Willibald Krenn wrote:
Tim O'Neill schrieb:
The original comment on support I think may be a valid one. If Novell wants to break into the desktop market in a serious
way then there should be more information available.
I have been trying to get a desktop to allow log in through a
Windows domain and have tried many things to no avail. I also was told that the question was a chargeable one.
Well, if you ask me, Linux isn't really ready for the ordinary Desktop user.. I don't think that SuSE -ahm Novell- can change that in the short term. I'm running 9.0 AMD64 here and browsing SMB has never worked for me. KDE never managed to set the volume of my sound device - I had to use alsamixer for doing that. Some programs I tried installing from SuSE supplied DVD just died SIGSEGV when invoking for no obvious reasons. The YAST kernel update once upon a time was working - today I do this by hand, cause I don't trust YAST on getting the job done (stalled downloads etc.). But I don't complain - that's Linux. Nothing's really working here as it should.. ;-)
If SuSE was to build a suitable desktop distribution, I fear SuSE would have to start from the very beginning and just supply a small set of applications that are widely tested, were found to be stable, are pre-configured in a sane manner, come with sufficient and in-depth documentation and can be updated over several years without having to install a completely new base system. However, this is impossible to achieve given the development process of open source software in general.
This in itself is fair enough however the big difference between Microsoft and Suse (for example) is that Microsoft have on their site a huge amount of information, white papers, and 'best practice' documents that anyone can look through. The basics are normally covered so if you are simply evaluating a product you can do this without real cost.
Actually there is plenty of free information available about Linux. You "just" have to invest enough time to find the right piece.
Willi
I have read this whole diatribe with amusement, the guy who started it is obviously a clown, & was tolerated far longer than I would have tolerated him. Kudos to SuSE for that. I disagree that Linux itself isn't ready for the average user, but I agree that the current level of 'fit & finish' of various distros (SuSE least among them IMHO) is problematic. I have worked on SGIs professionally for about the last 12 years. They are &%$*(&%$ expensive, but smooth as silk, well documented, everything just works, etc. I consider that 1 end of the 'fit & finish' spectrum. Then there is the Linux of the mid '90's, with cryptic installs, mismatched packages, for uber-geeks only, the other end of that spectrum. The Linux community has made good progress from that state, but still seems too tolerant of screwups that some (hopefully simple & cheap) check-outs at the distro level could eliminate or cut down. IMHO, that wouldn't/shouldn't be much of a hurdle, & I place responsibility for that with the distro packagers. They can/should do what they can to assemble their packages from parts that really do match, rather than always striving for highest package/kernel version #'s. The bleeding edge is cool, but only when it works. I realize that the distributed nature of the Linux developement model, as well as the nearly malignant variety of different hardware makes this integration difficult, but it seems that a (hopefully) small amount of extra attention there might clean up a larger fraction of screwups/package mis-matches/etc. SGI & most other high-end builders have sophisticated testing regimes (automated, software based) to verify that all of their stuff not only works, but works together with other included packages. With the increased power of machines today, it seems to me that some of the distro packagers could make some progress in this area, even if only by some brute force methods.