"Erik Jakowatz" <erik@xxxxxxxxxxxxxx>: >>I have a question to all members on the list - should we be striving >for >>binary compatibility (1) or for source compatibility (2). Since a >rewrite >>also means improving (or in the worst case cripling) the servers we >We must accomplish both. I'm probably starting to sound like a broken >record to some folks (eh, Marcus? ;), but if we sacrifice either at this >point, we run the very real risk of losing our userbase -- which would I wanted to listen only on this list, but this repeated argument forces me to say some clear words. Rewriting the OS will take a LONG time. If you're doing it in 2 years, it was FAST. And then you're just up to par. Whereas "par" is already 3 years late. That's what BeOS currently is: 3 years behing schedule. So beeing 5 years behind schedule and waiting 2 years for that will loose you userbase. A userbase that is already beyond recognition. Also, all those applications that benefit from binary compatibility will be 2 more years old - or they are still maintained and can therefor be easily recompiled. And don't forget that BeOS _always_ broke compatibility until R4. Dont' forget that the fragile base class problem will always be a time bomb for binary compatibility anyways. First thing when you're up to par will be a new networking, a new Media Kit. Which will break compatibility anyways. Also missing is a state of the art Interface Kit. Etc, pp. 60 people on this list, 12 who are interested enough to make public what they could to. This makes maybe 6 who stay active. That's just enough to debug and improve Tracker and Deskbar. Where have all the developers been all these years who should rewrite BeOS now? AtheOS is 5 years old. BeOS is 10 years old. Think about it. P.S: I would vote for improving on the R5 base, putting rewriting of old modules on second priority. This will give you binary compatibility for a long time and improvements which are needed to keep and increase the userbase.