The base can always be converted into an integrated library … And how anyone wants to organize this personal matter for everyone … In the end, disabling the unnecessary is always easier than finding the right one) Even so, I would like to see tighter integration with Altium and not with Eagle understandable reasons…
Well, Altium costs about 5000 USD per year, so they can probably afford to have a whole department just for sorting out the library issue. The repository that you linked to above is not complete, either…
But I agree with your comment that there should be a base that can be extended on an individual basis. At the moment, KiCAD’s library is a mixed bag.
Altium gives free libraries, even if not by the company itself, but by third-party manufacturers, I don’t really understand why this is not used for kicad. Write a converter to take the idea even in its own format … It’s like the industry standard dwg format for 2d engineers. The project on the github is solely as an example for implementation and it is free and by the way, the connection from the database with kicad is similar there I wonder what will happen to the libraries in version 8?) The main thing is that once again the format does not change
It seems folks are looking for a universal “plug and play” approach to symbols and foot prints. I’ve been responsible for PCB designs for more years than I care to mention. In EVERY case the target physical component datasheet is considered. And if multiple mfg are to be used each has to be verified. In addition the mfg process (if for high volumes) needs to be considered. (this was probably more important when wave soldering)
So it seems to be having a library that is all things to all people is a fantasy. Perhaps when AI comes to PCB design things will change but for now “the devil is in the details”
As for the current state of Kicad’s libraries, they are a good starting point. I often copy them when making a new part for my own libraries.
Then there is UltraLibrarian etc. I’ve use them often (mostly for the 3D files). But again they go into my personal library which if verify before saving for using.
This is partly solved by online services where the component is already drawn and even if it needs to (usually) be edited, it takes less time than drawing from 0 … And you still need to check this golden rule … On the other hand, if you stick to IPC, then this will reduce the likelihood of errors … For this reason, I suggested that the converter of libraries from different cad is one of the solutions that can work with Wishlist
@davidsrsb wrote:
Library maintainers have always been in short supply.
Kicad has many diligent librarians and first of all, I admire the former commitments of @Rene_Poschl Kicad library is doing its job, but functional improvements could replace many of the missing data. A better library concept could lever the maintainers efforts and shrink the library size by decades.
Introduction of inheritance was only a small step in the correct direction. From time to time there are some threads and ideas of a third kind of library between symbol and footprint what would be the correct attempt. This 3rd kind of library is pure Ascii information, links the required graphics from symbols and footprints and contains all other informations like swapping rules or vendor numbers. Very good concepts of proper library data structures exist for a long time. E.g. Library Concept | LibrePCB is using some of them and even old Ecad like Expedition show very good maintainable structures as they were introduced to save memory with limited PC hardware long ago.
Personally I am curious to play around and find out what can be done with the new database library approach what is also a small step in correct direction. Mainly such library concept it is not a job for a single developer. Same with the bug reports what always covers only one aspect, e.g. a padstack library what we dont have.
Assume all, with the experience of several company libraries, a library always blocks the creativity of a circuit designer. If there are too less components, the designer needs to switch his job with a librarian.If there are too many components its almost the same. E.g. if you have 50 diffrent 100nF caps and you cannot find a “normal” one as you do not know the history of all 50 variants, the situation frequently ends with 51 capacitors of 100nF in revised library and a purchaser what asks if he is allowed to order a 0,1uF model therefore. This way, best recommendation is to maintain its own personal library and consider Kicad as a free labority for new technology.
There is one big drawback with large libraries, especially when they are in the form of a database … With a size of, say, 2-3 GB, indexing and searching for components turns into flour with local ones, it’s a little better there, only indexing on disk, but after caching you can already work. For myself, I concluded that disabling sections that are not needed for the project saves the situation a little. This is observed not only in kicad. This is a problem when you have many small files.
You are assuming that people would have selected to have all the databases(s) available?
I don’t see why a better-organised database could not be smaller. Currently, most of this data is in the database. The original intent of the threat was to find a way to make it easier to add and update symbols.
Well, why are all available at once?) From the post above, the database is only 165,000 components) and there are far from all … it takes about 2.3 GB more to step models … When using a database, the search speed is less than that of a regular local one, and here base size has a big impact
The step models are an irrelevance for a lot of people, it is also something that does not need to be searched as individual elements would be lined from footprints. The only search that was needed was when the step was added to the footprint. Once the footprint was used it would be locally cached.
The keyphrase after it is used … To do this, you need to find it, and here the bottleneck in the database begins with a large and long delay, and the larger the database, the greater the delay.
Kicad libraries are simple and powerfull. When something is missing it is always easy to make your own symbols/footprints.
I think the libraries are more important to the occasional user. In that regard it could be argued that too much time is devoted to them. 
2 Likes
And here, not everything is so smooth, because with each new version of the kikad, the library format changes, which, frankly, is not very cool … I honestly don’t really understand why draw something that already exists or, let’s say, is available in finished form …
This was my favorite comment from last time the subject of libraries was aired:
Are people really still whining about the supplied libraries? Still wanting to be spoonfed?
All Kicad should be supplying is a basic set of symbols, and a set of common standard footprints.
For everything else - make it yourself. Its pretty simple.
I’m making new symbols and footprints all the time for Altium, and we have amassed a vast library as a result. To our specs. I don’t expect Altium to supply anything at all - and its an expensive cad package. (Hell - I’m just happy if it doesn’t crash).
If anyone is interested, here’s a link to the whole thread:
it may not be very cool that the format changes from time to time but changes are sometimes necessary to get improvements. also I use kicad since v4 and the libraries were one of the things which did not break between versions (yet). The only thing that did break were 3rd party tools, converters etc. which did not adapt to the new structure. this also would be a problem for a potential altium library converter.
because the “finished” form isn’t really a finished form as everyone has his own standards on what a symbol needs to satisfy. I can’t remember going to a different library, taking a symbol and having not to modify it intensively afterwards because it was missing information, did not comply with style requirements etc. therefore it is often simply faster to make the symbol from the ground up instead of breaking it down to the foundation and repairing all the issues, which is more or less a process of redoing the whole symbol anyway.
Well, yes, you draw your library for a whole year, use third-party services, create a database from them, and then a new version comes and everything stops working) there is no clear answer why you change the format)) You can do everything with your hands and feet, but this time and this is the most expensive thing for we have …
you exaggerate (again). the most essential parts (kicad working with the libraries) never stopped working and the reasoning for the changes are clearly documented 
Sure some dont use X. Though you have to understand that the STEP integration is the only reason i use kiCad in the first place. No step in libraries and software would mean i would use something else.
But like I say its a red herring, everybody and their friend want different things. Most people who complain about libraries want somebody else to do their work for them.
1 Like
Well, for example, the created libraries in version 6 were successfully converted to version 7, but some parameters require manual changes for each component, to say that it does not work is probably not true, it works but crookedly)