[haiku-development] Re: What's the status of Haiku?
- From: Simon Taylor <simontaylor1@xxxxxxxxxxxx>
- To: haiku-development@xxxxxxxxxxxxx
- Date: Mon, 25 Aug 2014 12:09:06 +0100
On 23 Aug 2014, at 19:44, Sia Lang <silverlanguage@xxxxxxxxx> wrote:
> On Sat, Aug 23, 2014 at 7:29 PM, Simon Taylor <simontaylor1@xxxxxxxxxxxx>
> wrote:
>
> >> Haiku works well on VirtualBox, and that for me is
> >> more than enough time for me to boot it up, play with it
> >> for a few minutes, remember the “feel” of BeOS that I
> >> enjoyed back in the day
>
> My goal is a BeOS clone that's a starting point for an OS that can be used
> daily for serious stuff
> on a wide variety of hardware. A hobbyist OS to "play around with for a few
> minutes" is not a very
> interesting goal. As Gassèe likes to point out, nostalgia isn't what it used
> to be.
My point is that it’s not the lack of driver support that makes it more no more
than a hobbyist OS for
me. It’s not the lack of support for esoteric hardware. It is purely that the
size of the ecosystem is not
sufficient. Fixing the lack of support for some modern or esoteric hardware is
not going to fix the
fundamental ecosystem issues. Using the Linux kernel is not a silver bullet
that will turn Haiku into a
serious desktop option and displace players such as Ubuntu (with a large
ecosystem, community,
and millions of dollars of investment behind it).
Therefore any potential switch to the Linux kernel is a balance between the
pros and cons. The pros
are easier support for new hardware using the larger driver community, and
(possibly, IMHO, citation
needed, etc) better performance and security. The cons are a lack of control of
the complete “OS
stack”, and a change in the character of the project. Essentially it all
depends on what your goals are.
If your goal is to support BeAPI applications on modern hardware, using a
userland layer on top of
linux is a perfectly sensible approach (and more sensible now than it was 14
years ago). That is not
necessarily the same as the goal of Haiku, and certainly not the same as the
personal goals of some
of the developers, for whom a modern, clean, complete desktop OS stack is an
interesting hobbyist
proposition.
> >> If you think there are a large group of ex-BeOS fans who
> >> are just waiting for the day when they can run their old
> >> BeOS R5 apps from their shiny new USB3 HDD, then I
> >> think you are mistaken.
>
> But then again, that's not what I said at all.
But then you have never clearly stated what your goals are and who the audience
is. I picked up on
your assertion that Haiku was being “disrepectful to BeOS fans” by not simply
taking the pragmatic
decisions that would make BeAPI apps work on their hardware in the shortest
timeframe possible.
However outside of the (small) group of ex-BeOS fans, that is not a
particularly interesting proposition
by itself. There aren’t many BeAPI apps (either from R5 or afterwards) that are
really worth using. And
despite it being a pretty good API, it was clearly lacking in several areas
(Layout and
Internationalization being two that Haiku have addressed, others remain). I
don’t think right now
there’s much of a argument for convincing people to write apps to the BeAPI -
the audience just isn’t
there, and many alternatives exist (eg Qt is a decent C++ API with good
cross-platform support).
So what’s the expectation? That switching to a Linux kernel will lead to a
massive influx of users
building modern, cool BeAPI apps? If we build it, they will come? Haiku have
basically built it. Haiku
works well for a large proportion of users. USB3 is not going to significantly
shift that equation. 3D
hardware support would be nice, but few current applications use it, so I also
don’t see that having
much impact on the level of user interest by itself. Obviously a re-imagined
Haiku Desktop
experience built around a guarantee of a compositing app server might bring
some fresh interest,
but that is clearly a massive design and implementation job in itself.
> I think becoming a serious alternative is Haiku's goal too.
I’m increasingly of the opinion that this goal is very unlikely to happen
without a massive investment in
the entire ecosystem. Creating a new Desktop ecosystem is hard, and expensive.
That’s basically the
thing that sunk Be in the end, and they were pretty well funded and weren’t
competing against the
same reasonably high-quality and usable Linux desktop distros that are around
today.
That’s why for me Haiku should just embrace its hobbyist status. It’s come a
hell of a long way for a
hobbyist OS, it has given plenty of people fun, interest, and experience of
working on all levels of an OS.
We do however need to accept that our ecosystem is too small for anyone to
invest time in creating
worthwhile apps. Until a Shuttleworth-type character comes along I’d suggest we
just continue to have
fun and just accept that world-domination on the Desktop (which was my hope,
back in the day), is not
going to happen until someone with a spare $100m dollars takes an interest :)
Simon
Other related posts:
