I'm honestly conflicted on this. On one hand, it's spam, as users are "receiving" messages they didn't solicit.
On the other hand, it's a kind of cool service, as an "enrichment" to twitter messages. If you could find some way of doing it for those who might find it useful without forcing it upon those who don't, I think this could be a great opportunity.
That's basically where I'm at now. Once my account was suspended I retired the spam bot. I even retired a similar project.
BookSuggest, in its web-app form, was the only way I could think of to use my existing technology in a way that wasn't, well... spammy. http://www.charleshooper.net/twitter/
The question I face now is: How do I take this application and deliver it to the user?
My initial idea, given the technology you have, would be to sell it to bloggers. So, let's say I've got a blog with a reasonable readership. I could go through and manually add affiliate links to relevant books, but that sounds like work to me. I enjoy writing; i don't enjoy jacking around with my templates to add affiliate links.
I'd suggest selling 'monetization made easy' movable type and wordpress plugins.
Just an idea; but it might make you money, and it wouldn't be spammy.
but that's the angle I would take. I mean, not only bloggers, but everyone who has content and users who wants an easy way to monetize.
hell, if you had an affiiate account with more than one bookseller, you could even choose the best deal (either the lowest price, or the highest commission, maybe chooseable by your blogger?) then take your cut before the blogger sees it...
Yeah... that's your freemium model. Give away a plugin that does it but with your affiliate id... give the blogger 50% or something. Then, for a higher fee, offer the plugin with the bloggers affiliate id, where they'd get the money direct from the affiliates, without you taking a cut.
That's an excellent idea, and one that I think xulescu was getting at with integrating the book recommendation engine into a bulletin board.
Part of the challenge is maintaining a large enough corpus/histogram to generate good results. Meaning, I'd probably have to host an accessible corpus interface (or make one downloadable) as part of this project. But maybe that's where the dollars come in?
yeah. if you do require a large corpus, that seems like an advantage to me... as if you do that, and maybe make more of the processing done server-side, it'll be harder to just reverse engineer your plugins, and even if people do, then they still have to come up with a good corpus, which presumably you already have.
On the other hand, it's a kind of cool service, as an "enrichment" to twitter messages. If you could find some way of doing it for those who might find it useful without forcing it upon those who don't, I think this could be a great opportunity.