I like the bit about the spaceman. :D
Originally Posted by airlied
But ya, essentially my plan to desktop Linux profitable is essentially to decentralize its maintence. It is much easier and productive to maintain a certain subset of code, then to maintain everything.
The semi-rolling release system was born out of neccessity. I needed to make a distribution that could keep up with the times in terms of video codecs and Bittorrent clients, while maintaining a stable core OS that was usable and safe for my users. Additionally, I needed a way for me to maintain the thing on my own without spending countless hours rebuilding packages.
I guess the best things in life simply come from pure need.
That does sound like a good idea. Since even if there was to be a roadmap (possibly with parallel moving since not everything depends on everything always), the current position(s) must be able to be tracked both on distro level and on upstream level. Then people can see that hey, I now have this and this is what I can expect when things get pulled from upstream. (a visual roadmap might actually not be a good representation, maybe possibly a categorized list of components like OpenGL 2.1, what it consists of, which parts are done, OpenGL 3.0, what it consists of, which parts are done, which parts are shared and which are not etc would imo be somewhat more informative and less of a hassle than trying to think which "road" we atm walk)
Originally Posted by bridgman
Keep us posted on how it works out in the next year or so. I'm a bit sceptic by nature but always interesting to see how new ideas work out. Start keeping a blog on your experiences with it, for example?
Originally Posted by darkphoenix22
First of all, i still can see no logic in your previous answers. Why would ATI need so much more time to do the same for fglrx as for radeon oss - when you look at xv. As the chips changed for nvidia with series 8 they had to implement xv differently and ATI had to do the same with R600 - this happened more or less the same time +/- 1 year or so. Why whould have got Nv more experience? Also interestingly the xvba lib appeared 1 month BEFORE vdpau and it is still not fully debugged. I think you reverse the history - thats just a lack of real interest not that you would have got less time to adopt changes.
first you need the produkt then you can sell it!
Originally Posted by bridgman
amd first need invest in linux after that they can ern money.
not first ern money and then invest in the moneyerning.
the situation is very bad 2 exampels:
if you wana sell hardware with installed fglrx the users watch the desktop and tell you something like composit isn't working very well and flash no not run well and at startup the system flickering.
and the viedeo acceleration isnt working well and you can't watch 'bluerays' (ok ok blueray isn't amd fail but they support that shit!) wine for exampel do an very bad job on fglrx...
ok other exampel:
you wana sell hardware with installed radeon the users watch the desktop all is fine flash isn't acceleratet but works fine viedeos do only have basic Xv but works fine and then aftr the good feeling you checkt out you can't play HON because openGL2.1 is needet and only openGL2 is stable and mesa7.9 isn't ready yet and wine dosn't run very well because wine need openGL3.2!!!! (no openGL2 is a joke for wine because fglrx do have openGL4 and works bad to) wine support stops on very basic thinks...
o yes FireGL catia works so fast.. LOL
first you get a flash because of the bad workspace feeling and you have to throw up blood because of the FGLRX but your Catia(3d-cat) works fast on FireGL! LOL!
radedon is more near by a good driver than fglrx because radeon only need openGL2.1 and then you get 'fun' (not only games).
the fglrx need a complete 'brainwashing' yes my be 'direct2D' but i think i have to throw up blood again on fglrx-'direct2D'.........