5 ms·

> And thanks to automated tools, the free offline-capable KiCAD library is now exceeding the scope of the paid online-only Altium vault. Well, you say this but

> And thanks to automated tools, the free offline-capable KiCAD library is now exceeding the scope of the paid online-only Altium vault. Well, you say this but as a KiCad dev I will say our library team is heavily understaffed and entirely burnt out. Very few library submissions have been processed in over a year. The automation is there to validate agaisnt rules but there's some that still need to be checked manually, i.e. actually confirming everything matches datasheets, etc. It's a thankless job.

fxtentacle 5y ago
That's unfortunate to hear. I agree that it's a thankless job, which is probably why companies usually pay for a service to do it instead. SnapEda charges $60 for schematic, footprint, and a simplistic 3d model. KiCAD could offer the same service but with the added benefit that components are added to the library afterwards.
- delfinom 5y ago
  Here's the thing. We aren't an organization and most of us would probably be prohibited from joining such under the usual employment contracts. Nobody is making money from it as is. The librarians themselves are entirely self organized and basically work independently of the code base developers. They are generally not programmers at all (and don't need to be). While there are some efforts underway to start slushing around the donation balance we have gained, it takes time when everyone is largely a volunteer with limited time.
contingencies 5y ago
I run a hardware company and mandate all boards are done in KiCad. That means I have invested a lot of money in to KiCad and given monetary incentive for EEs to learn it/use it. We love it. Automation, git integration, small size on disk, runs on all platforms, it's great! Thanks for your hard work. Recently due to chipageddon and imminent production of our major equipment we have been reaching for some less common (but available) components. Therefore we've started drawing up symbols/footprints/3D models more regularly. Even basic stuff is sometimes not there. I naturally want to contribute back but noticed: (1) https://github.com/KiCad/kicad-footprints https://github.com/KiCad/kicad-footprints is now read-only on Github but had 336 unanswered pull requests. Then moved to Gitlab, where it has 412 merge requests. Meanwhile, existing user pulls from Github aren't getting new data. (2) https://github.com/KiCad/kicad-symbols https://github.com/KiCad/kicad-symbols is now read-only on Github but had 217 unanswered pull requests. Then moved to Gitlab, where is has 304 merge requests. Meanwhile, existing user pulls from Github aren't getting new data. (3) https://github.com/KiCad/kicad-packages3D https://github.com/KiCad/kicad-packages3D is now read-only on Github but had 75 unanswered pull requests. Then moved to Gitlab, where it has 93 merge requests. Meanwhile, existing user pulls from Github aren't getting new data. Things I take from this: (1) Fundamentally the transition to Gitlab (which I support!) could have been handled better by KiCad, and this is harming community engagement - eg. you should either keep them up to date as mirrors of Gitlab or shut down the old Github repos and redirect users to Gitlab. (2) Gitlab requires SSH keys to allow cloning, which breaks for most scripts/automation (by design). This should have been considered deeply before the move with a strategy to mitigate the user impact. (3) The current de-facto UX of having to mystically discover a move happened, manually creating a Gitlab account, upload an SSH key and manually re-clone three repos in-place while preserving KiCad path configuration is a poor user experience to say the least, which less technical users will not be able to do, and should preferably have been handled automagically by KiCad itself using rsync or something. Asking devs to validate submissions is insane. Instead, consider to invest some time to create an off-git unified submissions process which enables users to verify one another's uploads, sort of like a captcha, ie. "does this pin on this footprint match this datasheet?" or "do the overall dimensions of this package match this datasheet?". This would allow for submissions but require some community engagement on upload to lessen the load for existing maintainers. To make up for potential inaccuracy, just fire the same content at a few people to verify, or even hand out open source experience library helperships to students: KiCad summer of datasheets. My 2c. Keep up the great work!
- delfinom 5y ago
  >Asking devs to validate submissions is insane. 1. The pulls got copied over to gitlab. 2. Yea the gitlab library transition was not clean...technically it was never finished. Once we get to RC1 for V6. I myself hope to switch over fixing up all our web content instead of chasing down Windows platform bugs. 2. The codebase devs actually have no involvement with the libraries. The libraries are already community ran. The problem is most people would rather file issues asking for new parts rather than finding a way to help. 3. Most everyone is a volunteer largely except for the two that make money doing paid support which is a job in of itself. We can't organize interns because none of us have the interest in burning more free time than we already do. 4. The current mindset of the librarians has been _quality_. Because bad PCBs due to component issues can be expensive and time consuming problems. And issues with components will not only harm trust of the libraries but also KiCad 5. Could we develop some sort of less-gitlab-dependent web system? Sure. Do we have the engineers with free time to do it? Absolutely not in a million years. I dream of doing it, I certainly have the ability to but I need to have a life too ;) The bigger issue is the industry, hardwarists can have little or no programming background or don't feel comfortable outside whatever low level programming they do. So I don't fault them but that's the hand being dealt. This is a niche industry tool with niche demographic.
  - contingencies 5y ago
    Thanks again. Happy to help on 5. Merge ideas @ https://github.com/globalcitizen/kicad-fully-secure-unified-submission-system https://github.com/globalcitizen/kicad-fully-secure-unified-...
    - delfinom 5y ago
      Not a fan of getting involved on github. There are reasons we use gitlab and it's mainly to avoid the daily github drama and noise. The biggest obstacle here is to get everyone else to buy in, mainly the librarians....who I'm honestly not sure how to hunt down for a conversation. Core devs probably don't want to deal with anything major/new until v7 next year at this point.

fxtentacle 5y ago

That's unfortunate to hear. I agree that it's a thankless job, which is probably why companies usually pay for a service to do it instead. SnapEda charges $60 for schematic, footprint, and a simplistic 3d model. KiCAD could offer the same service but with the added benefit that components are added to the library afterwards.

delfinom 5y ago
Here's the thing. We aren't an organization and most of us would probably be prohibited from joining such under the usual employment contracts. Nobody is making money from it as is. The librarians themselves are entirely self organized and basically work independently of the code base developers. They are generally not programmers at all (and don't need to be). While there are some efforts underway to start slushing around the donation balance we have gained, it takes time when everyone is largely a volunteer with limited time.

contingencies 5y ago

I run a hardware company and mandate all boards are done in KiCad. That means I have invested a lot of money in to KiCad and given monetary incentive for EEs to learn it/use it. We love it. Automation, git integration, small size on disk, runs on all platforms, it's great! Thanks for your hard work. Recently due to chipageddon and imminent production of our major equipment we have been reaching for some less common (but available) components. Therefore we've started drawing up symbols/footprints/3D models more regularly. Even basic stuff is sometimes not there. I naturally want to contribute back but noticed: (1) https://github.com/KiCad/kicad-footprints https://github.com/KiCad/kicad-footprints is now read-only on Github but had 336 unanswered pull requests. Then moved to Gitlab, where it has 412 merge requests. Meanwhile, existing user pulls from Github aren't getting new data. (2) https://github.com/KiCad/kicad-symbols https://github.com/KiCad/kicad-symbols is now read-only on Github but had 217 unanswered pull requests. Then moved to Gitlab, where is has 304 merge requests. Meanwhile, existing user pulls from Github aren't getting new data. (3) https://github.com/KiCad/kicad-packages3D https://github.com/KiCad/kicad-packages3D is now read-only on Github but had 75 unanswered pull requests. Then moved to Gitlab, where it has 93 merge requests. Meanwhile, existing user pulls from Github aren't getting new data. Things I take from this: (1) Fundamentally the transition to Gitlab (which I support!) could have been handled better by KiCad, and this is harming community engagement - eg. you should either keep them up to date as mirrors of Gitlab or shut down the old Github repos and redirect users to Gitlab. (2) Gitlab requires SSH keys to allow cloning, which breaks for most scripts/automation (by design). This should have been considered deeply before the move with a strategy to mitigate the user impact. (3) The current de-facto UX of having to mystically discover a move happened, manually creating a Gitlab account, upload an SSH key and manually re-clone three repos in-place while preserving KiCad path configuration is a poor user experience to say the least, which less technical users will not be able to do, and should preferably have been handled automagically by KiCad itself using rsync or something. Asking devs to validate submissions is insane. Instead, consider to invest some time to create an off-git unified submissions process which enables users to verify one another's uploads, sort of like a captcha, ie. "does this pin on this footprint match this datasheet?" or "do the overall dimensions of this package match this datasheet?". This would allow for submissions but require some community engagement on upload to lessen the load for existing maintainers. To make up for potential inaccuracy, just fire the same content at a few people to verify, or even hand out open source experience library helperships to students: KiCad summer of datasheets. My 2c. Keep up the great work!

delfinom 5y ago
>Asking devs to validate submissions is insane. 1. The pulls got copied over to gitlab. 2. Yea the gitlab library transition was not clean...technically it was never finished. Once we get to RC1 for V6. I myself hope to switch over fixing up all our web content instead of chasing down Windows platform bugs. 2. The codebase devs actually have no involvement with the libraries. The libraries are already community ran. The problem is most people would rather file issues asking for new parts rather than finding a way to help. 3. Most everyone is a volunteer largely except for the two that make money doing paid support which is a job in of itself. We can't organize interns because none of us have the interest in burning more free time than we already do. 4. The current mindset of the librarians has been _quality_. Because bad PCBs due to component issues can be expensive and time consuming problems. And issues with components will not only harm trust of the libraries but also KiCad 5. Could we develop some sort of less-gitlab-dependent web system? Sure. Do we have the engineers with free time to do it? Absolutely not in a million years. I dream of doing it, I certainly have the ability to but I need to have a life too ;) The bigger issue is the industry, hardwarists can have little or no programming background or don't feel comfortable outside whatever low level programming they do. So I don't fault them but that's the hand being dealt. This is a niche industry tool with niche demographic.
- contingencies 5y ago
  Thanks again. Happy to help on 5. Merge ideas @ https://github.com/globalcitizen/kicad-fully-secure-unified-submission-system https://github.com/globalcitizen/kicad-fully-secure-unified-...
  - delfinom 5y ago
    Not a fan of getting involved on github. There are reasons we use gitlab and it's mainly to avoid the daily github drama and noise. The biggest obstacle here is to get everyone else to buy in, mainly the librarians....who I'm honestly not sure how to hunt down for a conversation. Core devs probably don't want to deal with anything major/new until v7 next year at this point.