On Wed, 24 Jul 2002 12:22:34 +0200 Eduard Bloch <[EMAIL PROTECTED]> wrote:
> #include <hallo.h> Greetings :) Why is it that you think debian-installer is somehow worse than boot-floppies? Have you tried to build either? See below. > Nice. I have seen how motivation disappears when you are waiting > longer than 8 months for the next _freeze_. I expect that the whole > thing will slow down and we really won't get a stable installer until > one year is passed. Let's test that speculation, and I'm betting on the opposite, and not only that, but that the people who wrote d-i have done so with their own interests in mind: their time. I would also expect that they have done so with the interests of the other developers, as well, by making d-i easier to work on from at least the standpoints of the volume of work necessary to make changes to the installer itself and that of the time spent in getting to understand the internals of the installer (perhaps for the purpose of enhancing it, maybe in some major way). I'm speculating here. However, I want to see the actual results, so I would like to propose that several people step up as test subjects who have never seen the process by which either b-f or d-i is built. While I do intend to do the builds as noted below, I have to recuse my- self from actually reporting results because I have built and run b-f before, and have gotten extremely frustrated for days after as a result of the problems encountered. The task: On a fresh machine with DSL or approx equiv, install debian woody or sid, base -only-. Note: you will be repeating for the other installer, so if you start with woody, also use woody for the second test. Start a timer, begin learning how to build b-f through to the point of b-f building all the floppy images, do the build, installing what packages you need to as you go. Stop the timer, and record the time, and any comments. Then, blow it all away, install the base again (if you used woody before, stick with woody; same for sid), reset and start the timer, learn d-i, install, run to produce installer images. Again, stop the timer, record the time and any desired comments. Some should start with d-i, others please start with b-f. I'd especially like to hear about anything that either b-f or d-i has done to make it easier for you to do any part of the task as outlined above, or any part of any -other- task, such as altering the installer, what it presents, the order of things, add a hello-world message at any point or arranging for different packages or other initial material than is set up for the default install. How much time did it take for you to determine everything it would install by default? The foregoing would show how difficult and time-consuming basic development would be with each, and would also serve to shake down some d-i bugs :) The results would tend to motivate developers toward things that are better to work with. > > You're welcome to try to prove that this isn't the case, but > > everyone else has seen the way b-f's work a few times now, so don't > > expect anyone to believe your claims until you /have/ proved them. > > This is not a proof, but a speculation. Only the time will prove. So, Let's test! :) If you don't think this to be a good idea, please retract your speculation as being something you don't believe should be tested. > But I think that this is exactly this attitude that caused the fucking > long release period of Woody. And you are continuing working the same > way. Bzzt, wrong, nope :) Attitude didn't cause the delay... the freeze period was prolonged by a download server architecture problem, and remember that there was added the option of installing with linux-2.4.x kernel. In addition, several new structures have been added to the packaging infrastructure to allow major new features, one of which is the ability of apt to grab whatever is needed to -build- a given package. I know I'm forgetting other alterations to the base... but any time you do alter the base, that has far-reaching consequences. Also, debian as an organization is composed of a large number of sparsely-located developers. It takes time just to synchronize us, and you will be torn by sheer force if you try to maintain the sync! You can't blame just one person for a delay which actually resulted from work done building and testing new features. Which is the same every time debian is released... things get added, gotta test, gotta work on the problems, gotta listen to other users and developers, gotta take some or all of their comments into account, gotta do the work that represents. That takes time. Once and for all, -deal-! :) > Gruss/Regards, > Eduard. -Jim -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
