Slides, and image databases in general

31 Mar 98
I started thinking about the overall structure of the image part of the project, and got some ideas for the components which lead toward a functional spec for the things we need to build and link. I think I'll try to summarize the thoughts of the moment, with the intention to amend and refine.

We want to provide Larry with an electronic archive of his images, which is

The archive might be stored on CD-ROM (the drawback to that being that it's not all that easy to add images) or might be on the M: drive; if the database was on the M: drive it could (I think) be queried by a perl search utility, and the output passed (as thumbnail images) to the forms-based management/page creation tool.

The database is probably most easily created and managed with Excel, and I presume that perl can either deal directly with Excel files or that delimited ascii files can be produced pretty easily. What the fields should be is mostly up to Larry: what criteria might he want to search on in the process of managing and using the image archive? Species? Phylum? Season? Common name? Date? Source?

It may be that there is existing software for managing slide collections --we should at least inquire into its features, file formats etc., though I would expect Excel to be capable of all the functionality we'd want.
The [presumably forms-based] perl utility to extract and display [thumbnail] images and permit selection for display on web pages: this (1) searches the database according to various criteria (probably from some sort of menu), (2) uses the ID numbers of images to retrieve and display thumbnails, and (3) can then pass the IDs of a subset of images to

a utility [presumably a form] that allows (a) arrangement (and sizing) of images on a web page and (b) entry of commentary and (c) linkage to other pages.

Just who gets to build these is an interesting question, but they pretty clearly have broad utility once they're built.
Some links:

Kudo Internet Multimedia Suite

Seybold --industrial high end (just to keep an eye on how the Big Boys think) and another

Query By Image Content (IBM/UCDavis collaboration --again, a different scale from our problem but worth knowing about)

Knoware Image Databases

Image Finder (Docuphot)

Here's what Larry says:

I'd probably categorize the slides by interaction subject matter, e.g., predator-prey, competition, communities, ecosystems, etc. Genus and species, or description of what is happening on a slide would be a good cross-reference idea, if it is possible.

There are no consistent ID numbers on the slides at present. These are from very recent back to 1972, so most are not even dated. There may be between 300-400 slides at present (they are spread out among perhaps twenty or so slide boxes right now. Actually, if I get going on this thing I will probably double that number fairly quickly.

I asked Jeff Knudson about the practicality of this sort of linkage of tab-delimited file with image database with selection panel with annotation panel with html output and he said [essentially] 'solved problem', no big deal.