"Michael Phipps" <mphipps1@xxxxxxxxxxxxxxxx> wrote: > >Hi all :) > > > >I'd like to ask you developers about the status of OpenBeOS > >development. > > > >1) I see a lot of cvs commits for the kernel, and for the ppc > >platform too: does it mean that the NewOs kernel is ready for ppc > >platforms? > > Axel got a free PPC box in exchange for porting the kernel. > There is no promise to the general public that it will work or what > machines it will work on. :-) > I certainly will not promise to support old PPC hardware or any new > Mac hardware. ;-/ Apple is just unfriendly. > > >2) In your opinion, will it be possible to have OpenBeOS R1 ready > >for late 2004? > > I don't think that I could take another year of "when will it be > done" questions. ;-) > > >3) About multiuser support, for R2: there are a couple of types, > >the multiuser "a la Win98" (i.e.: different settings only for > >every user) and the secure one "the Linux way" (real multiuser > >mode) :) > >I would even like to see the first one, but if we/you can get the > >second it would be much better :)) > > I would argue that way 1 is no more or less secure than way 2. > In fact, way 1 is far *easier* to secure than way 2. > Think about it. For both you need real file system permissions (which > we have). > For both, you need a secure log in mechanism (which we don't have). > For way 2, you need to be absolutely sure that no process can > possible access the memory > of any other process. The problem is that the BeOS design precludes > that (see clone_area). > > The question that I always ask people who ask about multi-user is > this: > Do you really forsee wanting to have multiple people log in to your > OBOS machine simultaneously? Strong yes. I can't even provide a CVS repository on my machine, if simultaneous ssh connections don't work. I've been hoping for this for years... > If so, what would they use for a client? OBOS? If so, then what you > want is really file sharing. > Remember that our focus is a *workstation* OS. :-) > > >How many months of work would this require to the actual Kernel > >team? > > Way 1? Nearly none. > Way 2? Who knows? We would have to rethink parts of the API. It would > really all depend on what we want way2 to do. It would probably require massive work on the part of the media kit and app_server groups. Life is hard. Sometimes satisfying solutions require much work. CU, Ingo