>Subject: Re: Developers - which direction? >From: "Marcus Overhagen" <marcus@xxxxxxxxxxxx> >On Fri, 17 Aug 2001 10:57:27 PM, "Cedric Degea" <cdegea@xxxxxxxxxxxxxxxx> wrote: > >>* maybe a BeOS rewrite "from scratch".. was very reluctant to the >> very idea proposed by a friend a few montsh ago, but now >> I think I'll offer my help to him. It's been extremly late >> already for me, but it seemed indecent to me to envision >> a BeOS repalcement as long as the company was still "alive > >This is an option, too. But I think this would be the minimum that we would need to have: > >* The OS must be open source The one project I was thinking about http://blueos.free.fr isn't. Ouch, I can hear everyone's scream from here :-) But the reasoning is, free software is the way it is (with the only exception of Apache and Linux which you may consider success stories for their wide acceptance if not for their technical excellence) because one has no interest to invest his soul and heart in doing a kick-ass design for a work that will be basically given for anyone to rip/dissect.. BlueOS is a kind of odd concept in this day and age where "FreeWare" as defined in e.g. the amiga days seems to belong to history: it's a project where the source is "proprietary" (the one that will be added on top, not the modified GPL one of course) even though the "product" will probably always remain free-as-in-beer. As it was one of the first questions I asked Guillaume when I first heard about this, he pre-emptively replied to this concern in the FAQ, something like ~"we're a differnet kind of mothership [from Be] as we'lucky enough to be free from any obligation of making benefits, so even though you don't get the source, BlueOS's future is safe as long as there is user interest in it"... Thus maybe I'm already off topic for this list too, then, since it's dedicated to discussing OSS rewrites of BeOS only? >* Features like indexes file system, live queries, MIME based file typeing, UTF8 support everything we know and love in BeOS.. with the the very very hard work involved: BFS took more than 1 or 2 years Dominic to complete, not to mention BONE (and from a semi-recent discussion on BeNews it would seem integrating the BSD stack in a MThread OS is a no-no or what ?) >* Starting at some point when our OS becomes useable, we will have to provide > binary compatibility (not compatible to BeOS, but to earlier releases of our OS) > so that serious companies can release/SELL binary programs without source code!!! > I is impossible to force the use to recompile every program if he updated his OS, > and no serious company will create software for your OS if you keep on braking the > compatiblity. "we" should not be Linux with their infamous glibc.so, exactly. Much more like BeOS. >* It might be the best decision to take the Linux or freeBSD kernel and modify it > until it fits our needs, rather than write the whole thing from scratch (this would take years) BSD has a much much friendlier licence (and community.. try to picture the BSD crowd making our life miserable because they don't like what we're doing and threatening to sue our *ss to court.. you can't? right, I cannot either) but with Linux we get XFree w/ OpenGL (accelerated in HW) and more drivers... >* We would need a BWindow BView compatibility library, so existing BeOS programs > can be ported by recompiling. Yes.. imagine that, we're actually discussing the possibility of re-implementing the App&InterfaceKit and its 10*? man-years of work (count'em).. shiver. But it's the most important part, or second most. >* We should have a *much better* GUI from the programmer's side of view, fixed coordinates are > evil, for example font senisitvity is needed. if we want to be source-compatible, we have to provide the same BView ctors et al, period.. Which doesn't preclude having another layer on top of that. (sorry for the "definite" wording, GUI is a pet peeve topic o' mine :-) >* No longer maintained BeOS should be forgotten. We don't need BeOS binary compatiblility, > only to be able to run a few ancient programs. >* I personally do think that many programmers are available that would participate in such a project. > Most of them are not interested in BeOS programming nowadays because it is not free, and it has > no future. The new OS project wouldn't have such problems. exactly, people often talk of AtheOS, but Kurt isn't interested in having true binary compatibility, and that should be a big enough show-stopper for everyone to forget about the idea of pulling AtheOS from under Kurt's feet and "force" another version.. A _true_ clone of BeOS on the other hand, would "belong" to everyone in the community, and much liklier to ignite a grass-root effort, developers and non-devs alike working each on their stuff.. though, compared to apps writing it's a proejct hell-bent over coding rather than doc writing et al of course. Wow, that's what I call to concur with your point of view. (except for the OSS thing..) >* Oh well, no X-Server please. WE should be able to do better. ah, but with XFree we get 3D and 2D drivers.. if there was a way to have "the best of both" (a custom made frame buffer abstraction like AtheOS or BeOS have for a "business apps" version of openBeOS and one based on XFree for a gamers version of OBeOS) but it's probably even more unrealistic than the rest, coordinating people to write on top of the ugly X11 API will be hard enough w/o having two implementations live side by side as it is... cedric. >* And yes, I know about AtheOS. But it is not the future of BeOS. > >regards >Marcus Overhagen -- http://cdegea.free.fr/ | BeDev E-16870 "God exists, and she loves Bill" -- BMessage