Discussion:
RFC: Development forks plus an integration fork
(too old to reply)
Wo'O Ideafarm
2015-05-16 19:23:20 UTC
Permalink
Raw Message
Stylized facts:

(1) Two OW repositories are being actively developed: Jiri Malak's "open-watcom-v2" repository on GitHub, and the "official" OW 1.9 Perforce repository, hosted on a server provided by Perforce. Other OW repositories exist whose level of activity is not known to me.

(2) Jiri Malak is adding 64 bit functionality using a methodology that is perceived by the 1.9 contributors as risky and without an adequate QA/QC process. In turn, the 1.9 contributors are producing occasional bug fixes and isolate enhancements, using a methodology that is perceived by the 2.0 contributors (Malak and company) as cumbersome and ill suited to open source development with extremely limited manpower.

(3) The two development/maintenance efforts disagree about method, not about vision.

IMO, there is merit to the views of both contributor groups. IMO, what is needed is a QA/QC "department" external to both efforts, that focuses on (1) doing or coordinating testing, (2) emitting "pull requests" to encourage bug fixes to be accepted by both efforts, and (3) encouraging integration of each enhancement produced by one group into the sources of the other group.

In order to try this out, I contemplate mirroring the 1.9 Perforce repo to GitHub and also mirroring Jiri's 2.0 repo. Both mirrors would be owned by the "IdeaFarm (tm) Operations" organization on GitHub, which would serve as the QA/QC department. This QA/QC department would have a passive bug fixing role and would encourage (1) acceptance of bug fixes and enhancements by all competing development repos, and (2) preservation and development of both/all competing development visions.

End users would be able to go to a single place on GitHub to obtain binaries and sources for all variants of OW. Contributors who use Perforce would be able to continue to do so. Contributors who want to use GitHub but who prefer to contribute to the 1.9 effort rather than to Jiri's 2.0 effort would be able to do so.

WDYT?
Paul S. Person
2015-05-17 17:15:59 UTC
Permalink
Raw Message
On Sat, 16 May 2015 12:23:20 -0700 (PDT), "Wo'O Ideafarm"
Post by Wo'O Ideafarm
(1) Two OW repositories are being actively developed: Jiri Malak's "open-watcom-v2" repository on GitHub, and the "official" OW 1.9 Perforce repository, hosted on a server provided by Perforce. Other OW repositories exist whose level of activity is not known to me.
That's "Malik", BTW.

Shouldn't you know how to spell your hero's name?
Post by Wo'O Ideafarm
(2) Jiri Malak is adding 64 bit functionality using a methodology that is perceived by the 1.9 contributors as risky and without an adequate QA/QC process. In turn, the 1.9 contributors are producing occasional bug fixes and isolate enhancements, using a methodology that is perceived by the 2.0 contributors (Malak and company) as cumbersome and ill suited to open source development with extremely limited manpower.
I can only judge by what I see, and what I saw when Jiri was
"improving" wgml was quite enough to justify "risky and without
adequate QA/QC process".

Although calling it a "methodology" seems unduly polite to me. And,
instead of "adequate", I would say "any".

The only "cumbersome" bit about my methodology is that /I actually
test my code before committing it/. I find that nothing is more
satisfying than running that final test and finding that the code
finally works properly (not just compiles, although that is not a
given with Jiri either, but actually works properly). Try it some
time; you may find that you like it too.

(For my current effort, "works properly" means "removes the current
diff without introducing any prior diffs. Following diffs remain to be
worked on.)
Post by Wo'O Ideafarm
(3) The two development/maintenance efforts disagree about method, not about vision.
I'm not sure about that; care to compare the "visions"?

Our official list is located here:
http://www.openwatcom.org/index.php/Version_2_Todo_List

Does Jiri have an actual list, or is that too "cumbersome"?
Post by Wo'O Ideafarm
IMO, there is merit to the views of both contributor groups. IMO, what is needed is a QA/QC "department" external to both efforts, that focuses on (1) doing or coordinating testing, (2) emitting "pull requests" to encourage bug fixes to be accepted by both efforts, and (3) encouraging integration of each enhancement produced by one group into the sources of the other group.
One of the strengths (and weaknesses) of OW has /always/ been that
there is nobody in charge. People are free to work on what interests
them.
Post by Wo'O Ideafarm
In order to try this out, I contemplate mirroring the 1.9 Perforce repo to GitHub and also mirroring Jiri's 2.0 repo. Both mirrors would be owned by the "IdeaFarm (tm) Operations" organization on GitHub, which would serve as the QA/QC department. This QA/QC department would have a passive bug fixing role and would encourage (1) acceptance of bug fixes and enhancements by all competing development repos, and (2) preservation and development of both/all competing development visions.
That's an awful lot of complexity; and the QA/QC department sounds
like something that would be spending most of its time cleaning up
Jiri's little code messes for him.

Just out of curiousity, are you volunteering to do this? Or would that
be too "cumbersome"?
Post by Wo'O Ideafarm
End users would be able to go to a single place on GitHub to obtain binaries and sources for all variants of OW. Contributors who use Perforce would be able to continue to do so. Contributors who want to use GitHub but who prefer to contribute to the 1.9 effort rather than to Jiri's 2.0 effort would be able to do so.
End users can download whichever version they want.

And most of them should only need one or the other.

Well, provided Jiri's version hasn't regressed compared to ours, of
course. If it has, there might be a class of users that needs both, I
suppose.
--
"Nature must be explained in
her own terms through
the experience of our senses."
Peter Chapin
2015-05-17 17:39:57 UTC
Permalink
Raw Message
Post by Paul S. Person
Post by Wo'O Ideafarm
(1) Two OW repositories are being actively developed: Jiri Malak's
"open-watcom-v2" repository on GitHub, and the "official" OW 1.9
Perforce repository, hosted on a server provided by Perforce. Other OW
repositories exist whose level of activity is not known to me.
That's "Malik", BTW.
Shouldn't you know how to spell your hero's name?
It is, in fact, Malak.

Peter
Paul S. Person
2015-05-18 16:42:14 UTC
Permalink
Raw Message
Post by Peter Chapin
Post by Paul S. Person
Post by Wo'O Ideafarm
(1) Two OW repositories are being actively developed: Jiri Malak's
"open-watcom-v2" repository on GitHub, and the "official" OW 1.9
Perforce repository, hosted on a server provided by Perforce. Other OW
repositories exist whose level of activity is not known to me.
That's "Malik", BTW.
Shouldn't you know how to spell your hero's name?
It is, in fact, Malak.
I stand corrected.
--
"Nature must be explained in
her own terms through
the experience of our senses."
Wo'O Ideafarm
2015-05-18 07:56:33 UTC
Permalink
Raw Message
Post by Paul S. Person
are you volunteering to do this?
Yes, yes, of course, Mr. Person. My perspective, motives, and interests differ from Jiri's and also perhaps from yours. I am new to open source and have much to learn here, but I hope that you and Jiri and the other contributors find that my involvement makes your involvement more fun, not less. I am not a manager. I have no interest in controlling anyone.

Think of me as a user who is extremely motivated to ensure that OW remains alive and well. I am an independent C++ software developer who is so passionate about C++ and so filled with fury about the course that the software craft has been forced to take by that autistic sociopath William Gates that I wrote an operating system that I hoped would destroy Microsoft's monopoly. That operating system is strategically committed to OW. If OW dies, my own life work dies with it.

That's where I'm coming from. I think that I have something valuable to contribute to OW. Part of that value will, I hope, be to somehow get everyone back on speaking terms so that we are not working against each other.

Thanks for the informative response. It helped me to see more clearly what the interpersonal issues are and what the fundamental disagreements are.

Life is short. Let's all be friends. OW is a beautiful thing that deserves to survive. If we can work together, that will be another beautiful thing.
Paul S. Person
2015-05-18 16:51:10 UTC
Permalink
Raw Message
On Mon, 18 May 2015 00:56:33 -0700 (PDT), "Wo'O Ideafarm"
<***@gmail.com> wrote:

I owe you an apology: you did spell your hero's name right.
Post by Wo'O Ideafarm
Post by Paul S. Person
are you volunteering to do this?
Yes, yes, of course, Mr. Person. My perspective, motives, and interests differ from Jiri's and also perhaps from yours. I am new to open source and have much to learn here, but I hope that you and Jiri and the other contributors find that my involvement makes your involvement more fun, not less. I am not a manager. I have no interest in controlling anyone.
Then do it. Pick a bug from bugzilla or whatever Jiri uses and fix it.
And then the next. And then the next. No need to natter at us here.
Post by Wo'O Ideafarm
Think of me as a user who is extremely motivated to ensure that OW remains alive and well. I am an independent C++ software developer who is so passionate about C++ and so filled with fury about the course that the software craft has been forced to take by that autistic sociopath William Gates that I wrote an operating system that I hoped would destroy Microsoft's monopoly. That operating system is strategically committed to OW. If OW dies, my own life work dies with it.
So far as I can tell, there really is no immediate threat to OW's
health. Or your alternate OS.
--
"Nature must be explained in
her own terms through
the experience of our senses."
Wo'O Ideafarm
2015-05-19 00:49:21 UTC
Permalink
Raw Message
there really is no immediate threat to OW's health.
It appears to me that there is a very great threat. Mr. Malak is OW's greatest threat. He is also OW's greatest hope.

Critical mass and quality are essential. Critical mass is needed both in the number of contributors and the number of users. The competition is not standing still. All of Mr. Malak's work (and everyone else's) will be for naught if OW doesn't find its way back to being a viable product in its market.

There is much more to being a viable product than just having a handful of open source contributors who don't like each other's approaches, who are not on speaking terms, and who secretly or openly discuss the demise of each others' efforts.

That's where we are at this moment. That's where I see myself contributing. Fixing bugs is important. Adding 64-bit functionality is important. But the "soft stuff", the marketing, the administration, the interpersonal mediation, the leadership, those are all critically important as well. Approached properly, with people who are mature and likable, these "soft things" can make the code work more fun, especially once the product starts to take off.
Paul S. Person
2015-05-19 16:12:33 UTC
Permalink
Raw Message
On Mon, 18 May 2015 17:49:21 -0700 (PDT), "Wo'O Ideafarm"
Post by Wo'O Ideafarm
there really is no immediate threat to OW's health.
It appears to me that there is a very great threat. Mr. Malak is OW's greatest threat. He is also OW's greatest hope.
Critical mass and quality are essential. Critical mass is needed both in the number of contributors and the number of users. The competition is not standing still. All of Mr. Malak's work (and everyone else's) will be for naught if OW doesn't find its way back to being a viable product in its market.
You appear to be under the delusion that Jiri's version is part of OW.
It is not.

And a person who cannot be bothered to test properly is no threat at
all -- provided he keeps his hands off of OW.
Post by Wo'O Ideafarm
There is much more to being a viable product than just having a handful of open source contributors who don't like each other's approaches, who are not on speaking terms, and who secretly or openly discuss the demise of each others' efforts.
So far as I can recall, this group has never discussed the demise of
Jiri's efforts. He is free to do what he wishes and success is his to
obtain if he can find others to fix his little code messes for him.
We, however, have no interest in that activity.

Have Jiri and his followers been discussing our demise? I find that
unlikely; indeed, I find the very concept disheartening, as indicating
a very low level of maturity.
Post by Wo'O Ideafarm
That's where we are at this moment. That's where I see myself contributing. Fixing bugs is important. Adding 64-bit functionality is important. But the "soft stuff", the marketing, the administration, the interpersonal mediation, the leadership, those are all critically important as well. Approached properly, with people who are mature and likable, these "soft things" can make the code work more fun, especially once the product starts to take off.
OW is not a commercial product. There is no "market". There is no
income stream. The only forms needed are those needed to keep Perforce
(or whatever Jiri is using) happy. And "taking off" simply means more
downloads, not more work.

And it does not need a fuehrer. It is not out to conquer the world.

As with Jiri and wgml, it appears that you do not believe in
understanding something before trying to change it. This didn't work
for Jiri, and it isn't going to work for you.

If you want to help, get a shovel, and start cleaning up Jiri's code
messes. That will, at least, be useful.
--
"Nature must be explained in
her own terms through
the experience of our senses."
Wo'O Ideafarm
2015-05-20 15:27:20 UTC
Permalink
Raw Message
Mr. Person, thanks for your thoughtful reply. Although I respond here to only one of your points, thank you for voicing them all.
Post by Paul S. Person
OW is not a commercial product. There is no "market".
IMO, this is not the best way to look at things, but is a common perception within open source communities, which are typically comprised only of people who love to code.

One of the challenges that face every open source community is to remember that something is being produced so that it can be used by someone to obtain some benefit. Marketing people understand this because that's their focus. Engineers and technicians often lose sight of this because it's not their focus. People who like to code often really don't even care about the fact that they are producing something; they code because it is fun to code, much like pure mathematicians, who do mathematics for the pure joy of doing it, without any real thought about whether what they do will yield any practical benefit to anyone.

I am an accomplished mathematician. (When I was young, I extended the theory of differential calculus into the realm of discontinuous differentials.) I am also an accomplished coder. (I have written a "piggyback operating system" comprised of hundreds of thousands of lines of C++ code.) But I am also an economist. Yes, OW is a product, and OW must compete successfully in its market in order to survive.
Loading...