NOTICE: This version of the NSF Unidata web site (archive.unidata.ucar.edu) is no longer being updated.
Current content can be found at unidata.ucar.edu.
To learn about what's going on, see About the Archive Site.
NOTICE: This version of the NSF Unidata web site (archive.unidata.ucar.edu) is no longer being updated.
Current content can be found at unidata.ucar.edu.
To learn about what's going on, see About the Archive Site.
OK. I've run SunOS, Solaris, BSD, Linux, HP-UX, OS-2 and
NT. I'll group these together from a software engineering standpoint
since they are all designed around good basic principles:
simple kernels
layered (abstracted) device interfaces
prioritized multitasking
NT is the new kid on the block. It does not have the maturity that
30 years of evolution has given Unix. It doesn't have the optimizations
of Sun. Nor the rugged reliability of HP-UX, nor the simple support
of a world wide user community of Penquin geeks. But it does have
the MFC/SDK and some very good people producing some of the best graphical
programing interfaces in a solid visual programming environment. For
weather visualization, this is a clear advantage. The Xwindows interfaces
are good,
but, I will not call the shared memory interfaces ruggedly reliable without
a lot of semaphore and signal processing. In either case, data mapping,
overlays, filtering, blending, etc, use the same graphical algorithms.
For myself, I like my command line, ccom, make, link...in short
the Unix programming environment. But the Motif interface is just not as
slick as MFC, Windows SDK and Visual Studio. In the future, the vast user
base of Windows and NT will certainly breed a new mind set in some
programmers.
Tom, is correct in hedging his bets. The demand for Windows based
tools for the weather community is rising. The simple ease of use for
web delivery of weather maps, radar animations, and other data, prompts
most undergrad students to avoid the unfamiliar Unix based tools.
Any way, I wouldn't count out NT or Mcidas-X on NT, or something
very like it. The demand is there, regardless of my preferences. And,
the precepts of NT design are borrowed from Unix, soo, I expect it will
get better ("form follows function.")
Note of farewell:
This will be my last comment to this group. I have moved from
Atmospheric Science to Chemical Engineering here at Cornell.
The list has always been the best of the support groups I've
been privaledged to be a member of.
Thank you all!
jdm
At 02:48 AM 4/28/2001 +0000, you wrote:
>I think your last paragraph says it all to one of my points..
>I have a Solaris box that has the original Unidata apps from
>1998..the box has run 2.6, 7 and now 8 and never has
>anything broke with the upgrades. The cost thing
>about Linux is completely a non-issue anymore. You can
>download Solaris for free now. All the gcc tools
>are there..Cost is the same.
>
>We definitely agree on our dislike for Windows. I truly
>cringed when I saw McIDAS-X for NT. I am very glad to see
>that Tom is not supporting it fully yet..I think that's
>a bad path for met departments to go down.
>
>Anyway..that's enough for that. I got out my propganda.
>
>Robert Mullenax
>
James D. Marco, jdm27@xxxxxxxxxxx, jmarco1@xxxxxxxxxxxx
Computer Operations Manager
School of Chemical Engineering, Cornell University
Office: 361 Olin Hall (607)255-7312
Home: 302 Mary Lane, Varna (607)273-9132
ldm-users archives: