Post a possibly good app idea that you have no intention of doing yourself.

123 points by bemmu ↗ HN
At school we often had to draw class diagrams, basically boxes containing some text, with arrows pointing to other boxes. The free tool we were using for this was made with Java, and would often be very buggy and was generally hated. I couldn't really find anything I liked and ended up making the diagrams in Gimp, not fun. I was really pining for a small perhaps Flash-based web tool that would let me make diagrams and download as PDF.

255 comments

[ 3.1 ms ] story [ 307 ms ] thread
This is great, got to email the teachers about this.
Is it free forever, or free to try? When I need to diagram quickly, I use UMLet. But somehow I suspect that UMLet is the one he's complaining about.
No, Graphviz is truly free--it's open source, under the Common Public License Version 1.0.

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...

(comment deleted)
I would love something that aggregated all your karma points into an exchange of some kind. I would use it, but I would never build it.
What kind of 'karma exchange'?
A full-on reputational economy.
what about acting as gatekeeper for invite-only or beta sites, prioritizing invite distribution by a global karma factor?

or is that too eugenic?

That's fine, but I think invite distribution is too unimportant.
i guess, as w/ most things in life, it just depends on the context.
Email prioritizing based on karma?

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.

That's a perfectly good application of a reputational economy. There's pretty much infinite possibility. I'm not sure if it has been explored in fiction better than Doctorow's Down and Out Magic Kingdom, about a post-scarcity, post-singularity society with a reputational economy. http://craphound.com/down/ Available free online under a CC license.
gnus email/news-reader already provides, such a fature via scoring. You can score somebody up/down. At a configured threshold, you dont get to see the posts/articles from such folks anymore. love to have something like this in gmail.

[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....

How many Reddit points will a Hacker News point be worth?
Wait, I'm an idiot; of course you run a market to solve this.
This is smart but I'm not sure what the fundamental transaction on such a market is. Keep in mind that there is no rivalry for karma (in an economic sense) and that people always keep their karma after giving to others (or possibly even increase it).
Not sure I agree; I think people are already using ad hoc methods to "price" karma; for instance, I have used such methods to "price" a Hacker News karma point as 1,000,000 reddit points.

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.

I guess someone will have to build the karma derivatives trading platform...
Oh man... with 30:1 leverage I could make a killing.
Allow people on your aggregator site to manage the karma of the reputation sources. Open to gaming, but all karma is to some extent.
rapleaf.com is vaguely in this space.
I have been coding a prototype of this in my free time over the past couple of weeks. It's fun, but I'm not sure it actually makes lots of sense.

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.

I just don't see how it is technically feasible, as most sites don't provide a way to transfer karma? Or is the idea more that of an independent karma service that other sites could plug in to (like gravatar does for profile pictures)?
Something that takes blog posts from a related cluster of blogs and presents them as a threaded newsgroup.
Sounds interesting, could you expand on that?
he probably wants news.google.com but in a format like mail.google.com's threads. kind of like techmeme, i imagine.
I want blogs to look like Usenet, so that best affordance for a response is another blog post, not a comment. Comments are evil.

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

I'm interested to hear why you think comments are evil.

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.

They're second-class outlets. A blog author can say whatever they want; their posts are first-class elements in Google's corpus; the medium encourages stand-alone writing instead of meaningless squibs.

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 love that idea. I'd use it.
Would you help build it?

I'd really like to eradicate blog comments.

Possibly, but I'm not sure I've got the bandwidth. Drop me an email and let me know what you have in mind (dave@paperstack.com).
I've always wanted to see this in a Sphere-like way - I want to read HN's comments on random articles on BBC, for example, while I'm on BBC News. Or Read Reddit and Digg's comments in tabs beneath the original source article. I'd also like to be able to 'subscribe' to which communities' comments I'd be able to read.
fwiw, perfectlygoodideas.com is available right now. perhaps someone could use the uclassify api to classify & cluster similar ideas & recommend partnerships. no idea how well it would work, but it might be worth a an hour or two on a weekend to find out. im just thinking out loud here :)
I really like that idea. domain name taken :) -- i presume you didn't want to do this since you didn't buy it yourself.

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.

awesome! put it to good use :)

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.

There's something already out there similar to this, at ShouldExist.org (currently down for redesign.) Might be some duplication of effort, or an opportunity to strike while the iron is hot -- depending on your point of view.
Ha. We were required to use Poseidon for UML, but it was such a pain and crashed so often that we ended up making all our UML Diagrams in Dia (which is a nice enough and lightweight program, if slightly clunky; you might want to check it out). Yet instead of being pleased or even just ambivalent about the entire enterprise, our teacher got mad that we used a tool other than the prescribed one. I never really understood why people would get pissy about what tools you used instead of focusing on the actual output.
Maybe your teacher's goal was to encourage people in the class to learn Poseidon so he could hire them on a private project. Or maybe he/she was looking for a genius who would be able to figure out how to code Poseidon in such a way that it would not crash.

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.

N-party barter settling platform.
I think currency.com is already taken.
Relevant point, but money was created to abstract before we had general-purpose abstracting machines like computers with their associated tools. If it wouldn't get you in trouble with the IRS (big "if"), I think it'd be fascinating to see the new tool applied to the old problem. It might create an entirely new kind of market.
Bartering doesn't get you in trouble with the IRS, as long as any income you earn from it (be that a cow, an hour-long massage, or sixteen clay beads) gets declared on your tax return at "fair market value". And then, obviously, you pay income tax on it. The IRS will accept just about anything to satisfy your tax obligations, as long as "anything" is US dollars.

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

If people like the new bartering system enough, trying to enforce such onerous reporting requirements will incite a tax revolt.
An app that tracks your forum activity and displays all your private messages, thread subscriptions, etc. all in one app.
Kind of like a mint.com of internet activities? I like the idea. I'm not sure how easy it would be to be exhaustive. I can see how you could add the regular places you hang out at. Casual comments would be harder to track but a bookmarklet could help there.

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.

Maybe you are referring to Disqus.
A big problem I have is I want to post a single question to a forum and be on my way. I hate having to sign up. Same deal but worse with mailing lists. What if your app had login accounts to the top say 10,000 forums so people could pop in and post questions some how?
Some forum software accepts openID now. I agree completely, at this point I have so many forum accounts on random sites where I just needed to ask one question pertaining to some program or something. The other day I needed to ask a question on some forum and was about to make a new account when I found that I already had made an account. Now I've decided to only use openID for new forums if possible and just not join others unless I think I'll participate a lot.
Something that produces iPhone code out of Flex/Flash code and vice-versa? That sounds impossible (I don't know either one) but if it were, it could be cash cow, no?
There's a library called CoreAnimation that's similar in some ways to Flash, supporting keyframes and tweening and such. The iPhone also has (an unofficial) port of Rhino, to run the needed ActionScript. (Compiling that to native code would be a whole 'nother matter.)
I'd extend processing. There's already ports for java, javascript, flash, iphone.
This is actually something I've considered working on. And no, it's not impossible!

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.

Room service robot for hotels. It could basically be a motorized cart. It would be easy for it to get around in big hotels with elevators.
Might it be easier to just build a dedicated sort of line for each room? Like a part of the wall that opens up and is connected to the kitchen/wherever, to deliver your food. Would be harder at first but it'd be much easier down the line, and more efficient.
hotels rent space. your suggestion limits space. do you think that the advantage of robot server compared to humans would offset the huge decrease in available space to rent?
It is WAY cheaper to use a human to deliver the orders.
I actually worked in a hospital where they paid for robots to deliver medical records to doctors/nurses. The nurse would request records, and then the medical records department would fill it up and send them out. I think that didnt last too long, not efficient enough.
Video games for dogs?
Yes! Clearly: a harness with a slot for a Wii-mote. Implement for cats also, branded "MWii-ow."
OMG a fake bone where you can insert a Wii-mote. Imagine the possibilities ;-)
On a related note:

Porn games for the Wii.

Hm, know of any interesting porn games? My own venture into erotic games wasn't a big hit: http://3boobs.de

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

i think my dog would like a game that required her to quickly decide whether a particular newly introduced item requires growling, barking, or wagging. is it the trash truck, the hated dogs next door, or mommy? the dog often gets this wrong, so practice would be good.
Sell sunscreen in little packets (like ketchup packets) that can fit in your wallet. That way you'll never forget sunscreen. Should I patent* this?

* I hate patents.

But what do you do when you have sunscreen on your burger and mayo on your back? I think this might backfire.
You could invent some sunscreen mayonaise.
That would make it more compatible to the American metabolism.
Or make it so we sweat sunscreen when we eat the mayonnaise (on fries?).
There's an offshoot idea actually. Could you make a pill or food that turns you sweat and or skin oil into natural sunscreen?
Mayonnaise on fries is British. It's probably related but we'd have to market it differently.

(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.)

Well, you could have a weird color packet. Last I check all mayo packets were white.
Well now you only have a year under 102(b)... This counts as public disclosure.

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)

I'm 100% certain that I'd find a way to make one of these explode when I sat down, covering my wallet, money and ID in a sticky mess.

Which would make it pretty tough to get into bars.

Individually wrapped sunscreen wipes are already in most pharmacies.
Awesome idea. Even more awesome that it already exists. Ordering some right now...
I don't think you can patent this...my school gives these out at any outdoor event. My friends and I usually collect a bunch and I keep them around -- in my purse, in my car, etc.
I actually carry sunscreen with me at all times. I use it maybe once a year, and I can't predict when.
That idea's been done, but if you could develop a packet of goo that could be used as both ketchup and sunscreen, then you'd have something. You could put half on your burger, and the other half on your face. Huge time saver.
Scan in ISBN of a book, have it posted for sale on Amazon and craigslist.
Amazon already does this. Actually, I think it's required to list an item on Amazon.
You need the ISBN, but it's not as simple as scan it and you're done.

Delicious Library might do what you want?

I have a hand scanner for barcodes that works fine with Amazon.com. It looks like a keyboard to my PC and so when I scan a barcode it simply 'types' the ISBN into the appropriate box on Amazon.
Delicious Library does it with the camera in the mac, you just hold up the book.
FF extension - Search within a textbox.

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

Filter or customize your tv ads. I hate it when a horror movie ad comes on and I have to scramble to change the channel. Similarly, I may like watching John&Kate+8, but that doesn't mean I want to watch the typically associated feminine hygiene commercials.
Also, how about not having to see the same ad three times during a given show? At this point maybe we're getting too close to tivo for the advertisers to feel comfortable.
How about at least filtering the online ads. Just put a little link in the corner of the ad that allows you to say "no thanks, I never want to see this ad again."
adpinion is doing this fairly well
a bugmenot type service...except it would act as a proxy. The proxy will be used to hide the actual login info, and to prevent the users from screwing around with the actual account.(basically I see this as changing the username to XXXX, not letting users go to pages like usercp.php etc).

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

I assume the accounts go dead mostly because the content providers kill them, not because individual users kill them. There isn't much you can do to stop a content provider from figuring out which accounts are bugmenot accounts and killing them... It's an arms race to see who will give up first.
more reason to keep the login info secret. Sure they can figure out the account by multiple ips...but that takes a little bit more effort.

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.

You can not hide login info from the people you are logging into. All they have to do is install your plugin and start logging in. Whatever clever algorithm you put in place to expire canceled accounts works in their favor too. Less than a day's worth of scripting and I can fully automate this within Firefox.

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.

In addition to Jerf's comment, I will also point out that you can't hide the login info from the content providers. All they have to do is steganographically encode the user name into the web page content. Then they can use your bugmenot-like service to retrieve the page, extract the user name, and cancel the account.

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?

A Taxi service site and iPhone app. Lets you send your current location to the cab company with your iPhone and then track the cab as it arrives. Maybe even offer a prepay option so you can pay with your paypal account and just read off a confirmation number to the cab driver or something at the end of the trip for them to bill you with. Maybe in order to prevent confusion over more than one cab being called from the same location the app can light your phone screen up in a unique color for you to call the cab with as it approaches.

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/

Why are you disappointed that it exists? My reaction is more like "great, that problem is solved, now I can make something else".
I just thought I was original but it turned out I wasn't.
Have this idea/started web design, but busy with work. It's a crazy out there possibly good idea or bad.

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).

I've always thought about a system for teaching kids about programming by building AI players for online games. The system would consist of 3 components:

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?

A site dedicated to connecting Technical Business Analysts with Software Developers.
What kind of connection, like what purpose and what environment?

Edit: Cool name :P

A cradle that updates the baby's facebook page based on decibels, activity, dampness.
That sounds more appropriate for twitter...
Hmmm, what about a carousel/mobile that sits above the crib. The mobile has a camera in it that takes daily pictures of the baby's face so that the parents can see their child's progression.

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

Shouldn't the parent pay more attention to the baby than to facebook?

:-) Just Saying.

He's not saying that, he's saying that you should be living his parenthood vicariously through Facebook. I agree: you should.
+1

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.

My sister and brothers would disagree, since I pester them to babysit.
I've always wondered why social networks don't ally with music labels and promote/sell music. Small networks can contact indie labels, and huge networks can work with sony, etc ...

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

Isn't this basically what Myspace is doing?
Yeah but I think it has taken them a long time to realize. What I am getting at is a music advertising/sales marketplace where "any" social network could sign up and promote/sell music. This would be another possible solution to monetizing social networks. So if I maintained a 100k high school football network, There would be an easy way to integrate music ads. And when I say music I mean new acts in general, tours, specific shows, and of course actual product releases/merch sales. As I said, music is very social and everyone likes some kind of music.
(comment deleted)
An app that increases the number of hours a day has, e.g. makes time flow slower. That way I could accomplish more ;)
People in other industries solve this sort of thing already. It sounds like some sort of drug would help accomplish this, but I have concerns about the health repercussions.
So far the closest thing to this would be to hire someone else.
To increase the number of hours a day, make a watch with hours that last less than 60 minutes. This way you will work more hours in a day.

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.