Language Game

Here's an idea for a computer or console game intended to make learning a language interesting.

Imagine a fantasy-world game where the foreign language is used to cast magic spells. A microphone is needed to record what the user says. The game would teach the player to say phrases or sentences to activate spells. Introductory phrases such as those relating to time, location, or acquiring things lend themselves to this concept. For example, learning to ask about location or directions might conjure up a map, or saying "I am from ..." might return the user to that location. Other sentences might be used to activate objects. A user might have a magic bag that only opens when the user says "My name is ...". Basic written text would serve as the "magic language" to teach new spells at higher levels, or serve as in-game puzzles. Sentences or short poems could also be used for "magical combat". For instance players might freeze an opponent by saying "The weather today calls for snow and wind."

In games, the concept of "leveling up" is common, where the player gains more abilities over time as their character progresses (e.g. learning new skills, spells, etc.). This would work well with a language based game because the player would naturally learn more over time, leading to more sophisticated spells and puzzle challenges.

P2P-based web search engine

Web search engines are currently the domain of large corporations like Yahoo, Microsoft, and Google. However, one can imagine an open-source, distributed, P2P based search engine. Traditional P2P systems like Gnutella use distributed searching to locate files, and the web is nothing more than a collection of HTML files. The P2P-based web search engine envisioned here would take the functions that would typically be running on Yahoo, Google, etc. servers and distribute them to clients running on end-user systems. You might think of this as Gnutella for web content.

An ideal client for a distributed web search engine might be one of the existing browsers like Firefox. In that case, a new ranking system could be used based on the number of times a specific page is actually visited (e.g. as opposed to the number of times a given paged is linked to, as used by Google). This information would, of course, need to be anonymized.

Privacy screens

Portable devices (PDAs, phones) are widely used for business. An obvious side-effect is that business is increasingly conducted in public places, rather than just in the office. Especially on public transportation, privacy may be a growing concern. One solution is protective film that makes device screens impossible to read except for someone looking straight onto the display. Alternatively, privacy screens may be built into the device during manufacturing. Note that 3M is a manufacturer of privacy screens for PCs.

Parts that tell you when they are failing

Maybe five or six years ago I a radiator hose blow-out, just beyond the top of the "grapevine" pass on California's interstate 5. Apparently the radiator hose had been weakening over time and eventually split down the middle. A few years after that, the same thing happened again, just driving around town. I started to wonder why we don't have car parts that notify you when they start to fail.

For example, the government is working on clothing for military personnel that can identify when they are torn/pierced due to a fine mesh of wires or fiber optics. Some kind of simple version of this, built into parts that frequently fail (fan belts, radiator hoses), would save a lot of grief for just a few more cents of cost per part. The system could use passive RF for communication (ala RFID) with the car so no wires would be necessary. If the belt/hose expanded and broke the wire mesh, the part would report back possible failure to the car when queried. The mechanic would be able to identify the failing part, back in the shop, using the same mechanism.

Possible problems include ensuring the failure warning system itself does not become a source of false positives. Also, possible privacy concerns from RFID, but I think there are ways to solve this using proximity and/or generic part IDs rather than globally unique IDs. Cost of the monitoring system in the car would need to be low.