From: vassilii AT optimedia DOT co DOT il (Vassilii Khachaturov) Subject: Re: Beta-19 and configurations.... 9 Feb 1998 19:23:27 -0800 Message-ID: <34DF4CCE.3CC9DDF0.cygnus.gnu-win32@optimedia.co.il> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: Ronald Van Iwaarden Cc: "Gnu-Win32 (E-mail)" I suggest we all carefully think over and design WHAT exactly is to be included into the distribution of b19, and what options should the setup be given. Maybe, it's worth to do less things in the setup GUI, and package a minimalized perl+win32 support to manipulate the registry from upon the installation? Things I think are important from those that are not distributed currently, and were not cleanly announced previously as those to be icluded into b19, are (as understood, optional -- to be selected at the setup stage): * THE SOURCES -- ready to recompile all the tools in-place This could be great in boosting up the development. I personally think of some hacking around with cygwin.dll -- but still had no time with it, especially due to the initial overhead of the sources installation and compilation env. setup. This is the way they did it in Linux, and it worked. * (previously discussed): termcap/curses, minimal terminal programs * Natively compiled editors (vim, emacs --?) * a couple of terminal games compiled btw check http://www.wisdom.weizmann.ac.il/~vassilii/pub/boa/ * definitely the "remote" suite -- I personally have had several people seduced into gnuwin32 just showing them a colorful telnet into Linux with 'midnight commander' running there -- and a WORKING (as opposed to the one from the NT resource kit) inetd just with the in.telnetd capability. * think of package management and clean upgradability (wouldn't you like a simple script to turn the symlinks from b18 to be "system" when going coolview) * X11R6, clean instructions on setting up X server and links to all the tested X server packages. I know the gnuwin32 goal is more like "you have the kernel, do things yourselves" right now -- my cry is that this goal is old enough, and the things grew, so why not changing the goal? Setting up a primitive archive network of ported things, someone specializing on distributions/packaging -- and things will evolve faster. > Hmmm. I thought the install was relatively easy and I an new to WIN32 Well, this is exactly the point. I thought it's easy also -- I'm not afraid of setting environment variables manually, reading through manuals etc. I do have an experience of tweaking with various programs source distributions until they build on gnuwin32. I did lots of low-level and realtime hacking in my life and my hands can't get much dirtier. :-) However, what I am speaking about, is a number of suggestions to invest in gnuwin32 promotion by making "quick start" possible as the initial bundle setup becomes more automated and includes more components. Those that are not to be convinced that gnuwin32 is a useful tool -- like you and me -- are already using it, and can live w/o the nice things I had mentioned in my previous letter. I was mainly speaking about attracting new people -- and that implies later naturally turning some of them into another and another gnuwin32 promoters/developers. In addition to my old examples, think of the successful 'configure' script approach -- installation/configuration is much simpler now, and sysadmins are praising the autoconf scripts, that freed them from most of the makefile tweaking. Think also of the emerging KDE/Gnome initiatives. Enough just hacking for yourselves, make things attractive and usable -- and get the whole world praise in return (enough for me personally for all the free UNIX consulting/soft. devel. I'm doing in my spare time), or, if you feek like earning your living from it -- make it even more attractive, so that more people think of using it and buying support from you. - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request AT cygnus DOT com" with one line of text: "help".