I think that Trainz does have a future. This is not just a program, but a way of life for many people here, thus the concern that many people have for the viability of the organization. My company went through the same kind of downsizing - from 150 people to 12, and we're actually doing better and even posted a profit for the first time.
Anyway, the program its self can run stand alone, and really doesn't need an Internet connection other than for downloading content. So if the DLS were to poof out, the program will continue on its way, and as others have said here, the many user groups and sites will pick-up content and make it available.
If anything, I would love to see someone else host the DLS and keep the content updated. This would give Auran the extra freedom to focus on program development instead of devoting time to uploads and content repairs. The added benefit of hosting the content else where too would be to have a chance to revamp the whole DLS interface so that packs could be handled differently, and the content could be presented and organzied in a better fashion.
In regards to CMP, I've had very few problems that weren't solvable. First of all I found that by running CMP directly instead of through the launcher, I get better performance. I actually have a seperate shortcut on my desktop for CMP alone. I found that the launcher added to the memory overhead that CMP needs to run.
Like Ed and the others here, I have also spent some time removing or reparing faulty content. In many cases the errors were just typographical errors. In other cases, well the content went to the bit-bucket because it was missing components that went beyond what I could handle. After doing this, the whole Trainz 2006 suite has work like a champ for me. I've spent upwards to over 9 hours in Surveryor and Driver and exited without a glitch back to the destop when done.
Anyway, this is my 2-cents on this.
John