I'd love to have something on my phone that would tell me where the next bus was on a cool Google Map (or other) and an estimated time to arrival (which could be calculated by Google Maps). Of course, you'd have to get your transit authority to give you access to their GPS results.

But this would probably be one of the most helpful things a transit authority could do. The problem most people have with public transit is that it can be inconvenient, but often it is convenient if you have better/more complete knowledge of it. For example, when I get up in the morning, I have no idea whether a bus went by a minute ago and it'll be another 20+ minutes waiting or whether a bus is 2 minutes away.

This could also help trip planning. Often in a mass transit system, there are multiple routes to get anywhere. The trip planners we have today are primitive in that they often don't take current information into account (such as speed at various times of day, location of the next bus/trolly/subway train in proximity, etc.). Using GPS information would greatly change that.

I love this service. Very useful.

It's a trial, there are most (all?) trams and a few buses.

In Tampere on major stops we have information about when the next bus is coming in, some based on GPS and some on schedules, this distinction is indicated on the board.

Just sayin'

Something like tiava.com?

Of course, you'd have to find a way of guarding against malicious entries.

The system is operated in conjunction with the police. Stolen systems are required to have a police report number.

Orbicule's Undercover coordinates with police, built-in phone-home software, a Mac's built-in camera, etc. Nifty stuff.

This means that if anybody opens a stolen laptop in the vicinity of one of these APs it would be detected without requiring that the laptop be associated to the local network.

I think this is a really good idea.

I don't just want to not see ads, I want something with some visceral action, something to make the ads not want to come back! -- ie, something fun.

Using your phone to reduce carbon emissions.

Anyway, I know this may seem like common sense, but if someone was going to build an app like this, I would suggest that the best platform for it is the iphone simply because Apple actively market their technology as more "green" than other manufacturers.

* I hate patents.

(I've noted that the covers on the Harry Potter books in Britain seem to be drawn for intelligent people, whereas the books in the US are far more cartoony.)

Although, more seriously, I doubt it would be accepted. In order to be patented, the idea must be novel, useful, and not obvious. It meets the requirements for useful, but not-obvious and novel may be harder to establish. (Depending on any patents for single-serving condiment packaging, they may be phrased in general enough terms to be considered prior art)

Which would make it pretty tough to get into bars.

(Perhaps such an app exists; I'm too far away from an iTunes-capable computer right now to check.)

I was in a (non-life-threatening) accident last week and I can think of a couple of useful first-aid pointers lots of people overlook, like:

* so what is the local emergency number? (based on GPS position) * "let's give him something to drink to comfort him ..." (Don't !)

Basically a decision tree to help people with basic first aid. These protocols exist in the first aid world.

I like the idea of a decision tree. Big readable decision screens: "Does the person answer when spoken to? [YES NO]" etc.

Google e-IMCI to see some actual research in this area, maybe you'll be inspired, though that is focused on more formal procedures for nurses (like detecting what's wrong with a sick child).

All of these languages have a powerful reflection/meta programming facility (yes, Perl included). So technically, something like could be implemented:

1) The server (using Twisted in Python or POE in Perl) listens for requests. Requests are for objects.

2) The server looks at the request, serializes it into a s-exp describing the state of the object sent to it.

3) Emacs then looks at the s-exp and decides what to do with the object initially looked at (suggest a method to complete/use on? highlights it red if it's wrong? bring up HTML formated documentation, etc...)

I'm not entirely sure, but I think you might just have to implement a Swank server, and SLIME* would work without problems. It could be that you're forced to use s-expressions, but I doubt it.

Getting Perl code into an s-exp would be fairly difficult actually (as opposed to getting an s-exp from a Perl complex hash -- which would be really easily) - generally (with Perl5) the consensus is anything that can parse Perl, is Perl. Python, Ruby and Perl6 (which effectively is a Lisp, complete with macros in forms of first-order primitives for CFGs) would, I am sure, be easier a lot easier.

Yeah, probably.

1. A screen scraping/computer vision toolkit that would supply most of the AI primitives to interpret games 2. An interpreter for each game that would provide an machine usable API to interact with the game 3. An AI program that would then play the game

Component 1 would only need to be built once, Component 2 would need to be built once per game and then Component 3 is what the kids would get to build.

I always thought this would be an excellent, gentle introduction into the world of programming.

Take Fishy (http://www.xgenstudios.com/game.php?keyword=fishy) for example. It would be really simple to build a dumb AI that could do ok but to build a really great AI would involve some really sophisticated path planning and optimization algorithms.

What better way to get kids inspired and wanting to learn about programming than have them solve the games they're already playing?

This wouldn't really be necessary in NYC where cabs are ubiquitous but in DC it would be useful.

EDIT: Crud, already sorta exists: http://taximagic.com/

There's enough structure in blogs (blogrolls, tags, date, author, backlinks) --- especially if you mine comments --- to do take a stab at this.

Would there be any way to make that discussion-like structure accept input from anonymous, infrequent, or lazy users? Setting up a blog just to reply might leave out some decent responses.

The original vision of the "blogosphere" rejected comments. "Write a blog post, instead." Here's an old, well-known example:

http://www.markbernstein.org/Apr0401/Whyblogcommentsarebad.h...

The problem right now is that the affordances favor comments over posts. The first step is to adjust them; it just so happens that by making a threaded blog-group reader, you're also providing a lot of value to people who don't care that comments are evil, but do find it hard to follow 10 different blogs talking about the same thing.

I'd really like to eradicate blog comments.

Maybe the mobile could upload the photos to an online album.

:-) Just Saying.

Vicarious parenthood is cheaper, cleaner, and you can take a break from it whenever you want.

But nieces and nephews have a higher payoff than Facebook.

Right now bugmenot is pretty much useless, since all accounts are dead almost right away.

This way the passwords etc will remain private, and the accounts won't go dead right away

Honestly, I think the user side portion is a bigger problem. Its all the psychology, people change the login info so that noone else would be able to do the same thing and prevent their access.

I still see no benefit to the users changing the account info. This is to access sites with free accounts, yes? In that case, isn't it actually just as much work to assert control over multiple bugmenot accounts as it is to sign up for the free accounts in the first place?

I'll say it again: Whatever counter claim you think you have, you can not hide login info from the people you are logging into.

The main inconvenience I have with my shopping list on paper is that the things I forget to buy are not automatically on next week's shopping list. And I might forget less things if I could easily delete the items I buy while I'm in the supermarket. This sounds like a small easy-to-write application.

The application should also have the advantages of my paper shopping list. My wife should be able to add things to the list using her phone. Adding items to the list should be extremely easy. I hate writing text messages.

That said, I think there's an android app I saw recently on the market that does this. I'll install it later and see how closely it compares to the features mentioned.

However, I wouldn't be surprised if there is money to be made with an application that helps users their spend money. You could try to sell it to a phone company, or a supermarket. You could let supermarkets advertise discounted products. You could make people pay for the application, or for 100 updates to the list. If users update the list per sms, maybe you can share the revenue with the phone company.

I suppose the biggest problem is not the monetization but the distribution. This is solved for the iphone, and for android phones, but for other phones it looks like a nightmare.

Another problem is to make easy enough to use that it actually makes life more simple. I find typing on a phone cumbersome. Maybe using a dictionary of the items you have bought previously would make it easy enough. It could be nice to be able to take a picture of a product's barcode to add the item in the list. On the other hand every added functionality diminishes the ease of use.

You would set a means of money exchange and the kids could publish what they are earning money for to draw the adults into picking them.

(but who would voluntarily install that...)

Deaf people of course ! :-)

Ideally it would be able to track not only what applications you are using or URL are looking at, but also which files you are working with. Then you can tag entire directories in order to keep track of projects.

Pretty reports is a must

Bonus points if you provide programmatic hooks to automate filling of damned time tracking, ticket trackig and TPS tracking 3rd party solutions.

Look at something like Coda (www.panic.com) and think instead of having all the apps built in to the app you just have a tabbed window container that you add app windows to that allows you to quickly switch from one to the other (eg. tabs for text editor,terminal,version control gui,ftp client, pdf viewer)

http://wintabber.com/screenshots/

(Windows only)

This would be complete with futures trading and derivatives.

(Lots of other details required to make it work. Email me if you'd like to hear more)

http://fold.it/portal/

http://galaxyzoo.org/

http://noisetube.net/

http://qcn.stanford.edu/

http://stardustathome.ssl.berkeley.edu/

Not only is it really hard come up with a model to convert currency from many places into a "standardized" format (or even to allow for OpenID-style distribution) - it'll also not be easy to provide data to other sites that allows them to do useful things easily. Tags might be as good as it gets.

One other thing Y-people might not like: this SHOULD NOT have a revenue model, I think. Freemium completely screws you and ads on a reputation site would just feel wrong. Besides, it's a platform.

That's why I keep thinking about open-sourcing what I have, especially in order to try to solve the distribution problem before an alpha site might go online. Being a hopeless idealist I really want to see a service-decoupled reputation service that is NOT facebook/google FC.

Please let me know if any of you'd be interested to join me in playing around with this.

High karma could move your mail towards top of the pile for people that don't know you (VIP).

Or posts/mails could be filtered based on karma thresholds.

[edit]: the same can be done for articles/persons etc etc. for more info http://www.gnu.org/software/emacs/manual//html_node/gnus/Sum....

http://www.seriosity.com/attent.html

The side effect of the market would be to rank discussion systems by credibility, or signal/noise, in real time.

It's also geeky enough that I think you could keep liquidity, even though there's no dollar value to karma.

or is that too eugenic?

If I get to it this weekend, I'll submit to HN when I have something ready. Thanks for mentioning uclassify -- I hadn't heard of it.

ive pared down my list of "want to dos" to a single choice, but i definitely want to use uclassify. if i hadn't buckled down, i'd have a half-working django prototype in ~/src/ right now that i'd revisit to burn some time on every 3 months.

For example, suppose you didn't know what the little plastic thing on the end of your shoelace was called and you wanted to buy more of them. You could search for "shoelace" or "end of shoelace" and find a picture of an aglet. Maybe even some ads to where you could buy a supply of aglets. Now, no more frayed shoelace ends for you, and life is a little better. :)

http://www.zimbra.com/products/collaboration.html

http://fornicatur.com - a new way to meet people for well...

Though tech could be white labeled for any topic - beyond theme of domain above (GPS locate hiking buddies in central park, people who like your favorite movie, tons of stuff... interesting new way to meet people).

http://www.loopt.com/ ?

To make time flow slower, make a watch where hours last more than 60 minutes. That way you will accomplish more in each hour.

I prefer the second solution. Productivity is more important than hours worked.

For instance - I need an entrance ramp blocked so no more cars get on this highway - is there someone nearby who can park their car to block it and wave people around?

Or - how many people are stranded on the other side of this flooded road - can someone nearby give me an estimate?

There should be some well-known way to subscribe to the feed, and it should be able to recognize your physical location.

There are lots of other reasons why a teacher might set specific rules for an exercise such as this, too. Unfortunately you chose to ignore the instructions -- and you made the mistake of second-guessing him/her -- assuming that the teacher's goal was the end result and not the process of getting there.

The half bakery has some good idea, some mediocre ideas, and some very entertaining bad ones.

Already submitted here it seems but evidently didn't get voted up at the time - http://news.ycombinator.com/item?id=135677

The wipes would be warmed by the baby's body heat as well; babies don't like a cold wipe on their tushies.

It could be implemented as a flash llvm front-end with an objective C emitter. You'd also need to write a Flash VM in objective C, which I haven't looked seriously at but wouldn't be trivial.

Most importantly, a system like this would allow you write any application UI in Flash. A C developer, say, could then leverage Adobe's interface builders and the flex library without the need for disgusting hacks.

I feel so embarrassed for our faculty. I mean they have classes on UI design and usability, yet it's beyond them to make their own pages usable. This is the home page of our logic programming class: http://www.cs.uta.fi/~tn/LOGO_kotisivu_2009.htm I wouldn't mind the yellow background and the all-caps text, if every other course would just have the same information in the same place in the same format so that some sort of study planning could be done.

Be warned though, Blackboard basically has their whole system patented, though I think their patent for "using a CMS for school" was recently overturned. The patents may be less important outside of the USA.

Otherwise, Viso has a steeper learning curve, but probably has more tools for what you're doing. I know it's expensive to buy, but with a Bizspark membership (get sponsored and only pay $100 after 3 years) it's free for the downloading along with your MSDN membership.

And, since no one has described it here yet: aside from making pretty-pictures, the really neat thing about Graphviz is that it is declarative--you describe the nodes in your graph (which can include text, shape, color, etc.), specify the connections and voila--Graphviz does the rest.

Oh, and you can use it programatically too--there are libraries for many languages, so you could, e.g., have a user submit genealogy data and then spit back a pretty family-tree diagram.

Google around and you'll find various interesting things, e.g. this recent (12/08) article on using Graphviz + PHP to generate Oracle schema diagrams: http://oracle.phpmagazine.net/2008/12/drawing_er_diagrams_fo...

Porn games for the Wii.

It works on the Wii, btw (website in Opera).

1. Upload a ton of music to serve as seed data

2. Run a beat rate algorithm

3. Mix the vocals from one song with the rest from the other.

4. Present clips of two such mixes to users, who pick the better one.

5. Repeat until you have a best selling album of remixes.

if someone changes their avatar picture, email address, name (marriage), location, etc., they should be able to do it on one site and have it pushed to all of the other sites that user is a member of.

authentication would be a big problem, but you're smart, figure it out.

I want to be able to disable screen updates except for manually triggered refreshes of certain areas, or to block out areas I don't want to refresh until instructed otherwise. I want extremely low bandwidth low detail options that fill in like progressive JPG instead of line-by-line from the top. I want to be able to cancel data transfer and reset to nothing. I want the remote client to screen-scrape and send compressed text and outlines of windows and to be able to send it high level Windows support instructions like "open the control panel" and have it not refresh the screen until that's done.

I can orient myself in a Windows desktop with a small amount of information, but I can't stand hideously slow remote control connections.

have a mechanical turk type group calling people over voip...

customers submit polls online, charge a few hundred $, promise really quick turnaround, have 'turks' make the calls for less than you charge.

http://www.mr-lee-catcam.de/index.htm

2. An aggregator of free lectures (video/audio/text) on the internet

The Internet has provided a medium for new growth in the bartering exchange industry. This growth prompts the following reminder: Barter exchanges are required to file Form 1099-B for all transactions unless certain exceptions are met.

See: http://www.irs.gov/taxtopics/tc420.html

Could have priority markings and due dates integrated if it were a customized task list instead of google or RTM, and pop up the interstitial less often or less obtrusively when there's no important deadlines approaching, etc.

really last.fm should do this.

Edit: Cool name :P

Delicious Library might do what you want?

But isn't that kind of like a YC startup? I forgot the name though but I remember the website being a place where you could follow people's comments on blogs. Is your idea focused on forums?

I like the idea.

I saw a link here a while ago about encoding redundant copies of data with less duplication than you'd think...

Example: I take my car in for a squealing noise it makes. Obviously, I'm not going to pay $3000 for a belt replacement, but I have no assurance that the "Miscalibrated Magnetostat" isn't the correct diagnosis. I'm willing to pay $3000 for a magnetostat recalibration, and the going rate may well be $3k, but all the mechanic really does is replace a belt.

Though, saying it out loud, perhaps I should just implement some kind of rating system for mechanics as well.

This way I don't have to sync and stuff ;) does instapaper already do this? I tried it once before.. I forgot the reason why I got back to ruBooks and a custom python script to download websites from my delicious feed.

It just seems to me that music is the perfect social product...

(I hope I'm not posting too many ideas)

Imagine the kids running around the house taking shots of worms, spiders, ants, etc. and sharing them online.

http://news.google.com/news?q=aging+population+health