kicad-developers team mailing list archive

[Dick Hollenbeck <dick@xxxxxxxxxxx>]

On 09/26/2010 11:04 AM, Wayne Stambaugh wrote:
> On 9/25/2010 1:01 AM, Dick Hollenbeck wrote:
>   
>> On 09/24/2010 01:14 PM, Wayne Stambaugh wrote:
>>     
>>> On 9/24/2010 10:35 AM, Dick Hollenbeck wrote:
>>>   
>>>       
>>>> On 09/24/2010 08:59 AM, Dick Hollenbeck wrote:
>>>>     
>>>>         
>>>>> On 09/24/2010 03:41 AM, Lorenzo Marcantonio wrote:
>>>>>   
>>>>>       
>>>>>           
>>>>>> On Fri, 24 Sep 2010, Martijn Kuipers wrote:
>>>>>>
>>>>>>   
>>>>>>     
>>>>>>         
>>>>>>             
>>>>>>> I would be very happy with an export function, doing: Export all used components into a project library.
>>>>>>> This way I can send the schematics on to the next person and (s)he can adjust where needed, without having to look for some components/footprint differences.
>>>>>>> Of course, its export, so the component would still live in the system library to be re-used.
>>>>>>>     
>>>>>>>       
>>>>>>>           
>>>>>>>               
>>> <<< snipped >>>
>>>
>>>   
>>>       
>>>> Then if you give EESCHEMA the ability to treat any previously designed
>>>> schematic as a library source for a new schematic (remembering that a
>>>> parts list is a library that exists within a schematic), you can pull
>>>> from one parts list into a new one, using drag and drop also.  If you
>>>> want to extend this to the extreme, you can eventually do away with
>>>> library files in total and only have one file format, schematics.  This
>>>> extension is a separate optional design decision obviously.  You could
>>>> also do this and hide it by giving the libraries a different file
>>>> extention, but the formats could be identical.  Library files would not
>>>> need to have any component instantiations or wires.
>>>>     
>>>>         
>>> This is an interesting concept.  I was thinking about embedding the library in
>>> the schematic instead of a separate library file when you suggested a separate
>>> project library file.  I actually like this concept better.  The only down side
>>> I see is that it requires implementing the schematic structure to support this
>>> at the same time which may be too big of a change to implement in a reasonable
>>> time frame.  We could implement the component library first and design in the
>>> capability to embed libraries into other files as part of the library
>>> structure.  Then we could repeat this same exercise for the schematic structure
>>> and make the library to schematic transition then.
>>>
>>> Wayne
>>>   
>>>       
>> If we're basing the new library file on s-expression (aka DSN like) and
>> using richio concepts, I don't think it is hard to assume that that
>> makes loading and saving pretty much context free and relocatable, i.e.
>> nestable.  It should be easy to extend the library file to include
>> sheets, wires, instantiated components, and desired parts list columns,
>> and after all that, call it a new schematic format.
>>     
> I was planning on making the library file an s-expression using richio.
>  I didn't even consider anything else.
>   
>> And if need be a new branch can be started to do any of this work, to
>> avoid exposing folks to a phase where the program is not usable.  As
>> always, I see the rate limiting problem being one of qualified
>> man-hours, and who gets to donate them.
>>     
> Once this discussion dies down, I'll put together a bullet list summary
> and publish the new file format document.  Once that is fleshed out,
> I'll start working on it.  I am more than happy to push my dev branch to
> launchpad if anyone else is interested in helping.  I was thinking of
> using using conditional compilation to enable/disable reading and
> writing the new file format until the code is to keep the code base in a
> reasonably usable state.
>   

Conditional compilation sounds like a reasonable migration path for the
library work. 

I was mentioning the separate branch in respect to a new schematic
format and parts-list based UI, but it will take a few more emails later
in the game to solidify that path and those concepts.  Then we'd have to
find the qualified man-hours to complete the unified goals.


Dick