McIDAS-V Team Meeting - Wednesday, November 15, 2006
Present: Dee Wade, Jerry Robaidek, Scott Mindock, Jay Heinzelman, Matt
Lazzara, Rick Kohrs, Tom Rink, Tom Whittaker, Tom Achtor, Gail Dengel, Dave
Parker, Becky Schaffer, Kathy Strabala, Kevin Baggett, Barry Roth, Dave Santek,
Tommy Jasmin
Intro - Becky
This meeting is the first one where we're finally able to talk about what has
been happening with McIDAS-V, not just what we're planning on happening. There
have been meetings about the assigned tasks and some of the programmers have
been working on the tasks that we've talked about at past meetings. Today
Gail and DaveP will be giving demos and looking for feedback on what they've
done so far.
Development Meeting - What is McIDAS-V? - Becky
Becky talked about a development meeting that occurred, which consisted of
some of the PIs, some of the coordinating committee, and some of the programmers. The
goal of the meeting was to talk about the long term vision, and how to maintain
that vision. We all know that the slow down process has changed Mc-V
from a recapitulation of Mc-X to more of a PI driven project, but the other
thing that we talked about was the basic question of What Is McIDAS-V. We
keep saying that Mc-V will be Mc-X, IDV, HYDRA and other plug ins, but what
does that mean? Bob Knuteson came up with a graphic that
shows his concept of "What is McIDAS-V?" This graphic would have
to be modified to be a working model, but the basic concept is good. Bob's
conceptualization is that McIDAS-V is an island with bridges to McIDAS-X, Visualization,
Servers, Cluster Computing, and Databases. TomW likened this approach
to FireFox plug ins, where users can grab and use the ones you want. TomW
also mentioned that Unidata has a new SimpleIDV that is a stripped down version
of the IDV, which shows a bit of what is possible - we do not have to look
and feel like the current IDV package available from Unidata. MattL brought
up that this didn't seem that different from projects like AIFS or even SPILT,
which were not full packages, but rather glue that put together a number of
different applications. TomW responded that this is better, because
it has a better foundation. The IDV is built to exercise the capabilities
of the VisAD library, and that is the foundation on which we're building.
New Mc-X protocol - DaveP
DaveP gave a demonstration of a new communication protocol in McIDAS-X, which
will facilitate communication between McIDAS-X and other applications, like
McIDAS-V or a web browser. The web page shown included an image display,
a text output display, and a Mc-X command input box. In Mc-V, it could
be a data chooser that runs Mc-X commands and displays the resulting image. The
new protocol receives Mc-X commands and returns text and images only. It
provides a direct link between IDV and Mc-X, but does not require the user
to go back and forth between running commands in Mc-X and visualizing in Mc-V. From
the user perspective, everything is being done from Mc-V. Also,
it was mentioned that we are not throwing away previous bridge work - we are
still reading a pixel array and converting it to a data object. Not
memory mapped files. Still need SFU, but when using a web interface,
there will no longer be a need for an X server or tcl/tk. ScottM said
that this is a good solution for now, because it is evolutionary and not a
final solution. He also asked if this will have discovery
features and the ability for data computations. DaveP's answer
was "yes, hopefully".
Simplified Image Chooser - Gail
Gail showed her work on a simplified image chooser for quicker visualization. It
allows users to go through the normal image chooser, and then at the end, save
those selections for future use. Once saved, they show up under tabs
grouped by satellite name. Rick brought up that it would be nice to also
save everything about displaying the data, not just choosing the data. Then
when you picked your saved data chooser, it would know not only which data
to get, but also how to display it (projection, color table, map, etc.). It
was also noted that everyone always wants to tailor their own software to their
needs, and this is a great way to do this. Future work could also
include a more graphical way of choosing data, something like the current Mc-X
GUI.
Next Meeting - December 13th at 10:00 a.m. in Room 351