I have used GoboLinux for years, in dual-boot and virtualized environments alongside XP and Ubuntu/Mint. I have also used at some point SuSE, RedHat, Gentoo, Ubuntu and recently Linux Mint (itself an Ubuntu derivative). I truly value GoboLinux's vision and it's impressive the amount of work done and progress made by such a comparatively small community.
However, I have a question, which I understand might offend some purists (Goboites? Goboists? Gobozealots? Gobophiles?). Given that Ubuntu and its derivatives now enjoy possibly the largest installation base among desktop Linux distributions, and have a correspondingly large development community and software base, don't you think it might be easier to create an Ubuntu-derived distribution or a modification to Ubuntu to create a GoboLinux-like filesystem hierarchy, than to work on a separate distro, even one which is growing as GoboLinux surely is. Or, put another way, does there exist a 'market' of people who love the GoboLinux concept, but would feel more at home in a very polished, user-friendly distro such as Ubuntu?
I don't suggest that GoboLinux is extraneous, merely that it may grow stronger and find more converts working with this concept than by building GoboLinux separately.
I am not exactly sure what this leads to.
I believe the problem is not so much that there are not enough users on Gobolinux (there can always be more users of course) but more that there is not enough developers to make things work "nicely". I somehow believe that most users tend to stick to what is usable for them, i.e. the more tech savvy users will stick to Gentoo or Archlinux (GNU Sourcemage has a somewhat similar "problem" as Gobolinux insofar that new users come in only slowly) and the not-so-tech savvy will use Ubuntu, PC LinuxOS etc...
On the other side we have the huge behemoth Ubuntu. I dont think it is easily feasible to use an out-of-the-book Ubuntu system and make it "Gobolinux".
Just imagine how much work you need to apply - you basically need to start from 0, because the FHS on Ubuntu also means that all their .deb packages are spread over.
Now, ok, you can uninstall all, and then use the .deb package to install something into standalone dirs. That would be a start.
I am not sure anyone has done so yet? This ability would be a win-win for both Ubuntu and Gobolinux.
Also I think the majority of both Ubuntu devs and users does not like the Gobolinux idea. You should look at the brainstorm suggestions of Ubuntu, they were in majority turned down by the Ubuntu guys (think there were 3 suggestions that were very similar to the Gobolinux ideas).
Aside from the .deb issue, the only issue I could think of where Ubuntu could have a "stronger" hand would be to update large or complicated packages like Xorg, simply because they have more developers. (I.e. modular xorg, glibc, gcc tend to lag a bit behind on Gobolinux compared to other systems, i.e. Archlinux, or even Ubuntu)
Personally I think the biggest problems in Gobolinux are recipes & packages right now for various reasons, biggest one being is that they become outdated very quickly, or may have problems - or both. I think this is one thing Gobolinux should improve.
From my perspective (as a user) it is much more rewarding and interesting to help improve Gobolinux than use or help improve Ubuntu - while Ubuntu may be more polished and has a lot more users and devs, the idea in Gobolinux is simply better.
I'm not offended by the idea of merging Ubuntu with GoboLinux and I don't think any of the other devs are either. I think there are two main reasons this hasn't been done yet.
1) Why don't we support .deb on GoboLinux
Because we don't feel this is necessary from our point of view. Users might want it because of the large .deb repositories, but remember that we, the goboLinux devs are users of the GoboLinux system as well. There are numerous other ways to improve the aspect of repository size for GoboLinux. Simplifying contribution of recipes is one - that is one of the new features in the comming Compile release. Creating a build farm to create packages from the recipes in the repository is another - which we work on currently.
Officially supporting .debs on GoboLinux introduce other problems, specifically dependencies resolution. And that's not something I'd like to spend time on. It's not that hard to manually install a .deb, just unpack it into /Programs/Foo/x.y and then run SymlinkProgram Foo x.y. It works for most cases, if the dependencies are installed, due to the legacy symlinks.
2) Why don't Ubuntu use GoboLinux layout
Ask the Ubuntu devs. As shevegen pointed out the idea has been brought up several times in Ubuntu's brainstorm site, but it hasn't been recieved that well.
A third alternative would be to create a branch of Ubuntu/GoboLinux. But I guess that's up to those that are interested of such branch. Of course we (at least I) would help with GoboLinux specific support of the tools if such branch would emerge.