OVERVIEW: The quick version of this idea: make it so a user can automatically 'listen' to text content from the web by converting web pages into podcasts for him/her, based on criteria that the user specifies.
NOTE: The following describes this idea in the context of using a folksonomy bookmaring
application like "del.icio.us" it is assumed that the reader is familiar with the 'folksonomy bookmarking' concept that you see on sites like del.icio.us.
PROBLEM: Consdier this scenario. You are catching up on halfbakery posts, and you realize there are a lot of threads that seem interesting to you, but you don't have time to read them all. Since you are dilligent about not forgetting them, however, you post them to your del.icio.us account with the tags "halfbakery" and "_toread".
This works well because you can go back and read them later and they are all in one handy place.
The problem is, what happens when your list of "_toread" bookmarks gets too large? The typical outcome is they pile up, and eventually get ignored or forgotten.
SOLUTION: Text to speech is a well-baked technology, and so is folksonomy bookmarking. A great not-yet-baked way (AFAIK) to unite these technologies would be for a processing engine to go through all of your del.icio.us bookmarks with a specific tag (e.g. "_tospeech") and automatically convert (via periodic batch process) the text contained on those bookmarked pages into mp3 formatted text-to-speech files. This content could later be retrieved as a standard run-of-the-mill podcast.
This way, you would be able to 'listen' to all the content you bookmarked, but did not have time to sit down and read.
You could even use this kind of functionality as a substitute for sitting in front of your computer and reading from the screen.
ALTERNATIVES: This kind of functionality could easily be done by patching together some scripts with a text-to-speech engine that you program yourself on your own computer. The problem with that though is it would be nice if it could be done automatically on a central server. Not everyone knows how to write TTS scripts.
PITFALLS: This kind of operation would require a lot of server space and processing power for the conversion, ironically, this could be minimized by easily detecting whether a bookmarked page was already converted to mp3. This would obviate the need for duplicate conversions.
There are also copyright issues, but sites could 'opt out' of having their content convertable to speech, and most sites would welcome the additional ubiquity of having their content available through multiple channels.
There is also the technical drawback of the 'robotic' sounding voices that traditionally accompany text to speech. Some users may find this intolerable, but there are workarounds. (e.g., 'natural voices')
ADVANTAGES: Web users could optimize their time by only browsing graphics-intensive sites via the screen, and text-intensive sites via their personal podcasts.
Someone new to halfbakery could request a podcast of all the 'introductory' web pages for the site, so they would not have to sit down and read all the boring FAQs, but they would also be safeguarded from looking like a newbie because they at least listened to the 'introductory' podcast.
The power of tagging + text-to-speech + podcasting could open up a whole new queriable, portable interface to the entire text-based portion of the web.