Because applications require money or money equivalents like time doing boring stuff. To get that money Linux needs marketshare.
That's true for a few applications, like office products and the like. Linux seems to have gotten those just fine without the unified desktop environment you keep going on about.
Firefox got up to about 5% share because AOL wanted a browser alternative. It was able to crack the IE monopoly because it was able to rapidly gain share to generate Google advertising revenue.
Netscape (and then later AOL) did host and direct early development on the Mozilla team, it's true. That lasted what, maybe two years? Most of that time was spent throwing out the codebase Netscape provided and starting from scratch. Firefox is self-funding now.
Firefox broke the IE monopoly because IE was a horrible, bug-ridden piece of crap. Website authors hated it because it never moved ahead with the standards. Sysadmins hated it because every week another worm targeted it. Users hated how their machines would fill up with spyware, and the people they had fix their machines would reccomend Firefox.
Open Office was interesting to Sun because it offered the potential to do something similar to the Microsoft Office monopoly. Its failure to gain share is why it is no longer funded and isn't advancing.
I'm sure the LibreOffice people would be surprised to learn that it's no longer advancing. Seems to me that the lack of advancement under the OpenOffice.org name was due to Oracle's mismanagement. Volunteers kept improving it, hence LibreOffice.
KDE back in the early 2000s was getting money from the German government. KDE's success in gaining share would have translated into permanent funding.
Permanent funding doesn't exist, unless you have a very strong lobby group and some money to throw at the politicians.
The success of LAMP is the reason that is a Linux as at all.
Um, no.
Ask developers why they work on Linux. I doubt more than 5% would even mention LAMP. They work on it because they want to.
There are some developers who work for companies like IBM who do, yes, primarily work on it because of LAMP. Those are the minority.
That's a slightly different definition of share. And no they don't work for everyone. Being a native application is really important to look and feel issues and integration issues.
Look and feel is important for impressing the suits. Everyone else just uses whatever works.
Particularly on a desktop where you want: cut and paste to work with complex objects all the way up to object linking and embedding.
Cut and paste got figured out a long time ago, as far as GNOME->KDE goes. OLE is a nice dream, but let's face it; 1% of people actually use it. Generally what gets linked is data, not an application, and data generally has tools to work with it on both the GNOME and KDE sides of the fence.
The point was in an alternative history where Gnome isn't as aggressive and I think these things work out just as well.
GNOME isn't aggressive. It's included in distributions by the distributors' choice. Obviously, the distributors had reasons for choosing GNOME over KDE.
For example perhaps KDE gets control of their code base so they can relicense the code. Or they assert standing and make an explicit exception for use with QT in the license.
Yeah, except the KDE team chose a non-free widget kit that they didn't have any control over. KDE couldn't get control over QT, because it was proprietary software owned by TrollTech. Yes, eventually an exemption was made, and nowdays QT is pretty much free, but it took a very long time for that to happen.
Personally I thought at the time and still do that people creating commercial software being required to support the Linux widget set is a good idea.
Which helps your idea of increasing marketshare... how?
You don't pay to write commercial MFC apps. You don't pay to write commercial Cocoa apps. You don't pay to write commercial Motif/CDE apps. You think people who write closed-source but free as in beer software would want to take the hit?
The GNU problem I think could have been solved by the KDE group if they had more time.
It was years before TrollTech made the GNU people happy. The GNU people aren't known for their willingness to compromise their principles.
Almost immediately the Gnome people essentially reinvented a bad C++ to write Gnome in because they needed more structure. I'd consider the language issues a design flaw in Gnome.
Well, there I couldn't really say. I've written GTK+ software, and found the object model in it easy enough to work with. I've never written a line of C++ code in my life, and have no interest in learning (I'm not a programmer, but picking up some C is a good idea as a UNIX admin).
You seem to suffer from a grossly disfigured picture of how the free software world works, and you still haven't given me any good arguments for why GNOME shouldn't exist. UNIX is not, and never has been, a monoculture for development. You can't force developers who are volunteering their time and work. It's up to the distributions to make a coherent system, and most of them chose GNOME.
You are still free to use KDE. Nobody's stopping you. People will still continue to work on KDE. You're losing nothing. If people using GNOME bothers you that much, then perhaps free software just isn't for you.