< Fox and the Grapes
NYCB Readers Demand (or, "Whatever Mola Wants, Mola Gets") >

[Comments] (4) : I wrote a piece of software that is really neat. I think it's in the top quartile of the class of application that people could have turned into software companies back in 1999. I don't know anything about starting software companies, and it's not 1999, but I still think there's some way to make money off of it. Unfortunately, I have no idea how to distribute it even if I give up hopes of monetary gain. I'm hoping my readers can help.

My original plan was I would host it and let people have accounts on it, but as it turns out, it requires a lot of processing time to keep one user happy. (I have optimized this a lot but it's still pretty high.) I can't host it for more than ten or twenty people without buying a dedicated server, and a dedicated server will probably only let me go up to a couple hundred users. I don't want to enter the realm of running dedicated servers because that means the project has taken over my life, and unless it's my actual job to run the project I don't like that.

That's the one option. Now, this is a piece of software that sometimes reads other people's web pages. If somehow I got it all set up and everyone in the world were using it, it would use up about as much of your average web page author's bandwidth as, say, Technorati does. This I could live with. On the other hand, if I gave it out as open source and n people ran their own instances of it, it would be like creating n/k Technoratis for some fairly large k. If n got bigger than k I would have unleashed a monster upon the world. A monster! Usually you can count on nobody caring about the software you release, but I don't want to take that risk. The people I've shown this to really like it.

My current inclination is to just write a paper detailing all my secrets and making this someone else's problem, or at least making me look like a visionary when someone else comes up with the same idea. My ideal outcome would be selling the idea and the current implementation to someone, where I get some money up front and they make a lot more money by setting up servers and taking subscriptions and generally letting the project take over their life. I know that's not likely to happen. What's the next best option?

Filed under:

Comments:

Posted by Gary Benson at Wed Oct 20 2004 18:17

Aw, you can't leave us all in suspense like that. Spill the beans man!

Posted by cola at Thu Oct 21 2004 07:04

Aha! I think I know what it is.

Posted by cola at Thu Oct 21 2004 07:58

(I sent my guess to Leonard.)

I find performing the task that this object automates a real hassle. Therefore I heartily approve of this mystery object.

Posted by Yoz at Thu Oct 21 2004 10:30

I started answering this by suggesting that you gave the software out but made it check against a centralised registry to ensure that it wasn't already checking a URL that another install was. So it would go and grab from the install that was already pulling down that URL.
But, what if that install isn't reachable, because it's down or behind a firewall?
So I started thinking about how to build a big distributed cache.
But that's already getting silly.
And then I remembered that someone's done that for public use already:
http://www.scs.cs.nyu.edu/coral/
So, how about getting it to pull all its URLs from Coral instead?


[Main] [Edit]

Unless otherwise noted, all content licensed by Leonard Richardson
under a Creative Commons License.