[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

20001218: Unidata McIDAS access and update information (cont.)



>From: Leigh Orf <address@hidden>
>Organization: Department of Tesselating Kumquats
>Keywords: 200012121654.eBCGsDo24855 McIDAS-X tesselating Kumquats

Leigh,

re: who you got the McIDAS access info from
>You are right, I did get the password from someone here at UNCA (Jim
>Heimbach, who is actually the guy I replaced here).

Hmm... I thought that Jim had retired and left the area.  I guess that
he just retired :-)

re: download of 7.7

>That's me.

re: I have made a few small changes:
>Ugh !! I just rebuilt mcidas on four workstations! Aieeee! ;)

Sorry about that...  I was desperate to get a distribution out in FTP
land so that sites could do the install over the holidays.  The additions
after the original release date were not planned at first, but were
definitely needed.  ADDE support of the NOAAPORT NIDS products promises
to be a very useful thing, so I wanted to get all of the code mods and
speedups in there ASAP.

The only consolation is that the update process is very easy as you could
see by the instructions in my last email.

>Seriously, the past few days have been a crash course in getting past
>the hurdles of building and installing mcidas on Linux boxes and
>learning some of how mcidas actually works.

Did you have problems building the distribution on Linux?  Was it related
to UNCA using Slackware?  Did you make code modifications that make
the build proceed?  If so, can I get them so that I can incorporate
them in a future addendum (so that your life gets easier in the future)?

>I got my Ph.D. at UW-Madison
>and worked for SSEC for a while but rarely used mcidas... the last time
>I did it was on a cutting edge 486 running OS/2.

Smokin!

>Little did I know the
>the extent of the strange, sometimes baffling and horribly complex
>software that was being developed four floors below my office!

I can't disagree with your comments at all.  The good news is that I
am working with SSEC to make McIDAS easier and easier to use.

>Actually, I am pretty happy with mcidas now that it has the tcl
>interface. That is NICE.  I can open the console and see the complex
>commands being executed and learn more about them, modify them etc.

Super.  I am always glad when someone "gets it".  The reason for the
GUI console is so that people can see what the GUI is doing and learn
about what can be done.  BTW, MCGUI will be undergoing a total rewrite
in the coming couple of months so that it can take advantage of ADDE's
ability to access data from remote servers.  Stay tuned...

>This is good. I'm also already knee-deep in ldm and will soon plunge
>into gempak. As you can probably tell I'm the unix guy here in the
>atmospheric sci. dept. at UNCA and one one of the very few in the whole
>university.

I figured as much.

>I will certainly be bugging you here and there when I can't get over
>a hurdle. I am subscribed to mcidas-users and ldm-users and will
>participate in that, too.

I hope that is mcidas-x, not mcidas-users.  mcidas-x is the email group
at Unidata; mcidas-users is an email group at SSEC.

>I will do the mcupdate dance soon.

It should go smoothly, but if it doesn't, please don't hesitate to
ask questions.

>Here's one question for you (I'll leave it at one for now :). If I try
>to use the SSEC GUI interface (type GUI in the mcidas console) tcl dumps
>with a stack trace.

Is MCGUI running?  If not, can you send me the stack trace?

>It's the same stack trace on Linux and Solaris, so
>I'm thinking this is broken right now. Am I right in assuming this?

Probably, but I will have to see the stack trace to be sure.

>I'd like to see the SSEC interface even though I am happy with the Unidata
>interface.

I would recommend this.  The goal is for the GUI development efforts at
SSEC and Unidata to coalesce, so users won't have to make the choice
between the two.

>I can supply the text of the trace if you wish but I'm
>thinking you are probalby aware of this.

I was not aware of this, so I would appreciate the text of the trace.

>Just for grins I tried replacing the tcl and tk subdirectories in the
>mcidas7.7 distro with the latest (8.3.2 I think). It built fine but
>dumped core when I executed the MCGUI. I'm sure there are good reasons
>for why you still distribute 8.0.3, just kinda curious.

I have done the same thing and gotten the same results.  Finding and
fixing this (it is related to reparenting the McIDAS image window
in a Tk widget) is one of the things that I want to do for the GUI
upgrade(s).

>Anyway, thanks in advance for all your work on mcidas etc.... I can't
>wait to use this stuff in the classroom. Another point on the graph,
>I've been running mcidas-x on my Linux pentium 233 (128 MB RAM + 128 MB
>swap) at home with the /mcidas/data directory NFS mounted over a cable
>modem (16 kB/s upstream capped and not usually getting over 100 MB/s
>downstream over NFS)! It ain't fast, but it actually works acceptably
>well. And I can play with it in my home office, which is very nice.

I run McIDAS from home with no data being local. I did this for over
a year using a 56 Kb dialup and saw no problems (other than speed).
I have since upgraded to ADSL and have a smoking connection (1284 Kb
down; 322 Kb up).  McIDAS ADDE really is sweet: none of the data has
to be local!  As an example, try the following during the middle
of the day:

DATALOC ADD RTGINI ADDE.UCAR.EDU
DSINFO I RTGINI
IMGDISP RTGINI/GE1KVIS STA=AVL EU=IMAGE REFRESH='EG;MAP H'

(if you try this tonight, you will see nothing since it is after dark
there).

Also, for grins, give the following a shot:

<login as 'mcidas'>
cd data
ftp ftp.unidata.ucar.edu
  anonymous
  address@hidden
  cd pub/mcidas
  binary
  get AREA9000
  quit

DATALOC                   (note were RTGRIDS is to come from for later; if
                           the listing does not contain RTGRIDS, it is
                           implied to be LOCAL-DATA)
DATALOC ADD RTGRIDS ADDE.UCAR.EDU
CFILLCLP RTGRIDS/ETA T SFC 2000353 12 6 LAMB MYDATA/IMAGES.1234 Y Y


When all is done correctly, this will result in a color filled rendition
of the 6-hour forecast of DAY 2000353 12 Z Eta surface Temperature.  In
order for this to work, you would have had to FTP the global topography
file, AREA9000 and put it in a directory where it could be found by
McIDAS applications (~mcidas/data is a good place).

Don't forget to point McIDAS back to the original machine that RTGRIDS was
set to be read from:

DATALOC ADD RTGRIDS original_location_it_was_set_to

>Anyway, take care, have a great holiday.

You too.

Tom