Monday 9/9

The report is mostly done, except for the parts I'll add on after trying the new FM and image database stuff.

I think the main problem is that skandha connects by default to the old server. Whether it's always connected to the old one, or just connects there whenever there are problems, that would explain why we don't get descendants back for most of the structures. (They wouldn't have any descendants in the 'generic part' hierarchy, on the old server.) The only thing that doesn't fit is that we do get some descendants for Aorta... and I can't imagine how we could, if my explanation is right.

To do:

1. Talk to Jim and JJ about suspicions with the server.
2. Plan presentation for Wednesday, based on the report.

---

Tuesday 9/10

After talking to JJ and Jim about the server, and trying to change things in skandha with Jim for most of the afternoon, we decided this is best left for Kevin Hinshaw when he comes back. Instead, I'll work on a better integration for the FME, based on ideas that Rati and I had recently.

For the presentation: some important issues that I haven't spoken about yet.

• structure lists (reasons, in adding, removing, mouse clicks, saving/loading) - show in applet
• list generation for JSP
• new integration design for JSP
• show image manager function
• future improvements for everything

To do:

1. Redo FME integration.
2. Get ready for presentation.

---

Wednesday 9/11

I made several changes to the FME integration, so I can show the general idea at the meeting, if not all the details. Things that still need fixing are:

• image thumbnails (zoomable), and link to search
• navigational connections from images and scenes
• default scenes- cumulative

I also got a first draft of the report ready for comment, and sent it to the three professors.

To do:

1. Finish FME integration.

---

Thursday 9/12

The integration pretty much works, now.

Things to note in local documentation:

• images are slow the first time b/c they're being retrieved; later they're just being printed, but can still be kind of slow
• a limited # pictures would be faster
• doesn't currently have annotation clicking thing
• what structures work to test various things
• structure list might not be necessary
• for new FM: first letter upper case
• images might be causing out-of-memory errors sometimes
• README files for each program, with locations, class descriptions, how to run
• FM-dependent stuff (and how to change it and test it)

To do:

1. Revise report.
2. Local documentation and archive folder.
3. Go through stuff in detail with Todd.
4. Finish webpage and send zip to DMP.

---

Friday 9/13

Everything's pretty much finished. I'll be in touch with the professors in the future, because they'll be using my report as a resource for writing papers about these applications. But for now I'm done! :)