Twitter picture transcription
Feb. 1st, 2015 12:44 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
marnanelPeople often post images of text on Twitter, either to get around the 140-character limit or because they're posting a screenshot from another site. This is a problem for people who use screen readers and people who have images turned off.
I propose to create a web service which will allow people to associate images with URLs of pic.twitter.com/XYZ with transcriptions. For example, you could associate http://pic.twitter.com/2EaLCXaszP with the text "The Bishop of Dibley".
This would then be available on the Twitter site via a Javascript snippet in the browser.
Byte-for-byte identical images would automatically share a transcription, detected by digest. There could be a tineye-style similarity test, but that would make a simple idea much more complicated.
Users would log in with their Twitter account details, via OAuth. All their transcriptions would have their account name attached.
The biggest problem is of people maliciously adding incorrect transcriptions. I invite suggestions.
What do you think?
I propose to create a web service which will allow people to associate images with URLs of pic.twitter.com/XYZ with transcriptions. For example, you could associate http://pic.twitter.com/2EaLCXaszP with the text "The Bishop of Dibley".
This would then be available on the Twitter site via a Javascript snippet in the browser.
Byte-for-byte identical images would automatically share a transcription, detected by digest. There could be a tineye-style similarity test, but that would make a simple idea much more complicated.
Users would log in with their Twitter account details, via OAuth. All their transcriptions would have their account name attached.
The biggest problem is of people maliciously adding incorrect transcriptions. I invite suggestions.
What do you think?
no subject
Date: 2015-02-01 05:29 pm (UTC)Who would use it, and how would they build it into their workflow?
no subject
Date: 2015-02-02 08:09 am (UTC)