Hi,
I was on spring break when this email came and I have a note for
the Users Committee. I hesitate to email it to the usercomm address if
the meeting is over, if this comment is still appropiate please forward it
on or if it isn't please forward it to the bit bucket :)
RE: GEMPAK -- 8 bits per pixel limitation
I have been actively using GEMPAK for various projects at Iowa
State. I have been very happy with the support provided, so I hesitate to
complain, since I know that staff is probably allready taxed with other
things. My major complaint is with the 8bpp limitation of GEMPAK software
and the rise of low cost LINUX workstations.
I currently am involved with numerous weather data projects at
Iowa State and the National Weather Service in Johnston, Iowa. Up until
recently, ISU and the NWS have been using SGI workstations and HP stations
which only do 8bpp, this is makes using GEMPAK very convient. BUT now
things are changing, the old kludgy hardware are being replaced by high
performance LINUX workstations. These machines also include high end
graphics cards that have 24 bpp displays, now running GEMPAK is a
nightmare.
One solution, that does not pratically work, is to run two X
displays with one at 8 bpp and the other at 24 bpp and educate the users
to switch displays, but unfortunately, LINUX does not handle this
smoothly, at least in my experience. If there is a known solution, I
would be elated to know.
So I shrug and read the support archives at
UNIDATA. Unfortunately, when anybody raises the color problems, the
common answer is, "You need to run 8bpp, not run a window manager like
GNOME and not have netscape open." I wish that I would see, "You need to
run at 8bpp, we are currently working on fixing this 'problem.'" But I
have not seen anything like this and it is really depressing. I am not a
programmer, so I am not sure of the complexities involved in making GEMPAK
work with more colors.
I get local support emails all the time from students and NWS
people asking about the color problems and my response is, "Please use the
old and comparitively slow SGIs." (shrug) So the user usually just goes
to the web and does not bother.
Recently at the NWS, I have seen FSL D2D working on the LINUX
boxes and at 24 bpp, so I am optimistic that maybe I can install it at
Iowa State (I do not know if license problems will come up and I have not
looked into adapting our IDD feed for D2D use). If the 8bpp problem
continues to exist, I will investigate D2D further.
Sincerely,
Daryl Herzmann
Senior in Meteorology
Iowa State
On Wed, 14 Mar 2001, Linda Miller wrote:
~|All,
~|
~|There will be a Unidata Users Committee meeting in Boulder on 15-16 March.
If ~|you have issues or questions you want the Committee to address, please
send a ~|note to: <address@hidden>
~|
~|You can see the agenda items at:
~|
~|<http://www.unidata.ucar.edu/community/committees/users/agenda.01.03.html>
~|
~|Thanks,
~|Linda
~|
~|Linda Miller - address@hidden
~|External Liaison, Unidata
~|University Corporation for Atmospheric Research
~|P.O. Box 3000
~|Boulder, CO 80307-3000
~|303 497-8646 fax: 303-497-8690
~|URL: http://www.unidata.ucar.edu/staff/lmiller/un.act.html
~|
