Brad Ideas - Photo editors: Embed your text in the jpegs - Comments

Had not thought of that

brad — Tue, 06 Mar 2007 12:41:21 -0800

I had intended the function to be optional. I'm not a big fan of programs coming back to bite their users, even if their users might be evil. However, you are right that some spammers would forget not to include the text and thus their graphics would be more decodable. (Obviously images not intended for decoding like captchas would not have these tags.) I don't see this feature biting most users, however, other than perhaps if they don't want search engines to know what text is in their images. In theory the text would not reveal anything a human could not learn by just looking at the image, so it's not secret, but in practice certain types of very tiny or highly distorted text might be revealed which would not be visible to a human. Clever software could detect that but I think that's overkill.

I like the idea for this reason:

Solar John — Tue, 06 Mar 2007 07:54:28 -0800

Spam filters could then get rid of the junk, so it doesn't fill up my mailbox.

John

Photo editors: Embed your text in the jpegs

brad — Mon, 05 Mar 2007 14:18:23 -0800

Hey photo editing programs — I’m looking at you, Photoshop — a lot of you allow people to place text into graphic images, usually as a text layer. Most graphics with text on the web are made this way. Then we export the image as a jpeg or png/gif, flatting the layers so our artful text is displayed. This is how all the buttons with words are made, as well as the title banner graphics on most web sites.

So photo editors, when you render and flatten the layers, take the visible text (you know what it is) and include it in a tag inside the file, such as the EXIF information. Possibly as the caption if there isn’t already one. Let us disable this, including on just a single layer, but providing it would be a good default.

Then all the web spiders/search engines would be able to find that text. Web page editors could offer that text as a possible “alt” text for the graphic. And the blind would be able to have their web-page readers read to them the text embedded in graphics.