Mail Archives: geda-user/2017/12/09/09:14:22
Hello geda-users,
Part 1: what's happening now.
I started a process back in autumn 2016: I began to increase the distance
between pcb-rnd and gEDA. This was a gradual process. My project, pcb-rnd,
started as a "pcb fork with a few minor changes" back in 2013. It evolved
from a one-person "pcb fork with a lot of changes" and "pcb fork with
major upgrades" to an independent pcb editor on its own right, developed
by a very productive, growing community.
Under the hood many of the major infrastructure got refactored or even
rewritten/replaced; a lot of new code has been written. On the way a new
community gathered. We have reached a point where only 1/3 of our code
base looks similar to the code I once forked. We got a few side projects
(edakrill, genxproj, cschem) started up.
Our main project, pcb-rnd changed a lot over the past year. Time has come
to move on, and leave gEDA. Well, not that pcb-rnd was ever really a part
of gEDA in any practical way:
- Other than the mailing list, we did not use any of the gEDA
infrastructure
- The gEDA community did not really welcome us; just check how many clicks
it takes from geda main page to figure that pcb-rnd even exist, if you
don't already know the name! We got more contribution and support from
other communities.
- Although I tried multiple times with various strategies, we couldn't
really establish good cooperation with other gEDA projects (except for one
case, Roland merging the back annotation patch). In fact, we found better
cooperation with project/people outside of gEDA.
- Compatibility-wise: we are as compatible with gEDA as with kicad or
protel and lately even with eagle. By now the old pcb/fp file format is
not special to us, just yet another non-native format.
So why pcb-rnd leaves gEDA? The reason is simply the cost/benefit ratio.
I am sorrt if this hurts some gEDA users and contributors, but being
affiliated with gEDA is bad marketing almost everywhere outside the gEDA
community. This became clear as we built a lot of contacts with non-gEDA
users and contributors and projects.
(I've spent a lot of time thinking about it, figuring why, and I have a
theory. If anyone is interested, I am willing to share, but I don't want
to start a flamewar, and it's not my turn to try to change gEDA.)
In return, I never felt pcb-rnd got much from the gEDA project. The usual
misconception is that we got a lot of users because we advertised pcb-rnd
on this list. In reality, what brought majority if our current users and
contributors is a totally different strategy I started to use in 2016 -
and it brought people from random places, only partly from gEDA.
Another important aspect is that the number of users that can be reached
through this mailing list is very very limited. The bigger number of gEDA
users out there who may benefit from switching to pcb-rnd are practically
unaccessible to us.
Which is fine, long term I expect only minority of our target audience
consists of ex-PCB-users. I never intended to replace PCB or get tons of
PCB users. No offense meant, PCB for the win, hail for all the new
releases. But comparing the 1 year progress of PCB and pcb-rnd, I think
the loser of this setup is the user who couldn't learn about pcb-rnd
because they didn't subscribe to geda-user and thought the only pcb editor
gEDA offered was PCB.
I also think pcb-rnd was a great opportunity for gEDA to benefit from the
energy we have and switch gEDA to a higher gear - but for various reasons
the opportunity was missed.
What's next?
What does leaving gEDA mean in practice: we removed all references to gEDA
from the current version of our material, except for the history section.
We do not pretend we did not start from gEDA, we just don't regard
ourselves to be affiliated with gEDA in any way from now. In a sense, with
this move we only make the relationship symmetric (we still have more
references to gEDA this way than gEDA has to us, tho).
With pcb-rnd we are going to walk our own way, more than ever. Good bye
gEDA, and good luck competing kicad and others for new users!
Future of pcb-rnd: I managed to find a way to further increas my time
spent on pcb-rnd and side projects next year. Meanwhile our community is
growing too, more people spending more time on developing pcb-rnd. With
all our new tools and projects, we are building a new (yet to be named),
independent GPL'd EDA suite: pcb-rnd, cschem, edakrill and genxproj are
the framework. It will take years to accomplish this, but this is
definitely the right path for us. One may regard this effort as gEDA 2.0
as we reproduce a lot of ideas/values that once made gEDA successful - but
we also deviate in many values and choices, so it's better to take it as a
new suite.
Is the departure/split/fork inevitable? Almost certainly. In the past two
years I believed pcb-rnd proved that my strategy works well, while gEDA
demonstrated how the old strategy works. I can imagine pcb-rnd stays in
gEDA and we all start to work together on getting gEDA back on track and I
even regroup part of my pcb-rnd time to invest it in gEDA. But the only
construction I could accept for this would be totally unacceptable for
everyone else in the gEDA project. I guess we should just accept that we
want to run our project in totally different ways and split.
How gEDA will relate to us in the future is up to gEDA. I'll write part 2
about this in another mail.
Regards,
Igor2
- Raw text -