Mike, Doing work in the area of high performance computing, I really think that commercial UNIX is here to stay, no matter what. At this point, Linux PPC is still quite early in its development, compared to something like AIX. Linux SPARC is not as mature as Solaris. Linux MIPS is far from IRIX. Has anyone done any benchmarks comparing Linux Alpha to Digital UNIX? Going a little bit astray, I don't see Linux being used on a machine with thousands of compute nodes any time soon either. Some people can now squeeze computation power of gigaflops out of Linux-based clusters. But what about the teraflop stuff? Are there anyone from Berkeley on this list here? Or maybe someone from one of the national labs in the US? Maybe they can tell us what's the power of those machines, the architecture of those machines, the configuration, the design requirements, the policies of the authority, etc. Recently, I went to the GeoTriad, a convention of geologists, geophysicists, and well loggers. There, once again I was reaffirmed that commercial UNIX is going to be around for quite a while still. Linux is now running on the Fujitsu AP/1000. But that is just one of the many parallel machines in the world. It is not even one of the fastest in the world (at least not that I know of). What is needed in the industry, which time and money is of great importance, is a reliable solution. I have been told by quite a number of people at geophysical software companies that cost is not a problem usually. Reliability is. So until Linux is tested and proven on the large machines, it might be a bit hard to convince a lot of software companies as such to do it. I imagine that would be the case in a few other industries other than geophysics as well. As such, if commercial UNIX continues to live on, Motif will. And there will be people like me doing work on Linux, and then target the code for commercial UNIX. That's the reason why I am using Motif. Until KDE or GNOME or any other has a clear lead on all UNIX platforms, regardless of commercial UNIX or not, I will probably stick with Motif. Actually, very soon, I will be working on a 32-node Alpha cluster running Digital UNIX. On and off, I have used Digital UNIX on another machine and so far, I haven't had any complaints about it. What are the problems that you have with Digital UNIX? I use AIX on the IBM RS/6000 machines most often. I love the RS/6000 machines. For what it does, I think AIX is fine. Actually, recently, I am having some problems with some computation on the RS/6000 machines. In some odd cases, I have the problem of zero not being zero. But that may be because of my code. Until I sort that out, I can't say it's IBM's problem or my problem. I can only say that there is a problem. Other machines that I use often are the SGI machines. I think IRIX is great for what SGI machines do best too -- graphics. So as you can see, I don't see why people have problems with certain UNIX flavors running on certain hardware. I am also using Solaris, SunOS and NeXTStep as well. Do I have any problems with them? No, except NeXTStep. I can't stand the GUI. But I think that's just a matter of personal choice. Can anyone make me convert to anything else, move away from Motif? I am really open. Thanks a lot for your interesting input. Regards, Kenneth Tan P/S: I can go on and on rattling about OSes, draw in stuff like NetBSD, FreeBSD, SCO UNIX, OpenBSD, Hurd, etc. That's a large component of what I have been doing in the last few years. But I don't think I shall lead this list astray. Write to me directly if anyone wants to know what I know and what I think about them. ------------------------------------------------------------------------------ C. J. Kenneth Tan E-mail: cjtan@acm.org Telephone: 1-403-220-8038 cjtan@ieee.org 1-403-606-4257 URL: <A HREF="http://www.cuug.ab.ca/~tanc"><A HREF="http://www.cuug.ab.ca/~tanc</A">http://www.cuug.ab.ca/~tanc</A</A>> Facsimile: 1-403-284-1980 "An engineer made programmer is one who attempts to solve a problem, A programmer made engineer is one who knows how to solve a problem." ------------------------------------------------------------------------------ On Sun, 19 Jul 1998, Michael Lankton wrote:
C. J. Kenneth Tan wrote:
What I really don't quite get is what about Motif window manager, mwm?
There are just much nicer alternatives in linux. Plus motif is in the same boat as QT, no one in the linux world wants to support it because it isn't GPL'd. There is lesstif, yes, but gtk offers so much more flexibility, and wings (the widget set used by Windowmaker) is going to be really special as well.
What is going to happen to it? I know this is like fortune telling, but I just thought it would be something to think about. From what I see, Motif, CDE will continue to live on AIX, Solaris, HP-UX, Digital UNIX,
Commercial unices are going away IMO. Compaq has already announced the discontinuation of Digital Unix development (and good riddance I say, I use OSF1 at school, and it's no prize), and do you really have confidence in sgi's or Sun's long term success at this point?
Will the Gnome stuff be a separate stream then? Will Gnome stuff be the only thing on Linux then? I have no knowledge about Gnome at all, but what is it essentially? How different will > it be compared to Motif?
The easiest way to describe Gnome is Kde sans native window manager. That is one of the reasons why I believe Gnome will eventually succeed and Kde will fail. It isn't proprietary, and people like a choice. The other is license. Gnome is built on gtk (GPL'd), kde on QT (non-GPL)
-- ==================================================================== Michael Lankton <A HREF="http://tasteslikechicken.ml.org"><A HREF="http://tasteslikechicken.ml.org</A">http://tasteslikechicken.ml.org</A</A>> ==================================================================== - To get out of this list, please send email to majordomo@suse.com with this text in its body: unsubscribe suse-linux-e
- To get out of this list, please send email to majordomo@suse.com with this text in its body: unsubscribe suse-linux-e