Index: [Article Count Order] [Thread]

Date:  Sun, 28 Oct 2007 09:00:20 +0200
From:  Michael Stauber <bq (at mark) solarspeed.net>
Subject:  [coba-e:10991] Re: AW:  Re: Progress Development Bluequartz
To:  coba-e (at mark) bluequartz.org
Message-Id:  <200710280800.21549.bq (at mark) solarspeed.net>
In-Reply-To:  <40E1C71DFD600D4EA127F2840C6C1E9B028CCE (at mark) www.palei.com>
References:  <016201c818ca$f1e1b2e0$7002a8c0 (at mark) dell> <200710280044.07570.bq (at mark) solarspeed.net> <40E1C71DFD600D4EA127F2840C6C1E9B028CCE (at mark) www.palei.com>
X-Mail-Count: 10991

Hi Paul,

> Michael, I am not saying you are bad, heck no.
>
> Take any of the projects that have started as Open Source and spin out to
> commercial and you have seen a great deal of the community leaving. 

I have no intention of making any part of BlueQuartz commercial. To the 
contrary. There is plenty of code that I intend to contribute to the project 
under the GPL or whatever similar license the project choses to use. 
BlueQuartz itself has to remain open source as that has been what made it 
what it is today.

There are - and no matter how advanced BlueQuartz gets - always will be 
*special* needs of clients and users that are not (yet) covered by BlueQuartz 
itself. Which leaves enough room for commercial extensions or third party 
software. Or there are those that need professional help with BlueQuartz 
beyond the help that the user base or the mailing list provides. Those 
special needs will eventually get filled by one third party or another, but 
nobody will be forced to take that route. At the same time: Don't forget 
there are also companies that will not consider using software (open source 
or otherwise) if there is nobody around that can and will provide commercial 
help *if* and *when* they need it.

> With BQ that never happened, so we now have different people doing different
> things. Once again, I am not putting the finger nor blaming anyone. 

The problem with BlueQuartz is not that it lacks commercial background. There 
are enough very sucessful open source projects without large companies behind 
them. But the main problem that BlueQuartz has is indeed that there are too 
many people doing different things, because there is a lack of general 
organization and direction. 

Why is that the case? Well, as I expressed in private on the developer list: 

A TEAM of developers usually speaks among eachothers to determine a roadmap, 
to find out what needs to be done and splits up the work so that those most 
suiteable for a certain task (and who are willing to code and contribute 
things) work on certain aspects of it. If someone gets stuck or with his task 
or runs out of time to make contributions, others can look it over and pick 
things up. Once something is finished, the code is then reviewed and if 
deemed acceptable and tests out just fine it'll be commited to the official 
branch. Final decision lies of course with the chosen project leader and 
that's usually also the one that should then roll up the code and publish the 
updates for general usage.

This is how it ideally should be. BlueQuartz however is and always was far 
from that. And THAT has always been hurting.

Previous efforts to make BlueQuartz really a group effort with a team of 
contributors have been stonewalled and people that offered to help in one 
form or another have been turned away by simply sitting things out and by not 
communicating. This happened regardless if it was a push that (amongst 
others) included people with "commercial interest" (as you put it), or if it 
was the odd user here and there that had found a flaw, or wished to 
contribute a translation of the GUI that he had done for his own native 
language. That's quite a few missed opportunities as well.

How many people were involved in the porting to CentOS5 (5200R)? It was a one 
man show and that development branch hasn't seen any updates in four months. 
Like I said in private: That's not unexpected as it is a serious load of work 
and more than one shoulder *can* handle. After spending roughly 2000 man 
hours on a single commercial project of similar nature I can relate to that. 
But it's not that nobody else wanted to help. It's because nobody else was 
included or even given the chance to help. That's another missed opportunity 
and failure in communication.

But it's goot that this things are now talked about, so that a solution can be 
found which serves the best interests of the project, serves the user base 
and which helps to get things moving forward.

> Besides the obvious differences, are you and others going to include the
> roadmap for your own products into BQ? If not, then there is nothing wrong
> with that, but then lets called it a fork lets all be happy. 

I can't speak for others, but personally I have no problem if something gets 
included into BlueQuartz which so far used to be covered by one of my 
commercial offerings. Like I said earlier: Regardless how sophisticated 
BlueQuartz gets, there will always be special needs of people that are not 
yet catered.

As of right now a lot of my clients are running a pretty customized 
BlueQuartz. It is still based on the NuOnce v4.7 build, but about half the 
BlueQuartz related RPMs in that build have seen extensive modifications and 
improvements. Including the fact that Shadow authentication is now used by 
default and instead of PAM. But it also supports the old PAM based 
authentication, so people have the choice as of which authentication method 
they want to use.

I've always been interested in contributing these changes (and others) back to 
the BlueQuartz project, but due to lack of access to the repository I've been 
unable to do that. 

Sooner or later the point will be reached where it may make little sense to 
still tie that modified BlueQuartz into the official BlueQuartz repository if 
there are no signs of progress on the horizon. Once that point is reached it 
may as well be time to call the modified BlueQuartz no longer a "modified 
BlueQuartz" but make it an outright fork instead. If that happens it'll still 
remain open source and will be available free of charge. Not only as an OS 
template that runs on a commercial virtualization product, but also as stand 
alone install available as an ISO image that is downloadable and useable free 
of charge.

All things considered I don't like the prospect of doing a fork as it only 
leads to splintering whereas joint efforts are always a lot more effective. 

My personal roadmap is focussed on porting Sausalito to CentOS 5 and to 
implement features into BlueQuartz which I find lacking. I'd like to 
contribute those to BlueQuartz as well. I am here and am offering my support 
to the team and as part of the team. IF such a team is formed and it's no 
longer the one man show that it currently is. If that kind of help is not 
welcome, then I don't have a problem with that. Well, get things organized 
and if you want me aboard you know where to find me. 

> Don't make more of my comments because there is nothing else behind.

Based on your last one liner on the developer list I may eventually start to 
think otherwise, Paul. It's fine that you question my motives, but don't 
maneuver yourself into a position where the same can apply to you as well.

-- 
With best regards,

Michael Stauber
http://www.solarspeed.net