The Hacker News Experiment | Hacker News Reader

The Hacker News Experiment

hackernewsexperiment.blogspot.com

189 pointsroschdal52 comments

Neat. Even better if there was a Hacker News Startup Challenge, similar to Rails Rumble, where instead of mere 48-hours, you work for 1-2 months and launch, then you get reviewed by a panel of judges. This would be much more interesting to me, I guess.

It would be awesome if the winner would get a free ticket to fly to the Valley and join the current YCombinator class for one of their famous dinners, and get to present.

Wishful thinking, maybe.

That's a really cool idea. I'm not sure how many people would be available to participate though. Rails Rumble works partly in that no one expects anything that great from a weekend and a lot more people can make a weekend worth of time.

If I was starting a startup right now it'd be silly for me to wait a month for the "Startup Rumble" and I wouldn't want to cheat and get a headstart.

This would be really interesting, especially for those of us who are struggling to meet that special someone (i.e. a co-founder). For me at least, I think being part of a competition and having external interest to keep motivation up would be helpful.

Exactly, "put yourself in a position where failure will be public and humiliating".

http://www.paulgraham.com/die.html

meet that special someone (i.e. a co-founder)

Two words for you:

1) the co-founders meetup http://www.meetup.com/Co-Founders-Wanted-Meetup/

2) instead of lamenting about a situation, do something about it http://blog.fairsoftware.net/2010/06/10/will-you-be-a-great-... "entrepreneurs are the ones who do when everyone else is talking"

Literally, if you don't like something, fix it. Just do it yourself. Don't cry on HN that something needs to be done. To me, that's a tell-tale sign of entrepreneurs.

"Don't cry on HN that something needs to be done."

I read nothing in that comment that sounded like crying.

As to your first suggestion about the Co-Founder-Wanted-Meetup: Am I such a cheapskate that $65 in advance/ $125 at the door to attend a meetup sounds a bit steep? Perhaps I can't see the forest for the trees and the opportunity cost of missing a meetup like this is way greater than the attendance fee. So for anyone who has attended this meetup, is it worth the fee?

Sorry, you confused the regular co-founders meetup (price between $6 and $10 to get in), with the founder conference, a once-a-year event. Different size, different price.

The crying part: I was trying to anticipate the reaction of "but there is no such meetup where I live". Just do it.

i just thought about this and wah here you've posted it.

why not make a spreadsheet with each one's name and start date (may vary depending on our schedules)?

We can log our names and our app urls whenever it's done and we'll collaboratively set a final date. We'll then have people review our submissions.

Agree with the time period. 1-2 months is great. 48 days? Just to match railsrumble's 48hrs :)

The spreadsheet can be used to log our progress too. So if anyone finishes fast and and some other app is on the same track offering data which another app needs, then they can collaborate and release their data via APIs etc. Would be a great hackmonth :)

There's nothing stopping you from working for 1-2 months, launching, and then getting reviewed by a panel of customers.

Heck, if you win that, you'll basically get a free ticket to fly to the Valley and join the current YCombinator class, permanently.

At the review part right now. That's absolutely right. Why bother doing projects like this when there's nothing stopping you from playing the game for real.

There's nothing more valuable than going from A -> D (open for sales) and then realizing you have no fucking idea how to get to E. Excuse my language.

We did this. In fact, the first one ends in about 14 hours. It's been great so far!

http://59daysofcode.com

How about this? Build a general purpose "backend" interface that allows anybody trying to commercialize a web product to manage their users.

Your system should allow:

* the creation of superadmin accounts

* Map your user model to my user model

* CRUD users of my application

* Impersonate users of my application so I can see their data

* Expire / suspend / upgrade / downgrade users

* Search for users

* Notify a user via email of something

* Report top users by metric (in the mapping operation above we tell you where to look up these metrics)

* Report new signups per day

There are various Rails plugins that take care of most of this stuff. I've had great success with restful_authentication (http://github.com/technoweenie/restful-authentication)

You should check out AuthLogic. restful-auth's generator based approach is awkward sometimes, and it hasn't been updated in forever...

Devise has been my choice in recent times. but beware it has no roles. it's simple and bakes in everything though. so in 2 commands in the terminal you'll see a login system in your browser including forgot password feature that authlogic doesn't give you.

Making all of those features Samparicio said would be a little tough for code reuse. Engines feature was planned for Rails 3.0 beta but pushed to 3.x. That's the feature that can help code reuse. But it's possible to do this neatly on Django, which allows component-based architecture for apps.

Authlogic has no roles either, I use acl9.

Authlogic and forgotten passwords is pretty simple, though. It sort of gives it to you. Well, it gives you the tools. It's really only a tiny bit of coding.

I'll certainly check out Devise though.

+5, and make it plug and play with Spring/Hibernate.

I was already building something like this and your comment made me push the private beta :P.

http://news.ycombinator.com/item?id=1451758

Not sure if it covers all your bases, but check out Raible's AppFuse - http://appfuse.org

How about a 4chan style message board in the sense that posts expire very rapidly, but also in that only single topics are allowed for certain time intervals. For example the site could start off with a 24 hour cycle in that dialog can take place on a certain topic (such as BP oil spill) based on the first message that has come in that day. It'll need some means of aggressively pruning out (marking as spam) messages that are deemed off topic, but this can be a voluntary (karma based) or a machine learning approach that could also be tweaked or plugged in beyond the 4 week development period. The initial version would just keep topics on lock down for an additional 24 hour period, before permanent deletion. So basically there is one active topic, and one previous topic for review.

This isn't strictly a web application, but if you are a Java programmer then one project I'd recommend would be integrating the Twitter Search API with FreeMind for the purpose of facilitating real time curation of tweets. Real time curation is apparently the number one request for software that the twitter people get, and FreeMind is the ideal platform for curating tweets. All that's needed is a way to intelligently get tweets into FreeMind so that they can be quickly sorted appropriate buckets, and since FreeMind is written in Java it might be a good fit in terms of skill set and project size.

To give an example of how this would be used, the project described above would reduce the time of making this from about seven hours to about an hour: http://www.squidoo.com/Hacking_Education

Also, here are two relevant blog posts:

http://scobleizer.com/2010/03/27/the-seven-needs-of-real-tim...

http://shannonclark.wordpress.com/2010/02/19/curate-for-me-a...

Wouldn't this require Twitter to give you direct access to the firehose?

No, there's a streaming API [0] that allows for access to a subset of the firehose in JSON via search and filtering. I wrote a short ruby script that is a good demo of how it works [1].

0 http://apiwiki.twitter.com/Streaming-API-Documentation

1 http://gist.github.com/325757

What do you mean by FreeMind - the mind mapping software?

As someone who's been skimming "The Four Steps to the Epiphany", I note that your schedule is purely Product Development - there's no Customer Development. Since YC is about "building something people want", that second part is equally crucial. Arguably more so.

My take on "4 steps" is: do iterative/agile/RAD/MythicalManMonth development (you present a demo ASAP, so the client can tell if that's what they want or not), but to not wait til you have a demo. You iterate on "the customer problem you're solving" and "specs of the product". This is much faster than building a demo, which is faster again than building the whole thing - and only then finding out that the "people want" part is missing. (You build the product concurrently).

What you're proposing is an engineering project - which is great! Nothing wrong with that, challenging, lots of fun, lots of learning. Just noting it's focussed entirely on that half of an entrepreneurial project - a common developer approach.

Are you building a business or a piece of software? It sounds like you're building a piece of software which is totally cool - I just want some clarity before I contribute.

I first and foremost want to solve a real problem for someone. To do this I think that I should develop a web application which solves this problem. If solving this problem also requires a business, then that is something that I would consider. Andreas

Here is what I would like to see.

The Freecycle Network is brilliant - lots of stuff gets rehomed that would otherwise get dumped in landfills. People get their redundant stuff taken away, people get free stuff (computers, toasters, clothes...), it's good for the environment, it's all good. Technically speaking, it's in the dark ages though. It runs via mailing lists, which are operated by Yahoo and apply to discrete geographical areas. It works well enough, but I think it could be improved on, or at least supplemented.

I would like to see a system based around open standards. I envisage:

* A structured way of listing spare stuff - ie. hListing or some variant.

* A way of managing a network of people whose stuff you're interested in - Portable Contacts? Whatever the state of the art is, not sure. Ideally something that can be somewhat integrated with people's existing social network info. Or, screw it, just a list would do for now.

* A standard way of managing who can see your listings - OAuth?

* A web application that you can use to manage your listings (or in time any number of applications, since the end result is standards-based output)

* A web application that aggregates, filters, searches and displays other people's listings (these last two bullet points are probably functions of the same application)

This is a bit sketchy but If it's feasible, I think such a system has a number of advantages over Freecycle or the various balkanized web-based databases of free stuff. I vaguely intend to look into doing this myself at some point, but I'm hoping if I pimp the idea about a bit somebody else will do it for me (or with me).

I think it was suggested by someone else here on HN but the idea is to have a widget on blog that tracks how many pages a visitor has browsed. Then simply show him a message saying 'Hey, you seem to be a loyal reader. This is your 3rd visit on the blog and you have read 33 blog posts here. Writing unique content takes time and effort, so please consider leaving a small tip to encourage me to write even more.'

Also show some sort of leader board of donators and the content they consume. Idea is to come up with guilt-ware which presents a new revenue stream for dedicated bloggers :)

Regarding the use of Java, the language is dying. Maybe the languages developer base, and supporters refuse to believe it, but on the commercial end most of the major Java supporting software shops have wound down or died. Even sun is closing up development. So, if you are developing actively in Java you run the very real risk of developing a ton of unmaintainable code.

It's like the quiet little joke the industry refuses to address. Erstwhile, Google is now the only people doing any legitimate active development on a new vm... and even they aren't chaining themselves to Java.

So I ask you, do you really want your software written in a dying language? I wouldn't. Java apps were hard enough to support when the many and myriad vms were being actively worked on. I can't imagine the pain that will be brought down upon adopters now. =/

Really? - http://www.tiobe.com/index.php/content/paperinfo/tpci/index....

Web-based MMS message composer. Initial features:

0) Subject and Text 1) add image (gif,jpg,png) 2) add support for more than 1 image 3) add resizing tool for images 4) add support for other filetypes (no preview necessary - just add as attachments) 5) Add support for page breaks (separate pages) 6) Simple design choices (text and image together on page or on separate pages etc) 7) More advanced design choices (move content between pages)

The output should be zipfile (one file per image, one file for text, one file for design - SMIL language). If you want to be advanced then you can also output Nokia MMS file format.

I have written some C#-code for reading Nokia MMS file format if you need that.

I can coach and give feedback if necessary. I am working in the mobile technology industry as a software developer.

Interesting suggestion. What real problem would this solve for people? Sending MMS messages is expensive at the moment, so making it cheaper would proably make it more usable for people. Anything more?

Here's one idea: Take Open Text Summarizer (http://libots.sourceforge.net/) and turn it into a web application which takes in an URL and returns the same document summarized.

I love craigslist as an apartment finding tool. Most of the ads for apartments include a google maps link, so I can figure out if it's really in the area I'm interested in.

But every time I move I wish constantly for a service that would let me search craigslist by actual location. My mental picture is a location and distance, but some kind of region-selection on a google-maps overlay would be even better.. or a location and a maximum travel time. A feed of those ads as they are made would be even more powerful.

Thanks for all the feedback and ideas so far, everyone. I would still like to challenge you to share more real problems that I could develop a web application to solve. Andreas

To people here, the ideas themselves are probably more valuable than the effort required to realize them. It might be a good idea to reach out to communities that have good ideas, but not the technical means to implement them. Having said that, check out:

http://ycombinator.com/ideas.html http://openideas.eu/ http://www.reddit.com/r/SomebodyMakeThis/

I am reluctant to give you any of my ideas, but how about a web app for images, that particularly serves the needs of e.g. law enforcement, and allows storage and annotation related to crime scenes, cases, etc, or perhaps for the health industry, and does the same for patients with certain conditions, and you could tag them and other doctors could offer second opinions for money etc.

Interesting. I am open if you are considering collaboration which will give one additional man month to finish the project. I am familiar with C, C++ and Python.

Hi! This could be possible. Please sende me an email with info about yourself. Contact info in profile. Andreas

I want inline previews (first paragraph) or the full text for every news, so that I can read the stories without leaving HN. It can be a GreaseMonkey script that adds a "show full text" button, and, when you click on it, shows the full text.

Not sure, what is the legal status of such previews, though...

Update: I have publised a summary of the progress so far on my blog:

http://hackernewsexperiment.blogspot.com/2010/06/status-from...

I saw you're planning on doing something with public transit. You should check out Google's Transit Program, they have public feeds of transit info (stops, routes, stop times and what not) here: http://code.google.com/p/googletransitdatafeed/wiki/PublicFe...

Experiment Track Thyself: Make an experiment tracking database.

Create a micro book summary site. I have the entire thing planned out and will realistically never get to it as much as I think I could. It's mostly standard CRUD, with some bells and whistles, you can monetize it pretty well, and I know of a good domain for you if you want to do it.

I genuinely want you to take my idea, please contact me and I will give you everything I have.

I have been thinking about this for a while. Not the app, but the fact that book summaries would be really useful. Most books tend to be too verbose. Care to elaborate a little on your idea for the app?

Basically take stackoverflow and make it for book summaries, but put a word or character limit to the summaries. So eventually, you'll have a collection of the best 500 word book summaries

How about an open-source hacker news? Thereby destroying the same community that helped you to create it.

What? He said "the problem must not have been solved before"

The Hacker News software IS open source.

Not all of it

"Thereby destroying the same community that helped you to create it."

Downvoted because it's silly to think that an open source implementation (regardless of whether there already was one or not) would destroy HN. As if it hasn't been done a million times before. It's the community, not the technology.

But what if the (community of) hackers can actually modify the system during runtime, using same techniques as voting stories up/down to write the code?

I've already feature-frozen v0.7, but I'm going to look into this as a plugin for 0.8 - potentially very cool idea.

Reply on news.ycombinator.com