AuditImageLib : ensure image folders fit policy

This particular procedure is not handled in a satisfactory way at present and it a prime candidate for whatever technical help is available for us. This is a nicely defineable project that might be quite independant of the USMS database project.

USMS Archives have many, many images. Approximately 800 have been provided by members, usually from meets, and approximately 2000 have been provided as a courtesy of Swim Magazine. A very high priority for the H&A Committee is to be able to store, display and annotate these images appropriately. Annotation is especially important because we do not own most of the images. Swim Magazine only buys the rights to use photos in their magazine and our receiving courtesy copies does not include the rights to use the photos. Where we do use them, we must ask permission from the photographers and we must give proper credit both to Swim Magazine and to the photographers. (See: Top Ten Photo Gallery Photographers.) Additional annotation is needed so we know who the swimmers are and when and where the photo was taken. There may be other information important, like the fact that a record was set.

We have 3 ways to display images at present:

• "packed"  (this was created by the following syntax: CreateImageIndex "c:\swimgold\pix\MelDyck\" )
• "stacked"  (this was created by the following syntax: CreateImageIndex "c:\swimgold\pix\MelDyck\" )
• "floating"  (this was created by the following syntax: FloatIMG "c:\swimgold\pix\MelDyck\" )

None of these ways is satisfactory. "Packed" conveys no information about the image. "Stacked" can convey information, but it is inefficient in it use of screen space. "Floating" has captions and is efficient in use of screen space, but we are told it cannot be viewed satisfactorily in Mac computers and with older Netscape browsers.

So, a new method of display is needed. We believe XHTML has promise for a better method of display. See: Why Web Builders Must Move to XHTML.

The annotation which we must record for each image includes the following:

1. file name (extension optional ?)
2. brief caption
3. descriptive text
4. photographer (or artist)
5. copyright owner to be given credit (if different from artist)
6. (other ?)

We have no convenient way to capture this information. We need a solution where images can be displayed on the web with provision for browsing people (with permission) can enter annotation into a form which will then be stored in a database. We anticipate volunteers with no particular technical ability will type in the captions from Swim Magazine which accompanied the photos. For images sent to us by members, they or a friend can type in the annotation.

We would also like to automate the process of maintaining the image library. Our vision for a proper structure is that images appropriate for important web display ("web images") are stored in folders. These images are typically 360 to 480 pixels in height and are likely to be 40,000 to 150,000 bytes in size. Each folder which stores web images has a subfolder called "t" which has thumbnails of each web image. Thumbnails are 96 pixels in height and range in size from 4000 to 12000 bytes in size.

Our automated maintenance procedure should ensure the following and either fix discrepancies or print a report of them.

• every folder should have an index file,
• every "web image" should be of appropriate size,
• every "web image" greater than a certain size should have a thumbnail of appropriate size,
• if a "web image" does not have a thumbnail, one should be automatically created,
• if a thumbnail exists which does not have a "web image", it should be erased,

We also have a policy for how to store source art. These images might be 100mb in size and are likely to be in .tif format. Obviously these are too big to put on the web, but for future use we do want to save and protect them.

At the present time (before conversion to an SQL database), captions can be stored either on a text file or in an Excel file (files are named imgcptns.txt or imgcptns.xls) or captions can be generated from the TTDB database. This lends itself to our ability to circulate CD's whereby H&A Committee members can annotate images on their home computers with the CD and then send the text file to the websteward by e-mail. But all current procedures for handling images are subject to review by whoever takes on this task. Click here for information about our effort to identify an appropriate SQL database.

©Copyright 1997-2002 USMS. All Rights Reserved