FreeLists / openbeos / [openbeos] Re: Thoughts on XFree86 development and how it effectsOBOS

[Michael Phipps <mphipps1@xxxxxxxxxxxxxxxx>]

Far be it from me to gloat over the distress of the OSS community. This 
license craziness (be it XFree86 or SCO or whatever) sets everyone back. My 
reasons for not choosing X and Linux are well documented from that time 
period and were based on tech, mostly - the Linux kernel ***AVAILABLE THEN*** 
didn't look like something worth supporting. Today, I *MIGHT* make a 
different decision (actually FireFly looks pretty interesting to me - if we 
had to start over, that might be a logical direction). As far as X, at that 
time, I had never seen X used in a way that was fast enough for BeOS users. 
B.E.OS did some work that is interesting along those lines, but AFAIK, it was 
never what I would call a "complete" test - certain aspects were 
demonstrated, but not others. Not, I am sure, from an intent to mislead, but 
because things weren't done. 

Anyway... All of that aside... What I would really hope for (beyond all 
hoping) is that the OSS community realizes what a treasure that OBOS R1 is 
(when it comes out). That they put aside their interest in Linux on the 
desktop and come to the Light Side. In many, many ways, R5 is still better at 
an OS level that Linux, for end users. Configuration is one. One common set 
of libraries is another. The integrated widgets (instead of 10 sets of add-on 
widgets) is a third. BFS is a fourth. Media Kit is a fifth. Translators, a 
sixth. The threading a seventh. Is this to put down Linux/BSD/Hurd/Whatever? 
No. But I believed, believe and will continue to believe until I see 
otherwise that what is good for the Server Room is not necessarily good for 
the Desktop and that Choice is Not Necessarily Good. That is really what OBOS 
has to offer. An integrated package that contains lots and lots of stuff that 
makes development easy and being a user fun again. If we were to have a 
mission statement, that might well be it. 

This is turning out to be sort of long... Oh well. :-) Something else to keep 
in mind is that thanks to the long efforts of Thomas and Rudolf, we have 
drivers for practically every video card that you can buy. Sure, there are 
some high end cards and maybe some older low end cards that we don't support. 
But we have accelerated (2D) ATI, nVidia and Matrox support. Additionally, we 
support embedded ATI and (I think) nVidia. So other than embedded Intel, we 
support about everything that I have seen recently in new machines. I am sort 
of torn over supporting older machines. BeOS has a wonderful legacy of being 
great on older/low end machines. The problem is that it drastically increases 
the amount of hardware that we have to support. There is a lot to consider 
there, but, honestly, you can pick up a decent nVidia or ATI card for < 
$100...

So, finally, I wish the Linux and XFree86 people well. Just because we didn't 
choose their route doesn't mean that we wish them ill. There is a long way to 
go for all of us. I just (personally) think that the OBOS route is a smarter 
way, in the long term, than Linux for the desktop.

Michael


On 2004-02-29 at 18:47:50 [-0500], Solaja, Zenja wrote:
> When OBOS was but an idea on the bedevtalk mailing list, a lot of talk
> evolved around which subsystems to reuse in order to recreate the corpse
> still lying on the morticians table. Thats when NewOS was considered, as
> well as XFree86 as a display system.  OBOS went down an interesting part
> (deciding not to use XFree86), and a few people went their own way,
> grumbling about the unwise choice of not selecting XFree86 and a large pool
> of 2D/3D drivers we'd get for free.  With the latest news on the XFree86 4.4
> front (all computer news websites, about the split and creation of several
> factions or XFree86, which are too many to list here, just Google it), this
> actually brings OBOS into an interesting position.  With the disappation and
> fragmentation of the XFree86 world, it seems that it is less likely that all
> flavours/descendants will gain accelerated drivers.  This might finally be
> an straw which breaks the camels back (metaphorically), and force video card
> manufacturers to publish open specs for their hardware.  Fingers crossed...
> 
> The pace of advancement in other OpenSource projects looks like it will slow
> down dramatically, as far as video / display technologies are concerned.  An
> opportunity for OBOS, perhaps?
>