discuss-gnustep
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Thoughts triggered by these NeXTbuntu guys


From: Jeremy Tregunna
Subject: Re: Thoughts triggered by these NeXTbuntu guys
Date: Wed, 30 Aug 2006 02:31:01 -0400

Just before I go here, let me just start by saying that I don't represent the GNUstep project, nor am I an active contributor. So feel free to skip the rest of this message and label it as noise.

On 06-08-30, at 00:12, Tima Vaisburd wrote:

While many people here seem to think in terms of porting, I believe that new development bears greater potential. I believe there are enough people who have their own powerful ideas that would be happy to implement them with
our superior tools - if we provide ones.

The problem all comes in when you look at people who are spending time working on GNUstep, and those who think to themselves "it's not ready" and go on to something else. It's all these people passing GNUstep up for those reasons that GNUstep isn't further ahead today. Note that people who just don't have time, well, just don't have time so they're excluded from my description above.

However, we do not have such tools to show, because the core libraries -
I mean four of the packages taken together - are still incomplete.

Not all of it, especially when you look at what GNUstep's goals are: OpenStep compatibility. gnustep-base is pretty complete if not 100% complete, plus many additions.

Since I'm writing some image editor I happen to know that writing of
images in any format other than tiff is not implemented (and most people need jpegs or pngs instead). We do not have a single backend that works
perfectly. The arts backend is prohibitely slow on remote connection.
Xlib does not support rotation and poorly supports transparency. There are
cairo backend in development, but it's not released yet.

I often connect to my home machine from work, and use X windows over ssh and then cable internet provider. Since GNUstep is extremely slow on that
connection even with xlib it exludes any GNUstep program as a viable
alternative to other X clients (but emacs, for instance, works reasonably
fast).

These are things I can claim from my personal experience, but from reading this mailing list I got the impression that there are other unfinished areas.

You ever thought about doing anything to help alleviate those problems? I mean, it doesn't take a lot of effort to sign the copyright assignment and while you're waiting for the papers, you can easily work on your code to make it go. That said, I don't know what it would take to fix these issues.

Yet Gregory Casamento writes:
  "The libraries are complete in every sence that they need to be."

By what standards?

It seems to me that the project lost its leadership and its focus.
Good leaders can set the goals and show path from one goal to another.
I think we need that.

Instead, it seems, we start things like facelifting, windows port, desktop
environments without completing the foundation, beginning the new
project without finishing the old. That way there is nothing solid to show
to the mankind at any point in time - for years.

People get bored, and when you've only got about a dozen folks who are really committed to working on GNUstep, things can stagnate while real life gets in the way, and well, when they're as I say: bored. As I mentioned above, if more people would stop passing it by out of sheer "too much effort", then we'd be a lot further along.

Then if we get a person who is able to contribute there is another thing
that I consider the second most important problem - the copyright
assignment to FSF.

This topic deserves another message with a different subject line,
here I can only say that it presents a natural obstacle for people who,
like me, wants their work to be recognized by others.
Thus the Nextbuntu's idea on removing this requirement seems to me
quite appealing.

This I'm not so sure about. For a large project, requiring code to be copyright assigned to a neutral third party (or if run by committee, that group) can save you from cases where someone might get all pissy and decide they want to revoke distribution privileges for all their code they've contributed, setting the project back potentially many months. If you don't want to assign copyright, you don't have to -- you can maintain a separate patchset outside of GNUstep. But please, don't confuse assigning copyright to the FSF as not getting credit. All you have to do is look through the GNUstep source code, and you'll see "written by..." even if the module has been rewritten, you'll usually see a "original implementation by ..."; credit is given.

As I mention before, copyright assignment, while a bit of a pain, is a benefit to the project more-so than a hindrance.

To summarize: before going into new desktop environments, please finish the core gui/back libraries to let people learn and use them for their own purposes. Declare specific API version - being it Mac OSX 10.2.7 or not -
and make it work completely under at least one platform.

Different people have different desires; why should those people who want to work on a desktop environment work on -gui/-back when they may not want to?

Thank you and sorry for being so long.

No worries. =]

--Tima

--
Jeremy Tregunna
address@hidden






reply via email to

[Prev in Thread] Current Thread [Next in Thread]