Looking back througth this thread I think there is a lot of common 
ground, but it is abit of an ad-hoc discussion and I am perhaps the 
biggest culprit here.

Marcus's post was perhaps correct in listing the requirements, but 
before we go forward we should perhaps define the minimum steps required 
on the part of KiCAD, until we know what way that will be going it would 
be stupid to build on arbitary definitions.

Let's state that we need:

1) An unqualified and unlimited list of name/value pairs that can be 
associated with a component and that this listed will eventually be 
available as a data grid in the properties list rather than a list of 8 
fields.

2) It should be possible to define the fields at a library level such 
that default values are allready set when you use a component.

3) The library editor should be able to browse for components on the 
internet and allow us to use the compnent straight from the library 
(perhaps updating a local cache) or to review/edit the component and 
drop it into our personal library.

4) The components should be presented on the internet in a standard read 
only fashion. It is up to individual users to update thier libraries or 
for groups of users who maintain standard libraries to publish new 
releases of thier collections.

5) PCB footprints and editor works in a similar fashion, and adds the 
possibility to associate 3D views which may also have been fished from 
the internet.


At this point it would perhaps be the case to approach KiCAD developers 
with our proposals. I feel a bit cheeky as someone who only started 
using KiCAD yesterday and yet am allready making "demands" on the 
developers who have so kindly made this package available. On the other 
hand we are esperienced professional EDA users who know how EDA tools 
get used in the big wide world, I feel our feedback is important to the 
developers.

BTW, I am not even sure **how** we should approach the developers. Are 
any reading this post?

These requirements do not AFAIK actually require any changes to core 
KiCAD file formats and functionality, just extensions to the component 
editors, they could probably even be developed seperately using the 
KiCAD classes to visualise and edit the components in order to ensure 
future compatibility.

The big issue is the protocol used to present the libraries on the 
internet such that they may be browsed and accessed programatically by 
the editors.

CVS is one possibility, but it requires that either someone maintains a 
centralised multi-project CVS servers (who?) or that individual groups 
or users who want to share libraries must set up thier own server, which 
could be a problem for many, who perhaps just won't bother.

The alternative is the idea of presenting the files in a standard ASCII 
format via HTTP, such that they may be published on any website, 
including free homepages, and with the Wiki being used an index were 
users manually select sites to put into the library editors browsing 
list. In this case it would be up to individual groups to decide how 
they would cope with version management. This makes it is for individual 
users to publish thier work, whilst people wanting to form groups to 
maintain a centralized library will have a bit more work involved. But 
that seems the right way round to me!

BTW, of course it may well be that the developers have allready 
approached the problem of library distribution and maintenance and have 
thier own solutions in the works.

Once again, I think it is important that these issues are clarified 
before we go into details. We started off by saying "let's define those 
fields", but clearly 8 fixed fields just ain't enougth, and having good 
library support is going to be an important issue for the future of KiCAD.






Reply via email to