home.social

Search

1000 results for “airflow”

  1. Measuring the Time is a Breeze With This Air Flow Clock - If you’ve ever had surgery, and you’re over a certain age, chances are good you’re familiar with the... more: hackaday.com/2020/06/25/measur #2020hackadayprize #thehackadayprize #rangefinder #clockhacks #arduino #vl53l1x #blower #fan #pid #pwm #rtc

  2. Part 3
    Takamatsu Castle Park

    You can also go inside (after the usual shoe removing).
    It was so cool inside!
    I don't know how, but the older buildings like traditional, old houses (samurai houses), castles, etc. stay so cool and have a great airflow. It is awesome.

    #japan #nihon #日本 #takamatsu #castle #travel #photography

  3. On a related note, @MaQuest ...

    In #Thailand, a cheap bottle crate hack gives #tree #saplings a fighting chance

    Ruth Kamnitzer
    21 Nov 2025

    Excerpt: "Elliott and colleagues had seen examples of air pruning systems while touring nurseries in Australia, and previous studies have shown that air pruning can be beneficial for a number of tropical tree species. However, in Thailand, specialized air pruning equipment has to be imported and is prohibitively expensive.

    "So instead, the researchers cast around for a local solution. The answer was in plain sight.

    " 'If you go to a bar in the countryside, you know, a little restaurant, out the back there’ll be a rubbish pile full of disused bottle crates,' Elliott says. 'You can [also] go down to the #recycling shop and buy them really cheaply.'

    "Each bottle crate fits about 12 bags with saplings, and the structure creates a gap of 2 centimeters (nearly an inch) at the bottom. The researchers wondered if this gap would be enough to create air circulation for air pruning.

    "To test the idea, they raised saplings of five native tree species using three different methods: in crates placed directly on the ground; in crates placed on wire benches (for improved air flow); and the conventional method of a polyurethane bag on the ground. They then planted the saplings in a deforested site within #PraBangKhram #WildlifeSanctuary in southern Thailand."

    news.mongabay.com/2025/11/in-t

    #SolarPunkSunday #Reforesting #Rewilding #ReusingPlastic

  4. #climate #archetecture #Dubai

    "Last year, temperatures in Dubai reached a scorching 51C and outdoor 'feels like temperatures – which take humidity into account – were as high as 62C. Air conditioning is widespread in the United Arab Emirates' (UAE), accounting for more than 70% of electricity consumption in the country during the summer months. However, proponents of traditional, passive cooling techniques say these tried and tested approaches to shading buildings and managing air flow could help people beat the heat without racking up massive electricity bills.

    'The Emiratis built houses that were completely desert-proof," says Noor Ahmed, my tour guide. 'Wind towers were used to catch the cool air from outside and drive out the hot air from inside,' he adds. Gesturing at the labyrinthine alleyways of the old town, Ahmed points out how the high walls protect pedestrians from the blistering sun.

    After Old Dubai emerged from the sands of time in the 1700s, it became a popular residential area when the ancestors of today's Emiratis transitioned from a nomadic to a more settled lifestyle along the Dubai Creek, a natural saltwater inlet that runs through the city.

    The city's original inhabitants designed ingenious homes that could withstand the harsh climate of the Arabian Desert. They used technologies such as wind towers or wind-catchers known as barjeels, enclosed courtyards, latticed windows called mashrabiyas, coral stone houses, and narrow walkways called sikkas.

    'The beauty of Al Fahidi's architecture lies in the fact that it uses several passive cooling techniques that work in tandem to keep the neighbourhood cool,' says Ahmed Al-Jafflah, senior cultural speaker at Dubai's Sheikh Mohammed bin Rashid Al Maktoum Centre for Cultural Understanding, a nonprofit organisation that promotes Emirati culture and history. 'Our ancestors built a holistic architectural system that optimised wind quality, enhanced shade and minimised exposure to the Sun – all of which were crucial to maintaining comfortable temperatures in Old Dubai,' he adds."

    bbc.co.uk/future/article/20250

  5. It's now all set up and placed at least temporarily in my networking nook 😀

    The LanBoy Air might be over a decade old, but it still delivers - server now has airflow, expandability, and style 🖤💛

    Excited to upgrade it further, add more drives + Proxmox down the line.

    Thanks for coming on the ride, and what a ride it was! 😆

    #selfhosting #retrotech #pcbuild #homelab #Antec

  6. The LanBoy Air isn’t just for looks — this behemoth is also:

    • Fully modular, all panels removable
    • Mesh everywhere (even PSU zone)
    • Designed for max airflow
    • It supports up to 15 fans

    Another cool thing about it that it was built around a positive pressure airflow concept for low dust and high cooling.

    👇

    #pcbuild #retrohardware #Antec #modding #SelfHosting #HomeLab

  7. I've set up the rest of my vanilla beans in an informal version of the "turned off dehydrator" that led to vanillin developing on the one forgotten pod described in the post above. It's just a rack on my desk. This spot is dry, out of the sun, gets natural variation in warm and cold during the diurnal cycle, has maximum airflow, and since it's on my desk, I can keep an eye on it.

    The remarkable part is that even tho my house is basically outdoors, the area smells like vanilla. It's heavenly 😊

    Another interesting fact: the vanilla plant that produced the flowers that made these pods last year, well this year it has no flowers at all. If I recall correctly, the flowers were happening last April. I wonder why it hasn't made any this year. Maybe our long rainy season this year (it only just got dry a few weeks ago) delayed the start of flowers? I really don't know.

    #vanilla #gardening #food #plants #hawaii

  8. #prsm #artstudio
    Slowly coming together ! We can actually work in here and the air flow from other areas is enough to keep it warm! Currently switching gears to plumbing but after that it will be covering the ceiling with OSB and routing the air vent for the laser cutter weeee

  9. От таблицы в Excel до собственного инструмента: как мы строили в VK Cloud решение для работы с внешними партнерами

    Обмен данными между компаниями-партнерами при реализации совместных проектов — стандартная практика. Но часто есть сценарии, которые требуют особого подхода — например, из-за необходимости подстраивать формат отображения данных под специфику работы с информацией на стороне партнера. Более специфической такая задача становится, если готовых решений под такие запросы нет. С подобной ситуацией сталкивались и мы в VK. Меня зовут Елена Климанова. Я ведущий дата-аналитик в компании VK. В этой статье расскажу, как и почему мы прошли путь от использования excel-файлов при работе с внешними партнерами-вендорами до создания собственного продукта.

    habr.com/ru/companies/vk/artic

    #vk_cloud #анализ_данных #хранение_данных #BI #clickhouse #airflow #mytracker #redash

  10. The Enshittification of 3D Printing STL Sites: How Maker Repositories Became Content Platforms

    2,834 words, 15 minutes read time.

    There was a time when STL repositories felt like infrastructure. They were messy, imperfect, sometimes unstable, but they served a clear purpose. You went there to retrieve functional designs created by people who understood that tolerances matter, mounting points matter, airflow direction matters, and documentation matters. The mission was simple: share useful objects so others could build on them. That mission was grounded in the open hardware ethos shaped by projects like RepRap and reinforced by licensing systems such as Creative Commons. It wasn’t polished, but it was honest.

    Then the incentives changed.

    The growth of consumer 3D printing brought scale, and scale brought platform economics. Sites like Thingiverse, Printables, Cults3D, and MyMiniFactory evolved from archives into ecosystems. At first, that seemed like maturation. Better interfaces, better hosting, better visibility for creators. But over time, something more subtle happened. Utility stopped being the organizing principle. Engagement replaced it.

    When a repository begins optimizing for clicks, retention, and growth instead of engineering clarity, decay sets in. The interface starts resembling a social feed instead of a technical archive. Thumbnails become louder. Titles become bloated with keywords. Contests and reward systems appear. Download counts become currency. Visibility becomes gamified. None of this is inherently evil, but it fundamentally shifts behavior.

    This is enshittification in slow motion. First, the platform serves the user. Then it serves the uploader. Finally, it serves itself. Somewhere in that transition, the serious maker — the person trying to solve a mechanical problem or improve a machine — becomes collateral damage.

    From Engineering Repositories to Engagement Engines

    Originally, STL sites functioned like decentralized workshops. A model page typically included design intent, dimensions, print settings, and assembly notes. Comments focused on fitment, improvements, or mechanical feedback. The culture leaned technical because the barrier to entry was higher. Early adopters were often building or modifying their own printers. The audience expected competence.

    As adoption widened, the demographic broadened. That expansion was healthy in many ways. More accessibility meant more creativity. However, platforms responded to growth with mechanisms designed for scale, not engineering discipline. Search algorithms began prioritizing popularity signals. Trending sections surfaced based on interaction velocity. Reward systems incentivized uploads. Creator spotlights and contests encouraged constant content generation.

    The result was predictable. Content volume exploded. Signal-to-noise ratio dropped.

    A repository optimized for engagement behaves differently from one optimized for retrieval. Engagement systems reward what generates reaction. Engineering systems reward what functions reliably. Those are not the same thing. A flashy model with dramatic renders and broad compatibility claims generates attention quickly. A precisely dimensioned structural bracket that solves a narrow but real problem generates fewer clicks. When the algorithm decides what most users see, it doesn’t measure mechanical integrity. It measures interaction.

    That distortion shows up everywhere. Titles increasingly read like search-engine bait instead of design documentation. Descriptions stretch compatibility claims beyond reason. Remix chains grow without clear lineage tracking. Files are uploaded without test prints across common material types. Documentation shrinks while promotional language grows. The presentation improves while the engineering substance thins.

    Even when platforms attempt curation, the underlying incentive structure remains engagement-driven. Downloads, likes, and shares influence visibility. Visibility influences behavior. Behavior shapes culture. Over time, culture shifts from solving problems to producing content.

    This transformation doesn’t require malicious actors. It is structural. Platforms must grow to survive. Growth requires participation. Participation is easiest to stimulate with gamification and visibility rewards. But the more a repository behaves like a social network, the less it behaves like technical infrastructure.

    And infrastructure is what serious makers actually need.

    Incentives, Monetization, and the Collapse of Signal-to-Noise

    The introduction of monetization accelerated the drift. When creators can sell files directly, attention becomes revenue. That changes the psychological landscape immediately. The incentive is no longer just to share a useful design. The incentive is to attract buyers. That favors broader appeal over specialized utility. It favors aesthetic novelty over structural refinement. It favors marketing language over restrained documentation.

    Even in platforms that emphasize free sharing, reward systems distort priorities. Point systems, badges, and contests reward upload frequency and download counts. A creator who publishes five minor variations receives more visibility than someone who spends weeks refining one robust design. Rational actors respond to reward structures. The outcome is proliferation of incremental uploads with minimal differentiation.

    Documentation quality declines because it is not directly rewarded. Few platforms require structured metadata for tolerances, material testing, or mechanical validation. There is rarely a standardized field for recommended infill density under load or notes about heat creep in enclosed environments. Those details matter in real-world application, yet they are optional and often absent. The algorithm does not penalize missing rigor, so rigor becomes rare.

    Meanwhile, remix culture compounds fragmentation. Open licenses allow modification, which is essential to collaborative engineering. However, without disciplined version control and clear deprecation practices, remix trees become tangled. Users encounter multiple forks of the same design without clarity on which is current, tested, or abandoned. In software development, version control systems enforce traceability. In STL repositories, that discipline is largely cultural rather than structural. As culture shifts toward content velocity, traceability erodes.

    Centralization magnifies risk. When a handful of platforms dominate hosting, policy changes ripple across the ecosystem. Licensing enforcement varies. Terms of service evolve. API access can be restricted. Files can disappear if moderation policies change or accounts are removed. For a community built on open-source principles championed by organizations like the Open Source Hardware Association, that level of platform dependency introduces fragility. What began as decentralized collaboration increasingly relies on centralized infrastructure with commercial priorities.

    The consequence is not just inconvenience. It is cumulative inefficiency. Time spent filtering noise is time not spent designing, iterating, or printing. Trust erodes when files lack documentation or fail unexpectedly. Newcomers struggle to distinguish quality from hype. Veterans compensate by curating private libraries or retreating to smaller communities where engineering still dominates conversation.

    One personal example illustrates the friction without defining the whole problem. When I sit down looking for a specific printer upgrade, not browsing but targeting a known need, the retrieval process often feels like excavating through content layers designed for engagement rather than precision. That experience is not unique to upgrades. It reflects a broader structural shift in how these platforms function.

    The enshittification of STL sites is not about one bad search result. It is about the slow replacement of engineering-first infrastructure with content-first ecosystems. Until incentives realign around utility, documentation, and traceable iteration, the signal-to-noise ratio will continue to degrade.

    The Hidden Costs: Engineering Decay, Time Erosion, and Cultural Drift

    The most obvious cost of STL platform decay is wasted time, but time loss is only the surface symptom. Beneath that friction sits something more serious: the quiet erosion of engineering standards inside the maker ecosystem. When repositories stop functioning as reliable technical archives, they stop reinforcing good design habits. What fills that vacuum is “good enough,” and “good enough” spreads faster than rigor ever did.

    In a healthy engineering environment, documentation carries weight. You expect dimensional callouts. You expect notes about material choice. You expect disclaimers about stress concentration or thermal expansion when relevant. In the early days of open hardware communities shaped by the RepRap movement, designs were often shared alongside context because the people using them were builders. They were assembling printers from parts, tuning firmware, and troubleshooting mechanical tolerances. That culture naturally demanded explanation. The file was not the whole story. The reasoning behind the file mattered.

    As STL sites scaled into broader audiences, that expectation weakened. Many users now approach models as consumable objects rather than engineering artifacts. That shift is understandable. Consumer printers lowered the barrier to entry, and accessibility is a good thing. However, platforms did not compensate by raising documentation standards. Instead, they lowered friction for uploads. It became easier to post quickly than to explain thoroughly. When publication is frictionless and validation is optional, rigor declines.

    The technical consequences show up in subtle but consistent ways. Models are uploaded without real-world print verification across common materials. Clearances are tuned for one printer configuration and presented as universally compatible. Mounting interfaces lack tolerance guidance. Structural components omit orientation recommendations, leading to predictable layer adhesion failures. None of these flaws are catastrophic on their own. Collectively, they create a culture where mechanical nuance is secondary to file availability.

    That degradation compounds over time. New makers often learn by imitation. If the models they encounter lack documentation discipline, they replicate that behavior when they upload their own work. The repository becomes an echo chamber of partial information. What began as a collaborative engineering commons shifts toward a loosely organized content warehouse.

    There is also the issue of version instability. In software development, version control systems enforce traceability and changelogs. In many STL repositories, revision history is informal or nonexistent. Files are replaced silently. Remixes fork without structured lineage. A design that worked six months ago might be buried under newer uploads with minor cosmetic changes but no mechanical improvement. Without consistent version tagging or deprecation markers, users must reverse-engineer the project history through comments and timestamps. That is not efficient engineering practice. It is guesswork layered on top of guesswork.

    Licensing adds another layer of ambiguity. Creative Commons and GPL-style licenses were designed to enable sharing while preserving attribution and modification rights. However, as monetization enters the ecosystem, license interpretation becomes murkier. Some platforms mix paid models with open-licensed derivatives. Some creators misunderstand the scope of non-commercial clauses. Enforcement varies. The average user navigating this landscape must interpret licensing terms without legal clarity. For a community built on open-source principles, inconsistent license literacy undermines trust.

    Centralization intensifies fragility. When major repositories dominate discovery and hosting, they become single points of failure. Policy changes can alter visibility overnight. Search algorithms can deprioritize older content without warning. API restrictions can limit third-party archiving tools. Even if a platform does not collapse outright, its commercial priorities inevitably influence design decisions. That dynamic creates tension between community stewardship and corporate sustainability.

    The cultural drift is perhaps the most corrosive effect. When a repository feels like a content feed, creators start thinking like content producers. Aesthetic novelty becomes a differentiator. Iteration speed becomes a metric of relevance. The slower, methodical process of engineering refinement struggles to compete with visual spectacle. This does not mean creative or artistic models lack value. They absolutely have a place. The problem arises when the platform structure makes no distinction between decorative novelty and functional hardware. Without structural differentiation, serious engineering competes in the same ranking pool as viral trinkets.

    Over time, that flattening of categories shapes perception. New entrants may not recognize the difference between a mechanically validated component and an untested remix. Veterans compensate through experience, but the ecosystem as a whole becomes noisier and less trustworthy. The friction is cumulative. Every failed print due to missing tolerances, every incompatible mount mislabeled as universal, every abandoned remix chain chips away at confidence.

    This is how enshittification works. It is not a dramatic collapse. It is incremental degradation normalized through scale. Each compromise seems minor. Each engagement feature seems harmless. Each upload without documentation seems tolerable. Collectively, they alter the character of the ecosystem.

    The maker movement was built on iterative improvement grounded in transparency. When transparency declines and iteration becomes performative rather than analytical, the foundation weakens. The tragedy is not that platforms grew. Growth was inevitable and, in many ways, positive. The tragedy is that growth was not matched with structural reinforcement of engineering standards.

    Reclaiming Signal in a Noisy STL Ecosystem

    If the structural incentives of major platforms are unlikely to revert entirely, serious makers must adapt without surrendering standards. The solution is not nostalgia or withdrawal. It is disciplined navigation.

    First, it requires shifting mindset from passive consumption to active curation. Treat STL repositories as raw data pools rather than authoritative archives. That means verifying claims independently. It means reading comment threads critically instead of scanning download counts. It means examining geometry for obvious mechanical weaknesses before committing filament and time. In other words, it requires reintroducing engineering skepticism into the process.

    Second, it means diversifying sources beyond algorithm-driven discovery. Technical communities, open hardware forums, and repositories like GitHub often provide richer context than standalone STL platforms. Projects hosted in code-centric environments tend to maintain clearer version histories and documentation standards because those ecosystems were built around traceability. While not every hardware design lives there, the cultural norms encourage explicit change logs and structured updates.

    Third, it demands building a personal archive. When you identify a well-documented, mechanically sound design, store it locally with version notes. Archive supporting documentation. Preserve license information. Relying exclusively on platform availability is risky in a centralized ecosystem. Maintaining a curated library restores a degree of autonomy. It also reduces repeated exposure to algorithmic noise when revisiting trusted components.

    Finally, it requires cultural reinforcement. When uploading your own designs, model the standards you wish were universal. Provide tolerances. Document material assumptions. Explain orientation rationale. Clarify compatibility boundaries. Reference license terms explicitly. Even if the platform does not reward that rigor directly, the community benefits incrementally. Cultural shifts begin with consistent practice, not platform mandates.

    None of these steps reverse enshittification at the structural level. Platforms will continue to optimize for growth and engagement because that is how they sustain themselves. However, individual and community-level discipline can counterbalance some of the decay. Engineering ecosystems survive when practitioners insist on standards regardless of interface design.

    The future of STL hosting does not have to be bleak. There is room for curated repositories that differentiate functional engineering from decorative content. There is room for structured metadata requirements that elevate documentation quality. There is room for decentralized archiving that reduces single-point dependency. But those improvements require pressure from users who value utility over novelty.

    The enshittification of STL sites is not an irreversible fate. It is the predictable outcome of incentives misaligned with engineering purpose. Realignment will not happen accidentally. It will happen only if serious makers demand that repositories function as infrastructure again rather than infinite scroll feeds for printable plastic.

    Conclusion: If STL Platforms Don’t Realign, Makers Lose

    The enshittification of STL repositories is a slow-motion crisis. It is neither flashy nor catastrophic in the moment. It is incremental, structural, and insidious. When platforms prioritize engagement over engineering, when gamification and monetization distort incentives, when documentation becomes optional and remix chains chaotic, the ecosystem quietly shifts from utility-driven to attention-driven. Serious makers feel the friction in wasted hours, failed prints, and fractured trust. New entrants absorb sloppy habits as the norm. The open hardware ethos erodes, one low-effort upload at a time.

    That decay is not inevitable. It is the predictable outcome of misaligned incentives. Platforms exist to serve users and creators, but currently they serve metrics first. Unless that calculus changes, repositories will continue to favor clicks over precision, aesthetics over tolerances, virality over validation. As long as the architecture rewards performance in an engagement economy, the signal-to-noise ratio will remain unacceptable for anyone who cares about functional 3D printing.

    The solution begins with individual and community action. Curate your own libraries. Rely on technical communities where versioning and documentation are enforced culturally or structurally. Verify designs independently. Push for platforms to implement metadata standards, documentation requirements, and version traceability. Distinguish functional engineering from decorative novelty, and reward the former consistently.

    Growth and engagement will continue. Platforms will not vanish. But serious makers can reclaim control by refusing to normalize decay, by treating STL repositories as technical infrastructure rather than social feeds. If the maker community enforces standards, enforces rigor, and preserves institutional knowledge, STL sites can evolve beyond content-first ecosystems back into the engineering-first archives they were meant to be. That is the only path toward a 3D printing ecosystem that respects both time and craft, instead of turning precision into noise.

    The lesson is simple: stop letting platforms define value through clicks. Stop equating visibility with correctness. Engage critically, archive wisely, and insist on documentation. If we don’t, the culture of enshittification will become permanent, and serious 3D printing will be nothing more than a scroll past digital junk — endless novelty without engineering integrity.

    Call to Action


    If this post sparked your creativity, don’t just scroll past. Join the community of makers and tinkerers—people turning ideas into reality with 3D printing. Subscribe for more 3D printing guides and projects, drop a comment sharing what you’re printing, or reach out and tell me about your latest project. Let’s build together.

    D. Bryan King

    Sources

    Disclaimer:

    The views and opinions expressed in this post are solely those of the author. The information provided is based on personal research, experience, and understanding of the subject matter at the time of writing. Readers should consult relevant experts or authorities for specific guidance related to their unique situations.

    Related Posts

    Rate this:

    #3DModelDocumentation #3DPrinterUpgrades #3DPrinting #3DPrintingBestPractices #3DPrintingChaos #3DPrintingCommunity #3DPrintingCommunityStandards #3DPrintingCulture #3DPrintingFrustration #3DPrintingInefficiency #3DPrintingMistakes #3DPrintingOptimization #3DPrintingProblems #3DPrintingProductivity #3DPrintingStandards #3DPrintingTimeWaste #3DPrintingTrust #3DPrintingTutorials #3DPrintingWorkflow #algorithmicRanking #CreativeCommonsSTL #Cults3D #engagementMetrics3DPrinting #engineeringFirstSTL #makerDiscipline #makerEcosystem #makerMovement #mechanicalDesign #MyMiniFactory #openHardware #openHardwareLicensing #openSource3DModels #Printables #printerMods #printerPartUpgrades #RepRap #STLArchival #STLCentralization #STLContentOverload #STLCuration #STLDocumentationStandards #STLFileQuality #STLFileReliability #STLFileValidation #stlFiles #STLGamification #STLLibraryManagement #STLPlatformDecline #STLPlatformRisks #STLRemixIssues #STLRepositories #STLSearch #STLSearchFrustration #STLSiteDecay #STLUploadIncentives #Thingiverse #versionControlSTL
  11. Chrysler to go all-electric by 2028, starting with the Airflow in 2025 - Enlarge / Yes, it has a light bar, just like pretty much every other EV... - arstechnica.com/?p=1823858 #electriccrossover #electricvehicle #conceptcar #crossovers #stellantis #chrysler #cars #ev

  12. @photography

    A Sea of Lupines by Max Inwood

    The annual lupine bloom in New Zealand is spectacular, with fields of colourful flowers stretching across the Mackenzie Basin. Above the flowers, you can see the band of the outer Milky Way, alongside the constellations Orion, Gemini, and the Pleiades. Joining them are the bright planets Jupiter and Mars, with a strong display of green airglow visible along the horizon.

    #astrophotography
    #MilkyWay
    #lupins
    #BloomScrolling
    #NZ

  13. I just added some wings to my old hd in an old box and also screws in the bottom of the drives to effect a bit more airflow - better than the duct ape brick i had before #heat soak #heat islands #spin down #standby time #disks

  14. CW: Photo of meat. Guanciale

    The first piece of guanciale is ready. It is the thinnest one and already lost just over 30% of its weight. It also was on the higher hook which gets more of the air flow from the wine cellar refrigeration unit. The two thicker pieces look like they will probably need another week and the other thinner piece maybe 3-4 more days.

    I’ll call this another success.

    Might try to hang them somewhere else in the cellar away from the air flow so can cure a bit more slowly just to compare.

    #guanciale #curing #CuredMeat

  15. @AirSupportProject There's no reason we shouldn't be enhancing airflow in classrooms right now with #purifiers or #crboxes or other #aircleaners . We needed good #IAQ yesterday, and we definitely will still need it tomorrow. Read more at: airsupportproject.com/using-th

  16. Super Stealth Sunday! USMC F-35B of VMFAT-502 'Flying Nightmares" 168724 shows airflow over the wings at the Capital Airshow, Mather Field, California March 2025 #airshow #capitalairshow #aviation #AvGeek #spotter #photography #Nikon #nikonphotgraphy #CCA2025 #F35B #USMC #VMFAT502 #Hover #VTOL

  17. Airflow 3 is Coming

    Как-то один из самых главных контрибьюторов в Airflow Ярек Потиюк рассказал, что Airflow 3 станет новым золотым стандартом индустрии. Это довольно смелое заявление. Я же считаю, что в Airflow 3 еще многого не хватает, чтобы действительно стать стандартом. Если вы еще не знаете, что такое Airflow, то, к сожалению, это статья будет сложной. Давайте вместе освежим память. Airflow - это платформа с открытым исходным кодом для написания и управления рабочих процессов. Airflow была основана в 2014 году в AirBnB. С тех пор платформа прошла путь до версии 1.0 в 2015 году, стала Apache Top Level Project в 2019 и плотно обосновалась как Enterprise Production-Ready в 2020 с версией 2.0.

    habr.com/ru/articles/865674/

    #airflow #data_lineage #api

  18. Dagster или Airflow: что выбрать для оркестрации в DWH-проектах?

    Рассказываем, какие задачи решают оркестраторы в проектах внедрения корпоративных хранилищ данных. Выясняем, в чем разница между инструментами, и почему Dagster становится все популярнее в DWH-проектах, чем Airflow.

    habr.com/ru/articles/944284/

    #data_warehouse #dwh #airflow #dagster #оркестрация #оркестратор #data_engineering

  19. Tried #apache #airflow recently, and although it's a good tool, I'm beginning to change my mind for #dagster ...

    UI is quite good, with a lot of details, and the documentation, is quite good.

    Scheduling is less complex, in comparison with Airflow...

  20. Definitely inspired to try #Dagster from last night's #Python meetup here in #Canberra. It looks like a better fit than #Airflow for the #data pipelines I need to build.

  21. Интеграция с ClickHouse: NiFi vs Airflow

    На связи Никита Скирдин, программист 1С компании «Белый код». В прошлой статье мы уже говорили о загрузке данных для системы BI-аналитики. В этой же статье разберем решение задачи с использованием Apache NiFi — системы для автоматизации потоков данных. Хотя NiFi позиционируется как ETL-инструмент (extract transform load), позволяющий внутри себя осуществить необходимые преобразования над поступающими данными, ничто не мешает нам использовать его также для ELT-процесса (extract load transform).

    habr.com/ru/companies/w_code/a

    #интеграция #сравнение #clickhouse #apache_airflow #nifi

  22. Перенос данных из Oracle в PostgreSQL быстро и без потерь: как мы используем для этого Airflow и NiFi

    С необходимостью переноса данных из Oracle столкнулись многие российские компании: в июле 2022 года корпорация, создавшая этот продукт, ушла с российского рынка из-за санкций, как и множество других зарубежных IT-компаний. У пользователей из нашей страны больше нет поддержки вендора, а значит со временем система может перестать корректно работать. Система хранения Oracle была очень популярна в России: данные в ней хранили и обрабатывали даже компании из государственного сектора. И всем нам предстояло быстро решить, куда и каким образом перенести огромные объёмы ценной информации, ничего не потеряв в процессе переноса.

    habr.com/ru/companies/stm_labs

    #миграция #миграция_данных #postgresql #sql #nifi #airflow

  23. Greenplum, NiFi и Airflow на страже импортозамещения: но есть нюансы

    В статье описывается практическое применение популярных Open-Source технологий в области интеграции, хранения и обработки больших данных: Apache NiFi, Apache Airflow и Greenplum для проекта по аналитике учета вывоза отходов строительства. Статья полезна специалистам и руководителям, которые работают с данными решениями и делают ставку на них в части импортозамещения аналогичных технологий. Статья дает обзор основных сложностей внедрения на примере реального кейса, описывает архитектуру и особенности при совместном использовании решений.

    habr.com/ru/articles/810083/

    #данные_в_компании #хранилище #хранение_данных #хранилище_данных #airflow #greenplum #nifi #bigdata

  24. USMC F-35B of VMFAT-502 'Flying Nightmares" 168724 vapor across the wing allows visualization of the airflow. Capital Airshow, Mather Field, California March 2025 #airshow #capitalairshow #aviation #AvGeek #spotter #photography #Nikon #nikonphotgraphy #CCA2025 #F35B #USMC #VMFAT502

  25. USMC F-35B of VMFAT-502 'Flying Nightmares" 168724 vapor across the wing allows visualization of the airflow. Capital Airshow, Mather Field, California March 2025 #airshow #capitalairshow #aviation #AvGeek #spotter #photography #Nikon #nikonphotgraphy #CCA2025 #F35B #USMC #VMFAT502

  26. Агрегация данных для аналитики продаж с помощью DataSphere Jobs и Airflow SDK

    В маркетинге и продажах крупных компаний есть несколько аналитических задач, которые требуют регулярной обработки сотен тысяч и миллионов записей из разных источников. Например, это прогнозирование продаж или планирование рекламных кампаний. Как правило, их решение не обходится без построения длинного пайплайна обработки данных. ML‑инженеру или аналитику данных нужен ансамбль из нескольких моделей и сервисов, чтобы собрать качественный датасет, провести эксперименты и выбрать наиболее подходящие алгоритмы. Сбор, очистка и агрегация данных занимают большую часть времени и вычислительных ресурсов, а эти затраты хочется оптимизировать. В статье покажем, как мы ускорили построение пайплайнов обработки данных с помощью связки DataSphere Jobs и Apache Airflow™.

    habr.com/ru/companies/yandex_c

    #apache_airflow #datasphere #пайплайн #dag