Re: ANN: One Step to GNUstep - pre-release version 0.9

[Gregory Casamento]

RIccardo,

On Mon, Jun 13, 2011 at 5:57 PM, Riccardo Mottola
<riccardo.mottola@libero.it> wrote:
> Hi,
>
> thanks again for turning the announcement of a postive effort for GNUstep
> into a personal philosophic crusade where we always restate the same things.

I wasn't the one who did it.  This thread was well underway in this
direction prior to my response.  So, don't blame me.  It was, however,
important to remind people not of any philosophical debate, but of the
original purpose of this project and that the purpose of it HAS NOT
CHANGED under my watch and will not change under my watch for the
forseeable future.

I was seeking to make it clear once and for all what this project
stands for and what our goals are and they are not to be a Cocoa-like
framework but to be a Cocoa compatible one for the reasons I've
already stated earlier.  The attitude that GNUstep should be just
"Cocoa-like" forms a major impediment to the project and costs us both
time and resources.

> Do you really think this helps GNUstep?

As far as discussions like this go... I think it helps to get things straight.

> Our image?

Our image is the precise thing I'm trying to repair and correct.   We
are perceived very badly by the rest of the open source community
since we have this lack of focus.

> And overall, the morale of the current developers?

I'm not certain that morale is best served by putting another project
down due to the decision to use an obviously superior compiler as was
done earlier in this thread.  I'm also not sure if morale is best
served by continuing to avoid implementing or using "small features"
as you would put it because of the reservations of a few people.   As
lead I talk to groups and companies on a regular basis about helping
GNUstep and I hear again and again the same things over and over again
that GNUstep isn't complete enough, that it looks ugly, that it
doesn't support this feature or that feature.

Morale would improve immensely if there was more participation in the
project from more companies and more individuals all collaborating at
making this project better and that is what I'm trying to achieve by
travelling to California, Pennsylvania, New York and other places to
give talks about GNUstep and try to bring more developers into our
community.   I've also been trying to build community by encouraging
corporate participation and trying to convince other projects to use
GNUstep instead of alternatives such as libFoundation (even though
libFoundation has some of the same roots as GNUstep, things
contributed to it are not seen by us) or Cocotron.

So, please, do not speak to me about improving morale when I have
spent so much of my time and my own money trying to make it better for
this project.   I should mention here that these are things I don't
mind doing whatsoever...  I do them because I want to bring more
contributors to GNUstep.

>> This also emphasizes one other thing I've been saying repeatedly and
>> that is that small differences (i.e. lack of certain features)
>> presents a huge issue when it comes to getting new users and
>> developers. The features added to ObjC 2.0 in LLVM/clang are not
>> "small features", since lack of these features represent a serious
>> impediment to new developers who would like to either write new
>> applications or port existing applications from Mac OS X.
>
> Those features are not essential: you can program without them. They do not
> represent a serious impediment for new applications. They represent a hurdle
> only for the Mac crowd coming with their expectations.

This is not relavant.   Saying "you can program without them" doesn't
mean that someone who's seeking to bring an application over will
change their app to work with what we have.   We need to provide what
most people need out of the box and we also need to make it easier for
people to become contributors to this project.

> And to get back on track where Nicola left: sure, a lot of mac people
> *could* come over, but few of them will develop into free and open source
> enthusiasts. Most of them will look at GNUstep in a way to expand their user
> base.
>
> Granted, we are free, so I am happy if a Mac developer wants to use my
> stuff! Even for commercial stuff! That's the power of free software. However
> these are not the only user and developers who we can twinkle to.

Perhaps not, but as we saw with TestPlant, they DID become a
contributor to GNUstep and they continue to be one.  So, your
supposition is not borne out by actual fact.

>> Losing this potential new blood is something that cannot be allowed to
>> continue on this project.
>
> Wow! Right! Exact! But we do not only get visits from Mac developers but
> also from people genuinely interested in an alternative framework and/or
> environment. Several dropped into the IRC channel and several into the
> Mailing list.
>
> Several were turned away by the access to GNUstep, those who were able to
> run it, by the incompleteness and lack of maintenance of our applications.
> But that is new blood too. Do you think it is not "blue" as the royal blood
> from Apple? Maybe not, but perhaps it is more FOSS oriented!

I find the use of the word "several" somewhat amusing here, since I
can think of not several but many hundreds who have been turned away
by our incompleteness and an implementation of Cocoa.   I do agree,
however, the improving the stability of our existing applications is
very important, but this is also something that is best served by
having more contributors.   Having more people contribute to the core
of GNUstep would confer the kind of stability you're thinking of since
there would be more eyes to find the actual issues.   Lest you forget
that when porting TestPlant's application (Eggplant) there were a
large number of bugs which were addressed which made GNUstep more
stable and more compatible with Cocoa.   If we had several other
contributors such as this, it would be nice.

> The packages done by Richard for SuSE should help anybody in their first
> steps and the VM setup attract even lazier people :) It is also a show off
> how things can work (in the good or the bad) in a reference setup.
> Just through the making on it Richard helped pinpoint a quite large number
> of bugs, from the make systems to libraries and apps! This is precious.

Yes, it is.

> Wonderful rewarding GNUstep community.
>
> I'll just say thank you again to Richard.

I did thank him.  What he did was great.  Where this thread went was
not, but it did expose some of the major cracks in this community.

> Riccardo

GC
-- 
Gregory Casamento - GNUstep Lead/Principal Consultant, OLC, Inc.
yahoo/skype: greg_casamento, aol: gjcasa
(240)274-9630 (Cell)