20010913: Input on software future, hardware upgrades

>From: "Arthur A. Person" <address@hidden>
>Organization: Penn State
>Keywords: 200109131352.f8DDqZ110929 computing platforms

Art,

>We're nearing a time when we will need to do a major upgrade of our
>classroom computing facilities which is currently comprised of 10 Sun
>Ultra 1 workstations.  Given the current environment where users demand
>both Windows and Unix applications and the degree to which Linux has
>advanced, we are considering replacing these systems with PC systems
>which will dual boot either Windows or Linux.

A number of existing Unidata sites are moving in this direction, so
you would be in good company.

>However, we are faced with a dilemma...  the garp/gempak applications,
>which are our current mainstay for presentation, seem stuck in the
>8-bit graphics world.

This may change.  The reason I say this (this is no promise) is that
NMAP is supposed to be adopted as a base line application for AWIPS.
This seems to mean that it should be able to run on the same platform
as AWIPS.  If so, it seems that it would need to be modified to run
in 24-bit displays since AWIPS _must_ run in 24-bit mode.  Also, AWIPS
latest offering is FX-Linux, a Linux (RedHat 6.2 and 7.0 at the moment)
port of the mainstay that runs on HPs.

>The latest versions of RedHat Linux, when run in
>8-bit mode, use up so many colors that garp is unable to produce
>displays, making Linux unusable.

Actually, we ran the last set of training workshops under RedHat 7.1
Linux.  Chiz run under the KDE window manager since it uses a lot
less colors than Gnome.  He also twiddled some settings to reduce
the number of colors used by desktop icons and the like.  All-in-all,
the workshops were a great success, and the boxes used (733 Mhz PIIIs)
were nice and fast.

>With MS Windows, the X window
>environment is not directly supported so 3rd party software must be
>employed at additional cost and loss of the Unix environment.

We have been playing around with the free Cygwin software (not a
determined effort, just playing) to see if it would allow one to run
our existing applications under Windows (98, NT, and 2000).  Don Murray
demonstrated that Cygwin coupled with XFree86 4.0 allowed him to run
Unidata McIDAS-X from another box (X display on NT; application running
on a different machine running Linux).  It is my intention to test this
environment to see whether or not McIDAS can be built and run under the
Cygwin environment directly.  I figure that if McIDAS can run in the
environment, then GEMPAK can also.  This remains to be seen, of course,
and the "project" is ad hoc right now, so comments must be taken with a
grain of salt.

I just chatted with Chiz about the Cygwin concept.  He told me that he
already reported to you that one site had successfully built GEMPAK
under Cygwin on Windows, but that he (the user, that is) had to make
some tweaks so that LessTif would work.

>To
>complicate matters, many of our students don't feel compelled to use
>garp for weather forecasting but rather, immediately depart for MS
>Windows PC's and the abundant supply of data on the WWW.

This is also what we are hearing from other sites.  It seems amazing
to me that this would be the case given how easy it is to generate
displays in GARP/NMAP.

>This,
>perhaps, indicates a level of immaturity on the part of students who
>aren't able to effectively use the rich tools in the garp/gempak
>environment, but nevertheless shows the attraction and ease-of-use of
>the web in an MS Windows environment.

I agree totally.

>In order for us to effectively plan our hardware acquisition, it would
>be helpful for us to have some input from Unidata on the
>direction/future plans for Unidata software.  What is the future of
>garp/gempak?

GEMPAK and NMAP appear to be here for the long term especially in light
of the acceptance of NMAP as a base line AWIPS application.  We have
heard this comment both from Mary DesJardins (Chiz heard this) and
from the directory of FSP, Sandy MacDonald.

>Will the 8-bit graphics deficiency ever be addressed?

While we can't promise that it will, it seems like it must for NMAP to
be able to run on the same workstation as AWIPS.  We don't know for
sure yet what approach is being taken in the NMAP development group, so
they may be looking at needing to run a second X server on Linux: 8-bit
for NMAP and 24-bit for AWIPS.  This setup, by the way apparently works
well enough for more than one site to have adopted it.  We, on the
other hand, have not tried it out, so we can't comment on specifics.

>Garp would appear to be eliminated from our bag of tools if not.  If we
>can't use garp, what will we replace it with?

NMAP.

>What is the current status of the MetApps project?

It is proceeding nicely.

>Will this be a garp replacement?

Our intention has always been to create a suite of applications in
Java that will be attactive enough to make users _want_ to migrate
to their use.  The things being developed in Metapps right now
provide capabilities never before available in Unidata-supported
applications, things like 3D rendering of NEXRAD Level II data
and remote access to data (via ADDE and DODS).

>What would the time frame be?

This is hard to nail down exactly.  We are planning on making initial
Metapps offerings available by the end of next year.  I am not
the best person to speak for Metapps, so I don't want to mislead you
as to the timeframe for its availability.

>If not MetApps, what alternatives do we have
>if we wish to run in the Linux/Windows environment?

See above.  We are also going to be looking to see if it is feasible to
distribute FX-Linux.  Toward this end, we will be getting an FX-Linux
workstation setup here in time for the Policy and User Committee
meetings in October.  What we really want to find out is if FX-Linux
has the kind of capabilities that are desired by the Unidata community,
and if it can be supported within the current resource environment.
This investigation is being done in part because of responses we
have gotten from the User and Site surveys that we have been encouraging
users to fill out.

>I could go on with
>questions but I'm sure you get my drift... Can you give some guidance?

Please let me know if the above does not answer the questions you posed.

Tom Yoksas