Congratulations to Ali for trying to fork GNOME. Fair play to him for exercising his rights under the GPL. Now at least (after all these years) he has his own project space in which to rant on about his issues, rather than finding it necessary to keep winding people up on the GNOME mailing lists.
Seriously though, I do respect this attempt. In a way, he is finally attempting to do something productive about his gripes himself rather than persistently trying to persuade other people that what they think is wrong. Either way, the end result is that he is exercising his right to take what he wants out of the software and make whatever he wants out of it, use and develop it for his own purposes and make it available for others to do the same if they wish. If he feels this will be a productive route for him, then he's welcome to follow it.
However, as usual, the his post is unnecessarily bloated, is full of inconsistencies and goes off on weird tangents in places. The introductory paragraphs seem fairly straight and to-the-point, but when elaborating on the points in his bullet point list, he seems unable to explain himself clearly. I tried to objectively review some of the points he tries to make. Actually, some of his point are valid, even if they're not exactly breaking news.
I though that his complaint about UI consistently was fair enough, but the HIG was proposed long ago as the solution endorsed by the main developers. At the end of the day, we cannot force everyone who develops a GNOME application to agree on the HIG or develop or accept patches to conform to it, so it will take time. The core desktop apps should at least conform. There will probably be a few more revisions of the HIG out before we get anywhere close to that goal. However, unless anyone has any better ideas, progress is being made so we should move on. Again, he seems to be unhappy with progress and is harping on about old problems that already have a solution in progress.
And yes, Nautilus going spatial did seem like a step in the wrong direction. Nothing much I can say about that except - it's real shame. I remember seeing Nautilus demoed at GUADEC 2000 and thought - that's going to rock. Despite the *huge* amount of work that's gone into it, sadly I still find it very annoying to use. A moment ago, I tried using it to browse to a windows share on my network, and it insisted on opening four new windows to get there. Before, I could change this to 'open in same window', but now it seems I cannot. That's not progress in my opinion. However, at the end of the day, despite not working exactly the way I would like it to, it does work and allows me to do what I wanted to do. Nautilus could (and did used to) do better.
He also correctly mentions that some code forks can result in useful merges later (such as with gcc and XFree86). However, the people responsible for such useful code forks were highly skilled engineers, with a deep understanding of software engineering and of their goals, and such achievements were accomplished through a lot of hard coding work, and by presenting demonstrable, concrete results. In other words, these useful forks weren't accomplished by simply posting lists of personal gripes and a handful of patches in public places and hoping that others will join in.
I found it difficult to take the rest of Ali's post too seriously. For example, there's no way anyone would prefer Latex to Docbook for technical documentation. GConf is like the Windows registry because modern desktops need such a configuration interface, and (IMHO) GConf is an excellent implementation of this. His issues of bloat are matters of perspective and opinion (one man's feature is another man's bloat), and inefficient code can always be optimised, given enough time and effort. He says that 'GNOME must touch everything' and then goes off on one without a concrete example - I haven't got a clue what he's trying to say here at all! Not content with slagging off the GNOME and KDE developers and architecture, he also takes an unqualified swipe at the freedesktop.org standards, and saying they're a waste of time. Someone please pass me the heavy duty clue stick.
So, good luck to him, but I can't see an army of developers lining up to help him produce 'Ali's GNOME' any time soon, nor can I see many people wanting to use it if he succeeds.