Date: Tue, 27 Jan 1998 13:52:38 -0800 From: "Jordan K. Hubbard" <jkh@time.cdrom.com> To: hackers@FreeBSD.ORG Subject: /usr/src/release/sysinstall needs YOU. :-) Message-ID: <26477.885937958@time.cdrom.com>
next in thread | raw e-mail | index | archive | help
[And no, I'm not sure what the poster for this would look like so maybe we should just skip that part] I just wanted to make the point as 2.2.6 comes up for release here that sysinstall is NOT a sacrosanct piece of code that I growl at people for touching, far from it, and if you have a great yen to substantially improve FreeBSD's "initial out-of-box picture" then making the installation easier/better is a really good way of doing that. I'm always interested in seeing 3rd party improvements to this area of FreeBSD and sysinstall is no exception. In fact, to put it really bluntly, I'm *sick* of sysinstall and would be happy if I never had to touch it again. For those who haven't been following its long and sometimes sad saga, let me give you a brief recap: Back with FreeBSD 2.0 in late 1994, I decided that the existing shell-script based install was too arcane for most folks to cope with and so I prototyped (and I repeat: PROTOTYPED) a new installer based on the rather more visually colorful dialog library and some ideas that phk had worked up for front-ending the guts of disk partition/labeling through the libdisk library. Through sysinstall, I explored and in some cases abandoned various approaches to installing FreeBSD and generally used it for the purpose that all good prototypes should be put to, namely working out what the final product is going to look like. Unfortunately, two things happened in the years following which really threw a wrench into this scenario. One was that I became so much more tied up in project meta-issues like recruitment and arbitration with squabbling project members that my time for doing any actually technical work was substantially diminished. I can't maintain a highly public presence, deal with various project logistics *and* sit down to hack on installation bits for 8 hours a day, there just aren't enough hours available. The second thing that happened was that sysinstall worked essentially too well and there was too much temptation to simply extend it rather that start over. The prototype ssentially became the product and we ALL know what those sorts of products look like - they look like Windows! :-) So now we have essentially the Win95 of system installers, all flash and gaudy colors on the packaging with no real "guts" underneath that could be used to implement a substantially more robust installation and we're stuck with it until a wholly new framework (sysinstall, New Technology edition :) can be devised. We've also just hired Mike Smith at Walnut Creek CDROM to take this one off my plate and hopefully good things will finally start to happen for sysinstall in the near future, but it's not going to help us in the short-term. What would help us in the short term would be some time spent on improving our "Win95" of system installers and essentially take the same approach that any company does with legacy code - try and make it crash a little less often and deal at least with the easier to implement customer enhancement requests. That's all we need to do, just keep it alive a little longer. However, like I said, I've also done a lot of "time" with that code and hacking on it is not really my fondest occupation, though hacking on it in partnership with someone else always goes easiest. Since I also hear requests from time to time for "something to do" from our general population of volunteers- at-large, please let me be the first to point people at /usr/src/release/sysinstall and say JUMP IN! The water is muddy and there might even be some nasty diseases in there, but SOMEBODY's gotta do it! :-) Even if it's just to help me edit the docs in /usr/src/release/sysinstall/help and /usr/src/release/*.TXT, I'm all for a little help. Some of that stuff has gotten _really_ mouldy and could use a good housecleaning from someone who's not me (I've been looking at it for too long to see its more glaring flaws). Thanks! Jordan
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?26477.885937958>