[openbeos] Re: Bootable Test CD

  • From: Thom Holwerda <slakje@xxxxxxxxxxx>
  • To: openbeos@xxxxxxxxxxxxx
  • Date: Wed, 26 Mar 2008 08:52:55 +0100


On 26 mrt 2008, at 05:14, Jorge G. Mare wrote:

Howdy,

On Tue, 2008-03-25 at 20:47 -0400, Karl vom Dorff wrote:
It's hard to say, I agree with both sides of the story. I think maybe
we should make the CD version available to registered members only.
The public won't see it, and there's only >250 members at Haikuware -
I think they're all well aware of where Haiku stands at this point as
well.

The devs have made a conscious decision protect the quality of the Haiku brand by delivering only when a certain level of confidence is reached. Let's respect that decision. The live CD, the alphas/betas, and even R1 will eventually come; we just need to be patient and give the devs their
time to do their work.

Time to play devil's advocate.

This knife cuts both ways, you know. While I understand developers get cold feet with things like this, there is also a time when you have to "cut the damn [umbilical] cord already!" What I mean by this is that is this: in x months, we want to release an alpha, and more or less, we have decided it should be a working live CD. But, uhm, how are we going to release a working live CD when no one gets to test this feature? It's a fairly crucial part, so shouldn't we applaud any efforts to make the testing of it easier?

It is similar to the discussion we had a few weeks ago on whether or not the alpha should contain a working installation routine (partitioner+installer). My argument then was that this is such a crucial feature (if it fails, the Haiku experience is over), that exactly because of it being crucial, it should be included and enabled in the alpha. This falls under the same header if you ask me.

I'm not saying the live CD should be released *now*, I'm just saying that we have to think about the fact that we can't keep the umbilical attached forever, and then cut it two days before we release the alpha, and then expect it all to work just might fine, without it having seen any real-world testing. VMware is nice, but it shouldn't become an obsession. In the grand scheme of things, VMware is a crappy, stop-gap solution that we should strive to run away from screaming schoolgirlishly, instead of stick to.

This is all apart from the branding discussion, of course.

And fwiw, iirc, and imho and all that stuff.


Thom

Other related posts: