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

19990514: What is ADDE?

>From: Jim Heimbach <address@hidden>
>Organization: UNCA
>Keywords: 199905141945.NAA27523 McIDAS-X ADDE

Jim,

>     Things are progressing with the installation of McIDAS-X on
>a linux box (running Slackware's flavor).  I am having trouble
>with squaring away f2c, but I think that will be solved shortly.

What exactly do you mean by "squaring away f2c"?  Is it not bundled
with Slackware Linux?

>My question is, what is ADDE?  I prowled around the e-mail archives
>and the web pages, and remain clueless.  From what little I find,
>it appears that this will be replacing McIDAS-X ?  Anyway,
>could you explain what ADDE is and if I should install it whatever
>it is?  

ADDE stands for Abstract Data Distribution Environment.  It is the
method by which McIDAS applications request data from one or more
servers.  ADDE _is_ McIDAS-X; it is not separate.

What is going on is that old McIDAS applications are being replaced
by newer ADDE applications.  This has been going on since the introduction
of McIDAS-X Version 7.0.  The Unidata distribution of McIDAS-X still
contains a number of the non-ADDE routines that have already been
sunsetted (i.e. support dropped for) by SSEC.

In the ADDE way of doing things, one requests data from data sets.
The end user does not have to remember the arcane naming convention
that McIDAS uses for files.  Here is a very quick comparison between
an old way of doing things and an ADDE way of doing things:

non-ADDE load of a GOES-West IR image centered on 30 N and 130 W:

DF 130 1 EC 30 130 -2 EU=IMAGE SF=YES

ADDE load of the same image which can be found in the dataset RTIMAGES/GW-IR:

IMGDISP RTIMAGES/GW-IR 1 LATLON=30 130 MAG=-2 EU=IMAGE SF=YES

So, you may ask how/why this paradigm shift is beneficial to the user:

o first, the user does not have to know that GOES-West IR imagery is stored
  AREA files 130-139 (for example); s/he only has to know that there is
  a dataset named RTIMAGES that has a member called GW-IR that contains
  GOES-West IR images.

o the ADDE invocation as listed above automatically chooses the latest
  of the set of GOES-West IR images; in the non-ADDE example, you would
  have to look at a listing of the images to find out which AREA file
  contained the latest image (or you would have to use ALOOP which does
  the work for you)

o the RTIMAGES dataset does _not_ have to exist on your machine or
  on any machine on your local area network

The last item is the big win item.  One can "point" at another user's
machine (that s/he is allowed to point at) and get data loaded from
the remote machine.  For instance,  I can sit here in Boulder and
"point" to a machine at UVa and load imagery:

DATALOC ADD RTIMAGES windfall.evsc.virginia.edu
DSINFO IMAGE RTIMAGES
IMGLIST RTIMAGES/GW-IR.ALL
IMGDISP RTIMAGES/GW-IR 1 LATLON=30 130 MAG=-2 EU=IMAGE SF=YES

The DATALOC command tells McIDAS that when a user tries to access the
RTIMAGES dataset it actually lives on a machine named
windfall.evsc.virginia.edu.  If that machine is running the ADDE Remote
Server (something that you should setup; the information is in the
online documentation), and if your machine is enabled to request
the data (raw remote ADDE installations prevent nobody from asking
for data), then when a request is made, the information comes from
the remote machine.

The DSINFO command asks the question of what data of type IMAGE are
included in the RTIMAGES dataset. From the output of DSINFO, you will
see what kinds of products are available.

If the remote machine is setup as per Unidata instructions, you will
find images of type GW-IR.  The IMGLIST command (the ADDE parallel
for the LA command) will then list information about the images in
the remote dataset.

The IMGDISP command then says to display the latest image in the RTIMAGES
dataset in frame #1.

The very cool thing about this is that you can get data from other sites
when you machine missed files.  This is what I call a distributed, ad hoc,
short term data archive facility.  Cooperating users can act as a safety
net for each other in a transparent manner.  At the present time, there
are over 15 sites running remote ADDE servers that one could get data
from.  Etiquette demands, of course, that sites request access to other
users machines _before_ "pointing" to them for data.

To get more of a flavor of the difference between ADDE routines and old,
non-ADDE routines, please look through our online Learning Guide:

http://www.unidata.ucar.edu/packages/mcidas/mclearn/learn_guide.html

By the way, we are holding a McIDAS training workshop at the end of July/
beginning of August :-)

Talk to you later...

Tom