geda-developers team mailing list archive

[Evan Foss <evanfoss@xxxxxxxxx>]

On Fri, Sep 18, 2015 at 11:30 PM, Vladimir Zhbanov <vzhbanov@xxxxxxxxx> wrote:
> On Fri, Sep 18, 2015 at 05:05:12PM -0400, Evan Foss wrote:
> ...
>> > I had been waiting for responses of members of this list. None of them,
>> > besides you and Peter, has answered directly the proposed questions, and
>> > even your answer was indeed a question. So, as I see it, that voting has
>> > not taken place.
>>
>> Sorry yes I vote for the plan. I thought the vote was just for the
>> founding people/people who committed in the near recent past.
> Sorry, I don't understand, what plan are you voting for.
> No, I asked all the people already subscribed to the developer list.
> I thought (probably mistakenly?) they all have developer's privs in our
> repos.

In the older days geda-dev was for discussion of development and
geda-user was for users to ask questions, report bugs and chat. After
the move to launchpad this change was made. When we migrated off the
old lists I changed over to this new idea of geda-dev being more for
people with commit access.

>>
>> (For all DJ's talking about this being the list for people with commit
>> privileges because I don't think I have them to the gEDA repo.)
> That's is what I don't understand. I've always thought people in the dev
> team already have developer privs in one of our repo.

In the old days I was mostly on geda-dev because I was asking
questions trying to understand how the software was designed and why.
Most people were on both lists I think. I was trying to contribute but
keep in mind before the advent of doxygen a lot of questions required
asking people on the list. Back then I joined because I was intending
to go over geda and fill in more of those "Finish function
description!!!" messages in the TODO with actual stuff asking
questions on the list as I went. I have more serious ambitions now but
I still want to do that. I still see libgeda as a tool in it self and
there are other things I want to build off of it to add to the
project. I want to talk more about that but it is a topic for it's own
thread.

Things have been so flamewar heavy lately that I was hoping no one would notice
1. that you included me by name in the voting process with the other
people on geda-developers
2. that i really did not belong there
I figured drawing attention too it would lead to accusations, anger
over who wrote more, who wrote the most valuable parts and etc. I
would have been silent about the xorn & scheme things except that it
felt like what drove Igor2 and I off and I wanted to stop the pattern
from repeating for the good of the project.

>>
>> I guess DJ obtained because he is really more on the PCB side than gEDA.
> I don't understand this either. We are talking about gEDA as a whole
> project, not about its parts. Why we should separate them here? Do our
> strategies differ?

PCB and gEDA were separate projects on separate servers. They are most
often used together and mostly because of that a lot of people who
work on one also are involved in some way on the other. I think no one
ever tried / wanted to sync the two projects releases. They are also
of very different ages. PCB has been re-written a few times. I think
gEDA was written fresh.

The two tools meet at gnetlist and gsch2pcb. If you look at gnetlist
it was written to accommodate more than just PCB. Right now there are
toolchains that people like John Doty have developed that flow from
gschem to mathematica, protel, osmond and a lot of other stuff besides
just PCB. gnetlist supports all of this and I am pretty sure more
stuff via his many contributions in scheme. He has written a lot of
other scheme stuff that he posts on the geda-user list.

gEDA has always assumed everything is a tool.
* Originally libgeda, gschem, gattrib and etc were all packaged
separately because it was thought people will want to use our file
format or some other part of code and might just want that part of the
tool chain. Our current packaging system was picked specifically so
that we could transition back to that if we ever choose too. It also
makes it easier on people who are trying to do a recompile on just one
of the sub tools.
* This was taken so seriously that practically all notions of nets as
larger constructs than just a series of lines was kept out of libgeda
and in gnetlist only. The idea being that it was against the projects
philosophy to blur netlisting into gschem and etc. So gEDA has always
resisted integration with in itself. This is one of many reasons why
making back annotation/notation is so complex. We kind of need some
notion of nets to make gschem provide a meaningful display of the flow
back from pcb. Doty and I have gone back and forth debating this.
Please do *not* unstabilize this it is really hard to get to this
point and I think he might actually help code it when we get to that
stage. We should probably talk off list about it.
* I asked Doty as a joke about how it was that gschem is able to
export images and print schematics. He replied seriously saying it was
a compromise. Seriously purist.

gEDA puts a lot of effort into making sure everything is pure to it's
in keeping with this design philosophy. As you might have noticed John
Doty kind of considers himself to be the keeping of this concern. I
agree with him most of the time by the way. The code style is very
consistent.

PCB integrates things a lot.
* Look at the exporters if that was done by the geda crowd then it
would be done via some external pcbexporter that would have flags for
--gerber, --gcode, --bom, and etc*  * As DJ has said people write code
and if it is good enough it is included. So a lot of features that the
geda crowd would have made separate functions were built into pcb.
* I don't think anyone involved with PCB ever considered anyone
writing utilities around the

There was a push at one point to block contributions that were not in
a style people were going for and this is one of the things that drove
Igor2 away to his own branch back then. (Not to kick a dead horse,
just to give context) Prior to that pretty much anything that was
reasonable was added. DJ is a better source of info on this than me. I
was really obsessed with understanding geda inside and out.

I asked for pcb git access because
1. I wanted to try to track down some bugs.
2. I wanted to try to break Igor2's pcb-rnd branch into a series of
patches that could be considered by the group for inclusion in the
next release of the mainline.

OT: there is a project Igor2, John Doty and I were all getting read to
do after the next stable release.

> That is why I've talked about integration vs. separation of gaf and pcb
> projects before. I'd prefer more integration.

I prefer separation. I want more tools to fill the gaps inbetween and
I was working on some plans for them which we should talk about in one
or more other threads.

>>
>> > Please consider my previous answer (about "majority") a joke.
>> > I understand my position is overwhelmingly minor and I'm biased to agree
>> > with you wrt having xorn in the geda-gaf repository.
>>
>> You have a bias but I was overboard about pointing it out. Sorry.
> Please could you say all this using other words, I really don't
> understand your English here. Thanks.

I went too far too. Sorry.

>>
>> > Thank you Evan,
>>
>> :)
>>
>> > and thank you all developers for your input.
>>
>> It is frustrating that they are all ignoring this. I guess we could
>> import the people we were talking about adding and then put it to
>> another vote but that feels electioneering or rigging the ballot box.
>>
>> I think we should just adopt the plan and move forward.
>
> Agreed. That might be a document we could refer to if we had some moot
> points.

Include a link at the end as a footnote to the email you suggested this plan in.

> Thanks,
>   Vladimir
>
> --
> Mailing list: https://launchpad.net/~geda-developers
> Post to     : geda-developers@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~geda-developers
> More help   : https://help.launchpad.net/ListHelp



-- 
Home
http://evanfoss.googlepages.com/
Work
http://forge.abcd.harvard.edu/gf/project/epl_engineering/wiki/