Re: Plans for change.... (minor correct of previous post)

[Henrik Mikael Kristensen]

On 17/12/2006, at 23:10, Philippe C.D. Robert wrote:

> On 16.12.2006, at 23:24, Gregory John Casamento wrote:
> > > I now ask these questions: What was the original goal of NeXT with
> > > their OS? Should that goal not also be the same for GNUstep?
> >
> > The original goal of OPENSTEP was to create a crossplatform set of
> > libraries which could be easily used.  These platforms consisted of
> >  Windows (OPENSTEP Enterprise 4.2/Windows), Solaris (OPENSTEP 1.1/ 
> > Solaris), &
> > Mach (OPENSTEP 4.2/Mach) were the implementations of this created by
> > Sun and NeXT while NeXT was still in business.   There were  
> > proposals to
> > have an OPENSTEP implemented under HP-UX on the PA-RISC architecture,
> > but that didn't happen prior to the buyout by Apple.   On each one  
> > of these
> > platforms, mainly windows, OPENSTEP was made to look/act like the  
> > operating
> > system it was on.   As you can see, GNUstep's purpose is  
> > *precisely* the same
> > as OPENSTEP's.
>
> And NeXT clearly failed with this strategy ;-)

Ah, the answer I expected from the question was the original goal set  
by NeXT for their first products.

The original goal was to make a high quality workstation usable by  
science institutions, universities, students and also in enterprises,  
I believe.
What they did was price the machinery out of the range for precisely  
those customers to buy them.
Had they not done that back in 88, they would probably have gained  
more foothold than they did. I don't think OPENSTEP particularly  
failed, but Windows was already gaining too much marketshare for  
OPENSTEP based desktops to catch up and the technology that it had,  
didn't matter that much anymore to the customers.

> I guess the difficulty here is that there are some who understand  
> GNUstep as something like OPENSTEP Mach 4.x, an entire OS or at  
> least desktop environment running on a Unix/Linux OS, whereas there  
> are others who understand GNUstep as an implementation of the  
> OpenStep API specification (with some - but not all - Cocoa  
> additions/changes) which integrates seamlessly into its host  
> system. In this case a Windows port is what probably matters most  
> (business wise). Right now GNUstep is a mix of both which makes  
> nobody completely happy.

Yes, this is also the way I see it. I'm just looking at it from one  
side where Gregory wants it to be on the other side. I guess there is  
a big difference in goals with the early NeXTstep OS, which is what I  
would go for and the OPENSTEP goals which is what Gregory wants.

The reason for my opinion is, I'm looking at the current state of  
Linux desktops and where MacOSX seems to be going. Being a Mac user  
on a daily basis, Windows user second, Linux user third and Amiga  
user fourth, it looks to me that MacOSX is going to be the big winner  
in terms of richness and power of its technology and the leverage  
that developers have to create new kinds of applications very  
quickly. MacOSX certainly doesn't need any of the others to do well.

MacOSX works on its own, most likely due to smart and disciplined  
people, many of those working at Apple. A crap load of money also  
helps a bit, of course. :-)

Linux desktops on the other hand are a mess. They are moving along,  
but very slowly and Gnome still doesn't appear to be polished with  
many features as crippled and defective as they were when they were  
introduced 4 and a half years ago in Gnome 2.0.

In the case of KDE, it has a much better foundation, but severe  
usability issues and "featuritis", IMHO.

Those two are the Linux desktop that have the greatest momentum, and  
they are both doing very badly at conquering market shares, because  
MacOSX is going to steal the show as the true alternative to Windows.  
Think about that.

I would just like to see GNUstep technology available in its own and  
purest form to create a very different desktop without compromise on  
standard PCs.

Gregory mentions somewhere else in this thread that such stand alone  
projects are never successful, such as AROS and Haiku OS. I think  
this is because developers are not catered for properly as is the  
case with GNUstep. Without developers, these projects are quietly  
strangled to death. They will never grow popular beyond a very small  
group of fans.

There is also the "nice toy" effect to overcome (boy, that looks  
really nice, but can we go back to Windows again?), something that is  
very common in small projects that deal with new/different  
technologies. I struggle with that every day when using REBOL for my  
development activities. Who in their right mind would use AROS in a  
serious company environment for critical tasks?
None of these projects have a "serious flavor" to them, while GNUstep  
has, being already in use in companies for serious custom applications.

I think it's possible to create a true stand alone desktop and  
development system with GNUstep in a way that none of the others  
really can.

--
Regards,
Henrik Mikael Kristensen