On Fri, Dec 03, 2004 at 11:52:08PM +0100, Eduard Bloch wrote: > From my point of view, we could have released Sarge one year after Woody > with boot-floppies. The only thing needed was a bit more man power from > the porters. Instead, most of the "core team" and the BFs porters > stopped to work on it. And without manpower (motivated people), there > will be no development. Well, y'know, if they weren't willing to work on it, we can't exactly twist their arms, can we? Regardless of the reason, if there were an insufficient number of people willing to work on b-f, the reality is that we *couldn't* have released that way. If you were unwilling or unable to do the work yourself, and no one was willing to give you the help you needed, then to say we could have released "from your point of view" misses the big picture. > OTOH many new people jumped into the boat following the ideas of having > a cool, fresh-fashioned installer. IMO (only IMO) most of them had bad > memories about their first Debian install so they doomed BFs without > any closer look at the source. Speaking for myself, I had looked at the boot floppies source not too long after woody's release, to try to include an updated kernel for my employer. I found BF's monolithic structure difficult to grok, and time-consuming to manipulate; I believe in the end it was not worth the time for me to follow through on the project, instead muddling along with the stock installer and manipulating the kernel on each server post-install. In contrast, d-i's structure made it very easy to start hacking on, to the point that at the end of last year I had put together an XFS-capable install CD using a separate kernel in the space of a few weeks, and then (as someone who had never been part of a porter team) went on to write about 90% of the alpha-specific code for sarge d-i. I couldn't imagine myself doing the same for BF; and if I did, I'm afraid I have to say I think I would have enjoyed it much less. However good the code might have been in BF, the positive knowledge transfer and shallow learning curve that come from using a package-based design like d-i do have an effect on its accessibility to outside contributions; and when the primary blocker for BF was that people weren't helping, well... > Well, the outcome of d-i development is good, but not that impressive > (no GUI, no jumping penguins, confusing partitioner tool). Sometimes > they have made the same errors we did with BFs, sometimes new things > have been invented. I can agree with you that d-i is not perfect, and yes, you've pointed out two of the sarge installer's most notable weaknesses (I don't know anything about jumping penguins, though :). OTOH, I've also seen comments like "the sarge installer is the best Open Source installer ever" and "definitely easier to install than Windows XP", so clearly *some* users are impressed; and considering even the most fervent Debian advocates have always regarded the installer as a weakness, I think that's saying something... -- Steve Langasek postmodern programmer
Attachment:
signature.asc
Description: Digital signature