[haiku-development] Re: To __BEOS__ or not to __BEOS__?
- From: Luposian <luposian@xxxxxxx>
- To: haiku-development@xxxxxxxxxxxxx
- Date: Fri, 9 May 2008 00:33:49 -0700
On May 8, 2008, at 11:23 PM, Fredrik Holmqvist wrote:
Sooner or later we probably want to go 64 bit as well, which will
definatly break things.
Now THAT would be a welcome breakage. 64-bit Haiku with a 64-bit
journaled file system! Sweet deal indeed!
Bring it on! The sooner, the better, as far as I'm concerned.
However, given how little "mindshare" BeOS R5 has nowadays, except
among us rabid loyalists, would it be more prudent to simply make the
"compatibility break" here and now, as we approach R1 and say "sorry"
to backwards compatibility? It's a nice idea and all... but will it
simply make things HARDER to break away from, in R1.5 or whatever? How
many of us are really willing to live with a Haiku that isn't all THAT
much different from the BeOS that it was spawned from? Do we risk
losing more people to "gee, that's nice..." type of glance and
departure, because Haiku isn't a big enough difference from BeOS, if we
go with GCC2?
Maybe it would be best if we just apologized now and made a clean break
from all BeOS compatibility (source and binary) and released Haiku as
an OS that stands in appearance similitude to BeOS, but it's honestly a
different beast underneath?
I think the sacrifice would be worth it, in the end. I think we've put
to much into Haiku... made too many improvements... added too many new
drivers and features, we risk hurting the overall health of Haiku
(ultimately, the "mindshare" of end-users/developers, etc.), just to
please a few people who want to still hold onto the past, while looking
to the future. We may not be able to have our cake and eat it too,
without ruining the whole cake in the process, by tripping over
ourselves in trying to satisfy both sides of the equation.
As much as "binary and source compatibility" was the intended goal, I
think "feature creep" has snuck in too much (too many improvements over
and beyond BeOS) and made it more difficult to still accomplish that
"intended goal", while keeping an eye to the future.
I think it's time we faced facts and admitted that we added too much
extra, while building Haiku. Had Haiku been released as R1 a couple
years ago, with only one or two small improvements over BeOS R5, the
purpose of "source/binary compatibility" would have held validity. But
who can honestly say Haiku isn't so much better (and going to be, all
the more so, the further towards R1 we get) than BeOS R5 was, it's
almost not even worth considering source/binary compatibility anymore.
Hmm?
Luposian
Other related posts:
