On Mon, 05 May 2008 13:46:33 -0400, Stuart Brorson wrote: > My suggestion is to look back 3 or 4 years > and see what was on common Linux distros back then and make that the > minimum. For common Linux distros, I'd suggest: > > Fedora > SuSE
Suse in 2004 was still commercial and quite different from OpenSuse now. > Debian Debian is notorious for their long release cycle. Four years ago Debian stable was represented by Woody which got the feature freeze in autumn 2001. I'd rather not run a desktop on > Ubuntu wasn't even born four years ago. > But remember that we also support other Unices: > > FreeBSD (Dan) > Mac OSX > > So we'd need to see what dependencies exist on those platforms. The gecko engine uses cairo and Mozilla/Firefox works on these platforms. Is it naive to deduce that cairo support for geda will work too? > In any event, my key points are: 3 -- 4 years since that's the > lifetime of a computer on the typical desktop, The notion of not upgrading a computer is common among windows users, but but not so among the linux crowd. The reason is twofold. Upgrades are cheap and development is fast and generally for the better. Four years ago, sound on linux was still a pain, graphics acceleration a pipe dream and openoffice no viable alternative to MSOffice. None of my linux friends still runs the same overall version of their software on their desktop like 2004. In fact, most of them are in step with their distro of choice, even if they run it on dated hardware. The oldest supported Ubuntu is dapper drake from 1 June 06. The oldest supported Fedora is moonshine released on 31. May 07. Do you really want to restrict the geda project to resources of distros that have been abandoned by their maintainers? > and look to see what > pre-exists on all the common distros (not just one of them). ...and make sure, geda/gaf and pcb get packaged by them. This is way more important than depending on the least denominator of libraries some time in the past. > Remember that lots of gEDA users are not hackers. They're EEs who want > to use the tools without apt-get or yum. Yes, most geda users are not hackers but ordinary users. They typically use the precompiled binary of geda/pcb that their distro provides. It is the nerds and hackers who like to path there own way through software wilderness and actually compile from source. Ordinary users already feel bold if they use Ubuntu and instead of Microsoft. IMHO a dated look and feel of the geda/gaf suite will result in less users. When I decided not to shell out 6 kEUR for a protel license three years ago, I was put off by the look of the gschem GUI. Fortunately, the alternatives including eagle didn't look and feel any better. But the others improve with time. Last time I checked Kicad didn't feel like a toy any more. Eagle on linux recently got a boost regarding pan and zoom speed. Seems to me, it is faster on linux than on win now(!). > Indeed, they may not know that > these tools exist. Too often they are turned off by the complexity of > getting gEDA running on their particular platform. This is true for windows users. They have viable alternatives with eagle and the various not so free EDA tools for win. Where do solaris users turn to if "make install" of geda insists on failing with errors? > This is exactly what Ales' binary distro does. We probably want to push > that harder as the preferred install package on the gEDA website. If you do, then please make the approach more linux distro friendly. That is, ask the user to install missing libs from his distro first. Only after this fails, revert to compiling a private version of the libs. Back, when I did my first steps with geda/pcb this was a major source of pain. ---<(kaimartin)>--- -- Kai-Martin Knaak http://lilalaser.de/blog _______________________________________________ geda-user mailing list [email protected] http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
