the-box!

https://github.com/numere-org/NumeRe-Packages

@libreoffice @concretedog #NumeRe can do that too

Tables in #NumeRe will now finally also handle the "duration" data type correctly. Together with this change, column types will now prefer "value.f64" over "value.cf64" and the user-triggered auto-conversion will now also try to convert a column to categorical values.

We just pushed the implementation for bit-level logic and manipulation functions within #NumeRe.

We just pushed changes to the static code analyzer and the code parser, which will improve the detection of types and errors in the next version of #NumeRe: https://github.com/numere-org/NumeRe/issues/386

Bonus feature: the path object has now methods for handling the path parent.

We added two new objects:
- the "kind-of-missing" dict object (https://github.com/numere-org/NumeRe/issues/376) and
- a ready-to-use logger object (https://github.com/numere-org/NumeRe/issues/397)

Both will be availake in the upcoming full #release of #NumeRe in a few weeks.

Any ideas, which further standard objects might be necessary?

After we released the current release candidate of #NumeRe, there is now the real possibility for community localisations. Right now, we are working on a process for contributing as well as on the actual plug-in for getting the localisations in the first place. And yes, the plug-in will be written within NumeRe.

Just designed a new #hoodie with a #NumeRe specific design. Tell us, what you think: https://sprd.co/3xDxeJq

The production will take about a week, we'll keep you updated, once we have it in our hands.

Some of you might not have noticed that we have a built-in #issue reporter in #NumeRe, so consider this a periodic reminder.

If you have a #feature request or experience some troubles, just use the "Report a problem or a feature request" (for you, the GUI will most likely in English) menu option in the help menu. That will open up a dialog window, where you can enter any details necessary for understanding your inquiry. Our bot will then create the issue for you right in our #Github repo

At least the standard #NumeRe #installer is now available via #winget.

There seem to be problems with the portable one, which we do not really understand, because the contents of both installers are identical, only the installation process is different with the standard one requiring elevation.

The publication of the #NumeRe #release to #winget takes again some time, as they seem to have problems with our installer - which should work perfectly fine - at least it does on our various testing machines.

Do you experience any troubles? Please let us know!

Using the new matrices in #NumeRe, some calculations are now also getting faster like the #flame #fractal below.

What are the major features in this release?

- We moved all matrix operations into the standard calculation and deprecated `matop`
- #NumeRe is now #Unicode aware and handles that by encoding all strings and files using #UTF8

WE DID IT!

Finally, after many months of waiting, we published a new #release of #NumeRe this morning: https://www.numere.org/home/releases#h.2zto4ol8vozo

The binary can be downloaded from Github or SourceForge. We comitted a new manifest to winget, but the PR has not been merged right now, so you need to be patient here.

With the next #release candidate of #NumeRe we will also ship three new example scripts showing

- how general binary files can be read

- how one can translate C file structures into NumeRe code during file import and

- how #JSON files can be traversed using DictStruct instances.

The next two days are reserved for #Nightly tests of #NumeRe on a non-dev machine. We're doing this to ensure that as much bugs are detected as possible

We all know that catching all bugs is not even remotely possible...

We just completed the tests regarding the new deeply integrated matrix calculation mode for #NumeRe (making `matop` obsolete).

The tests for a few new object methods are also completed so as the tests for standard maths functions.

One major change will be the total switch to UTF-8 encoding for all files. Note that this will affect you, if your used non-ASCII characters within your files.

We just completed the tests regarding the new deeply integrated matrix calculation mode for #NumeRe (making `matop` obsolete).

The tests for a few new object methods are also completed so as the tests for standard maths functions.

One major change will be the total switch to UTF-8 encoding for all files. Note that this will affect you, if your used non-ASCII characters within your files.

We have a poll for your!

What do you prefer as update cycles?

#NumeRe #release #updates #softwareDevelopment #lifecycle

How do you like the following outputs of matrices. Can you notice that one of the matrices is an higher-order tensor?

#NumeRe #Matrix #maths

@ataustin Don't think that's necessary. She will most probably use some nicely written #NumeRe code to analyze the data later 😉

@mirsadra we are writing #NumeRe in C++.

@hackaday #NumeRe is also a good alternative, much like GNU Octave and @LabPlot it comes with many features already built-in, but also provides a powerful programming language.

Fun fact, we've also written a blog post about when NumeRe is the better spreadsheet: https://en.numere.org/home/blog/when-numere-is-the-better-spreadsheet

@alternativeto mentiones some of the main new features brought to you with the current #release of #NumeRe called v1.1.7 "Chandrasekhar"

Take a look at: https://alternativeto.net/news/2025/8/numere-1-1-7-brings-new-array-email-and-encoding-functions/

The #Chandrasekhar limit does not influence this new #release of #NumeRe, besides being this version's namesake. It's only about the mass of white dwarfs.

Get the unlimited release via winget or directly at GitHub (https://github.com/numere-org/NumeRe/releases/tag/v1.1.7.2508) or SourceForge (https://sourceforge.net/projects/numere/files/v1.1.7.2508/)