Plans for Bob demo

Pedro Szekely ([email protected])
Fri, 21 Jul 1995 15:10:21 PDT

Everyone seems to be fired up about doing this demo, so let's go for it.
Pablo and I talked a bit about the steps we need to do to get from where we
are to the demo. Here is a plan of how I think we can proceed so that if we
don't get everything working, we will at least have something to show.

1- Pablo will integreate the Browser with the cpp version of the model server.
This will allow the model server to read an MTF file at startup, and pop up
the brower to look at the objects. This is DEMO 0.

2- Pablo will enhance the browser to be able to receive reports from the model
server to update the display when the attributes of objects change value.
This was working on a previous version of the browser about one month ago.
Since then Ewald did some changes, and some work needs to be done to get this
working again. THis will still be within the cpp version of the model server
and browser.

3- Pedro, Pablo and Kurt will figure out how to get a simple Corba version of
the server working. This simple version will consist of the model server and
a client that creates some objects and changes their attributes. This is
similar to the examples that Ewald was working on before he left. The work
basically involves getting Ewald's examples to work with our standard
directory structure and libraries. This should not be too hard.

4- We need to come up with an MTF file that we know will load correctly into
the model server. I think this MTF file should have a few objects from each
sub-model. I suggest using the example that I used in the Web pages. We need
to define a couple of tasks for it, and a one application object. Can Noi
define the tasks, Spencer define a small application model, and I will make
the presentation model. Kurt's job will be to make sure that the MTF loads
correctly. If we need to, we should write the MTF so that it loads, even if
it is not the MTF we really want (time is short).

5- Kurt and Pedro (mostly Kurt) can work on the attribute setting programs. We
should start simple, so I suggest just three or four programs here:
set-value object-name attribute-name value
add-value object-name attribute-name value
delete-value object-name attribute-name value
clear-values object-name attribute-name
If Kurt is comfortable doing something more ambitious in the time we have, it
is fine with me. Kurt is really the only one who has a whole week to work on
this. The other ones will be busy talking and drinking coffee Wed through
Fri:). These programs will be tested using the simple Corba version developed
in 3.

6- Once we do 1, 2, 3, 4 and 5, we make sure that the attribute setting
programs work together with the combined server/browser server.
This is DEMO 1. Kurt and Pablo.

7- After 6, try to split the browser and the server into two processes, so
that the browser is also a client. This would be very neat because we can
have multiple copies of the browser running on separate machines. This is
DEMO 2. Kurt and Pablo.

8- Graphical interface to set, add, delete and clear values. Would be nice to
have it, but I think it might be too much to try to do. This one is up to
Pablo. THis is DEMO 3.

9- Make the attribute setting programs a bit nicer so that they can take more
complex l-value specifications. This is DEMO 4.

----------------------------------------------------------------------------
This is a lot to do, but I think we can do at least DEMO 1, which is already
pretty neat. We have most of the pieces working. The challenge is to put
them together. Anything beyond DEMO 1 is a bonus.

I will touch base with Kurt as soon as I get to the office on Monday. One of
the first things for Kurt and I to work on is step 3. I will get started on
it right away, and let you know how it goes. This is fun!

Pedro Szekely
USC/ISI, 4676 Admiralty Way, Marina del Rey, CA 90292
Phone: 310/822-1511, Fax: 310/822-0751
http://www.isi.edu/isd/szekely.html, http://www.isi.edu/isd/Mastermind