Unified Media Search

This is a simple idea that I've been formulating for some years:

What happens when people want to look back and find something? Interactions from people lost long ago?

Sure, each social media site allows you to go back and find things/people/whatevers you posted from a long time ago, but that doesn't help in the long run, because they come and go.

For this I have an idea, some kind of a background service that checks your accounts, automatically adds contacts to its file, and your correspondence, chronologically. All of this wouldn't just come from one source either, emails, texts, voicemails, and forms of communication we haven't even imagined yet, mixed together with a base platform with rich plug-in hooks to process the data.

It would be a tool for social engineers, scholars, detectives, and regular people tired of managing their digital lives across so many services.

The basic service would be fairly simple to build, initial extension points I see would be:

• Inputs/Parsers (create "messages" or "contacts" from files or services)
• Reports (visual ways of viewing/manipulating stored data and analyzers)
• Analyzers (Something that creates tentative links between things, could be by phone numbers, IP addresses, lat and longs, facial recognition, writing styles, etc.)

I suspect the initial data would be fairly simple to structure as well, each message would have:

• Date
• Sender(s)
• Receiver(s)
• Format
• Extracted text/Resources
• Original Message

Basic outputs I would like to see are:

• Who knows who? A map between people likely to know each other (good for looking up someone you don't remember)
• Conversation history, show all conversations between two people on a timeline, from all media sources.
• Happenings, mapping when people got together and where, creating "events" in a timeline.

Sleep Sort

A kind of neat sorting algorithm if, for some reason, you had a machine that had unlimited interrupts, and unlimited running time, but had almost no CPU power to do other sorts, perhaps biological computing?

1 2 3 4 5 6 7 8 9 10 11

#!/bin/bash function f() { sleep "$1" echo "$1" } while [ -n "$1" ] do f "$1" & shift done wait

Purely Pedantic Password Affirmation

PBKaC (Problem between keyboard and chair). Yes, people are the source of all problems in Computer Science, a computer does exactly what it is told to. But sometimes they can be the solution to problems too.

What if, when you entered a password, three things were sent back to the server?

1. A password hash.
2. A list of the times taken between keypress events for the password, hashed or something.
3. Some identifying information for the computer, plugins, whatever.

The server could check the headers sent back to see if this is a common computer used by the user, and if so check the password and move on with its day.

The server could also check the times between keypresses in the password for relatively spaced times in the way the user normally enters the password, providing a fingerprint for a particular user.

Suppose it normally takes me 30ms to reach from "F" to "T", and fifteen to get from "O" to "."; if the password was entered differently the user should be redirected to a secondary question page; either they broke a hand, or someone else entered their password.

Think of it as non-random password screening. It would stop bots in their tracks, and would create only a small problem for users.

Idea: Frensel Lens Foundry

While browsing around the interwebs, I came across the fact that Frensel lenses have a great ability to focus sunlight. Naturally, I was surprised to find that such a great invention is being so under-utilized. (kind of like deck prisms which have just been "discovered" recently)


Anyhow, it turns out that these powerful lenses can be found in projection televisions, and best of all, people give these things away on Craigslist, because the TVs are broken.

A rather devilish plan came to me: why not melt aluminum cans in to small ingots? Sure, the use of these ingots would be rather limited, but you never know!

A quick trip to YouTube confirmed that things like steel could be melted with a lens about the size of a television, it has a melting point about twice that of Aluminum.

And now comes the really cool part: maybe it could be automated, an Arduino to track the sun and move the lens, a chute to feed aluminum cans, and some kind of a pot to catch the molten metal.

I like the idea of dumping the liquid in to a muffin tin, ensuring the metal comes out properly, although it could simply be poured back in to another soda can with the lid removed.

If for nothing else, the ingots could be used to save space while you gathered enough cans to take them to the recycling center.