[haiku-development] Re: R1/a4 initial planning
- From: Sean Collins <smc.collins@xxxxxxxxxxx>
- To: haiku-development@xxxxxxxxxxxxx
- Date: Wed, 22 Feb 2012 17:28:02 +0000
Stephan Aßmus wrote:
Sometimes it is, sometimes it's not. There are plenty of projects out
there where the destination seems to be more important than the
journey and that's perfectly fine. After all, we need results and
something to work with. However, it's not the point of Haiku to be
build just like another Linux distribution, only with a less capable
kernel and yet another API called Be API, similar to Qt only much less
complete.
But let me tell you something about only putting together parts. The
way Linux distributions work, they want this steering wheel and those
brakes and this engine but with that exhaust pipe. But unlike a car
(hopefully), software stacks don't necessarily fit together. And so
when they want this steering wheel, it only works with another set of
tires and including these certain brakes means they also have to
include a second pair of wheels. And so what you get is a car which
does drive, but somehow ended up with three different sets of tires,
combined with two pairs of wheels, somehow made to work at the same
time. But it works "good enough" and even looks pretty, since they
managed to hide the cludges with paint.
But if we wanted that, we wouldn't work on Haiku. Doesn't mean the
Haiku way doesn't come with its own set of problems, like for example
that the car is still unreliable, doesn't work with all kind of
gasoline and can only be driven on certain roads.
Best regards,
-Stephan
I understand what your saying, but even in racing, we have to often take
dissimilar parts and make them work. I wouldn't go bringing every
platform to Haiku, but given the prevalence of QT based software, and
the fact that with a good solid QT port Qtwebkit becomes a reality, I
think for some things, it might make life much simpler and lower the
overhead for the Haiku dev team. IE , maintain QT is easier then
maintain webkit, if QT is up to date, the qtwebkit browser can also be
updated much more simply. I built the latest Qtwebkit trunk last night
with a modest amount of effort, and thats really a modest effort for me,
it'd be no work for any of you who are dramatically more skilled.
But thats neither here nor there.
As to your comments about the BeAPI, if it is truly that behind, then
what other good options exist for the near term to bring the capability
the users,donor community are hoping for. To play the devils advocate
for a moment.
Just trying to spur a bit of conversation to get Haiku whipped into a
viable platform for beta.
Sean
Other related posts:
