Potential Projects

June 28, 2003-

Take these ideas. Please.

These are projects I'd like to work on but dispair of finding the time for them. I would really appreciate hearing about similar projects already started.

• Virtual email folders.

  A piece of email doesn't just fit into one "folder", and MUAs are doing us a disservice in assuming they do. A piece of mail should be able to belong to many folders. These folders must be instantly searchable and browsable. I would like to automatically "tag" email according to any number of criteria, such as "if message participants include any of these mailing lists, it's tagged as 'perl-list'". It should also flag particular authors, subject-lines or text keywords, and allow me to add additional tags to the message at any time. Outgoing replies to incoming messages could be tagged with all of the tags of the message I was replying to. Within my MUA, I could select all messages matching one or more tags, ideally much more quickly than grepping 30MB of raw email for certain keywords, finding a list of folders, then performing the same search within each folder. An example MUA query could be something like "find all non-work email about Boston and humor newer than 2000." (Try grepping for that... even with 'formail' to split the messages, it's still an unweildy mess.) Tags could also be easily added and removed at will, for particular projects or more complicated searches. The "tags" could be kept in a database with proper instructions to find the original messages, wehther they are in a local file or remote IMAP folder.

• Email-sorting help.

  My email directory currently contains 30MB+ of partly useful and partly crufty data. I'm lazy about sorting incoming mail. I want suggestions for where it should be filed. What incoming mail is periodic and therefore deserves a "folder" of its own? What mail in my inbox have I not replied to for over a week, is from a friend's addresses but not via a discussion list, and so seem to have been overlooked replying? Together with the above database, tag certain mail as "deep-storage" and let me weed it out once a year.

• Regional Business Reviews database.

  Searchable, browsable. fun to read like "Mechanic X-Files" section of cartalk.cars.com. Able to support setting up Review Clubs of like-minded individuals, such as Restauraunt Review clubs or auto clubs. Browse by club, business type and subtype, neighborhood, reviewer, and/or rating.

• Mailing lists for "thinking globally, acting locally".

  Provide web/email infrastructure to make study circles for interested persons to work on social or political change. They would plug into existing national/international groups to encourage and germinate "cells" of local groups to allow people to meet for coffee, have in-person book discussions, show solidarity at local events of interest, etcetera. At the same time, the network would have national/international support for information exchange, book reviews, seeding new local groups, etc. This would be quite useful for national/international mailing lists that want to set up local groups, but lack the infrastructure, which is largely replicated between groups anyway. I hope this already exists, but if it doesn't, it should.

• "Handshaking" protocol to determine common interests.

  I read about a japanese toy that will signal users if somebody who has the same "favorite color" comes within range. Well beyond "favorite color" or "compatible zodiac sign", PDAs (or instant messengers) should have a protocol for interacting and deciding what was "common or interesting" between two people, and selectively disclose ONLY the common elements, such as favorite music, sports teams, cities to vacation; or the "interesting" elements that were chosen by both as interesting, such as "home city". There would need to be security to prevent mining other peoples' data, such as progressive disclosure of bits of information that matches. The biggest problems I see: it isn't clear how it should react if people have different comfort-levels for disclosure, or different selectivity toward the idea of "favorite". Also, there's the potential for sliding further down the slope of confusing networking and honest-to-god friendship. On the other hand, part of me thinks it would be funny as a heads up display for your car windshield. Über-customized bumperstickers!

• Fuzzy GPS

  ... or commercial application of cellular tracking. I am not sure if this is a good idea or a terrible idea. Some phones have the capability of tracking which cell tower they are closest to. Assuming 'always-on' internet, one could send that tower ID to a server (via SMS or email) which would map it to the physical location. This could allow numerous services which depend on general location. Including: rough global positioning if you're REALLY lost; a service to send broadcast messages to other phone-users within a general area ("Did anybody loose a wallet?"), or tracking somebody's location (who, hopefully, has chosen to make it public). Like I said, it might be a bad idea.

• Life Rewind.

  Do you ever hear a conversation/presentation and wish you could rewind 30 seconds to catch what somebody said? Or think, "Wow, I wish I had gotten that down" just after the fact? If you carried a digital recorder with a two-minute loop, you could. The loop would constantly erase the beginning of the recording as it records at the end. You'd press "rewind 30 seconds" and hear what you missed; and possibly dump it to long-term storage. This might be another one of those terrible ideas- do you want to "always be recording live?"

• Personal audio recorder.

  Fun with a PC radio card. It would be relatively simple to record radio to MP3. Very simple "personal audio recorder" computer application to grab scheduled shows, particular DJs, etc. Because not every show has free online recordings like "Car Talk" and "This American Life" do.

• What if this were paired with "Life Rewind"; your computer could always be taping your favorite radio-station with a one-hour loop. When you're in the car, you might hear something you really wanted to write down or save. If you can get home in an hour (or if you're set up properly, to any networked computer), you can tell it to save the last hour.

• Link authoritativeness.

  Google has released an API for programmed searches. This has a number of possibilities. What if you wanted to track and display the authority of the links away from your site? You could easily show the world that your off-site links are useful by displaying little graphics next to each of the links. A database service could handle daily google-queries on a search-term, then searching the results for a particular URL or URL fragment. That ranking would be the URL's authorativeness on the search-term. The ranking could translate to a small graphic color code, such as "blue" for low-rank and "green" for high-rank. A site-owner would then reference this service to display the generated image for each search-term and URL; and also let their users find other similar results by clicking on the graphic.

• Video driving directions service.

  There are often situations where a map or written directions just don't cut it. Large cities often have poor or missing highway signs, especially in areas under constant construction. Many directions rely on complex cues, such as what lane you should be in, and what unmarked exit to take. But these routes are successfully taken by many thousands of people daily. Could you take advantage of their experience? What you need is more or less a visual walk-through of the tricky parts. I'd like to see a video directions service, possibly offered through AAA or On-Star. Subscribers/buyers would download a video clip with video of the difficult parts of the trip. They would watch it before the trip, taking 10 minutes to learn the route as if they had been there before, and potentially saving hours of confusion. The neat part of the idea- you could partner with this service if you can demonstrate a brand-new or updated route. You'd be mailed a web-cam and tiny storage device, which you'd stick to the dash. When you're done, mail back the storage device, and get paid for essentially taking your normal commute. And, presumably, keep the web-cam for future route updates.