Re: SGML Text Embedded in Image Files

Frank Dunn (mailto:frank@BRAZEN.DEMON.CO.UK)
Sat, 11 Feb 1995 11:06:16 +0000

Message-Id: <mailto:199502111118.FAA24551@library.wustl.edu>
Date:         Sat, 11 Feb 1995 11:06:16 +0000
From: Frank Dunn <mailto:frank@BRAZEN.DEMON.CO.UK>
Subject:      Re: SGML Text Embedded in Image Files
To: Multiple recipients of list IMAGELIB

>I have several questions about the embedding technology.

My experience is limited to textual data with news images and their transmission. Its also conditioned by the pre dominance of the Mac in this field.

The standard tagging used is IPTC (International Press & Telecommunications Council), which defines a series of records that describe the image(s) in terms of the colour space, binary type, input device etcetc. The whole bundle inc. image is an envelope. The earlier IPTC standard you will see in use as basic caption data/text as part of the images bit map, invariably cropped off any published wire pictures.

>1) Is there a utility that will do it for TIFF files?

PhotoShop up to version 2.5.1 had a caption capability. As of 3.0 it has full IPTC descriptors.

>2) Does JPEG compression of a TIFF file destroy the embedded text?

Not if its a Mac OS file, as the text is in the resource fork of the file. I'm informed that there is an equivalent de facto standard in the PC world.

>3) Does conversion of, say, a TIFF to a GIF destroy the embedded text?

As 2 above.

>4) Does AIIM have any preferences or strong feelings about this technology?
>Is it a standard practice that has a reasonable chance of some longevity, or
>an extension that will not survive the millenium?

The wonderful thing about standards is that you can always choose which one you want to use ;-)

IPTC take up in the UK has been variable. Not all the major wire providers have moved to it yet, one failing is the lack of *any* baseline mandatory fields in the - to me - critical record that has keywords etcetc. Some users still haven't moved away from analogue picture transmission, IPTC is digital.

Reuters for instance have implemented parts of IPTC, which we then parse out of the image file and drop into our image database record. It works. OTOH we 've had major problems with AP files at the reception side, which is now back in their court to fix.

PhotoCD doesn't support IPTC, which is puzzling to me as Kodak have a vested interest - market share - in publishing. My guess is that an SGML based standard that addresses cross platform issues and the real world use of images might see wide usage assuming it broke out into the commercial world.

>I would welcome comments to the list or to me personally from anyone who has
>knowledge or experience with the issue.
>
>
>Bob Rosenberg + Supposing is good,
>Thomas Edison Papers + but finding out
>Rutgers University + is better.
>New Brunswick, NJ 08903 + --Twain
> mailto:rarosenb@gandalf.rutgers.edu

mailto:Frank@brazen.demon.co.uk News International Newspapers Ltd., mailto:Frank_Dunn@delphi.com 1 Virginia Street,London,UK, E1 9BD. CompuServe:100012,23 +44 171 782 6384/6 +44 171 782 5213 fax Library Manager (Images). "Ars longa, vita brevis"