Mail Archives: geda-user/2017/02/12/13:08:36
On Sun, 12 Feb 2017, Britton Kerin (britton DOT kerin AT gmail DOT com) [via geda-user AT delorie DOT com] wrote:
> On Sun, Feb 12, 2017 at 1:02 AM, Roland Lutz <rlutz AT hedmen DOT org> wrote:
>> On Sun, 12 Feb 2017, gedau AT igor2 DOT repo DOT hu wrote:
>>>
>>> just watched your 2016 FOSDEM talk - funny thing is that we are actually
>>> doing a lot of things in pcb-rnd that you were talking about in that
>>> presentation about how EDA tools should cooperate. I'd say there's a 3/4
>>> match and only 1/4 mismatch (or disagreement) between what we are
>>> practically doing and what you described there.
>>
>>
>> Yeah, I noticed that, too. I think that if all people in the project
>> spelled out their intentions in a high-level way, we would find that they
>> aren't very different at all.
>>
>> The problem is that we waste a lot of time and energy on bike shed issues.
>> When people see that someone addressed an issue which they themselves want
>> solved, too, but they would've solved them in another way, then they're
>> trying to make that solution go away, either by ignoring it or, if that
>> doesn't work, by attacking it.
>>
>> Take for example the experimental branch I posted some time ago which
>> implements some netlister features which have been on the wishlist for quite
>> some time: working buses, parametric subschematics, and customizable power
>> symbols without a trailing ":1". For every one of these features, people
>> have criticized me for not implementing it their own way: for buses,
>> everyone had their own idea about numbering pins, someone preferred to use
>> simple net objects instead of buses, and someone didn't like the idea of
>> working buses at all. For parametric subschematics, we disagreed about
>> which parameter name and separator character to use. And even about the
>> trailing ":1", which I had expected to be really non-controversial, there
>> has been an argument yesterday.
>>
>> So I basically have the options to either merge my patches even though there
>> are people opposing them, or to not merge them and not have the features in
>> gEDA/gaf.
>
> Not that it matters, but I vote for merging them. I've pretty much
> given up on contributing to pcb in part because the effective
> resistance to merge is so high. Even with pretty exhaustive test
> cases mainly bug-fix stuff that I've spent quite a bit of time on
> never makes it in. I'm not saying it's all intentional hostility that
> causes this, but whatever the reason the contrast with pcb-rnd is
> painfully clear
Thanks, I'm actively trying to keep the barrier low for pcb-rnd.
> (maybe only because pcb-rnd is still largely a one-man
> show, I don't know).
Haha, this again... I know it will be me again getting CoC'd, and
not those who spread things like that, but...
pcb-rnd facts:
- at the moment there are 12 users with commit right
- just today there were commits by 4 users
- this is happening all the time - we have at least 4 users commiting
every week
- all these easily add up to 500..1000 commits per release
- plus we have some users/testers continously contributing (they
are not in the above statistics because they don't yet have commit rights)
- currently we have less than 50% code lines that came from the original
mainline at fork and didn't see significant change (or is not totally new)
So, debunking some misconceptions I see popping up from time to time:
- pcb-rnd is not a one-man-show anymore. It does have leadership, which
might be a strange concept in some projects, tho
- pcb-rnd is not just small changes anymore
- pcb-rnd is not something small or slow or temporary; it's going to say
for long and is one of the most lively places around here. Hate it as much
as you want, it's still happening. You should better learn to live with
it.
Regards,
Igor2
- Raw text -