I’m taking some time off from Findka feature dev this week so I can work on documentation for Biff (and a little code cleanup). So far I’ve finished an example app (Tic Tac Toe). See the readme for instructions on running it in dev and for some good code exploration starting points. Also, please let me know if you want to learn how to use Biff. Here’s a paragraph from the readme:
#biffon clojurians.net for discussion. By the way: I want to help people learn web development with Biff. In particular, doing so will help me improve Biff and its documentation (I also enjoy teaching). So if you want to use Biff, come say hi. If you've got an idea for a web app you want to build, I'd be happy to do some 1-on-1 mentoring/pair programming to help make it happen.
My next item of business is to start working on one of those fancy single-page-documentation sites for Biff (e.g. book.fulcrologic.com).
The on-demand recommendations feature I released last week has been working. Findka has received over twice as much rating data from users as it did the previous week.
However, I’ve come to a mini conundrum. I’m not sure if I should keep adding more features or if I should just focus on getting more users. I want Findka to be so good that you think “wow, Findka is really good.” But I wouldn’t say it’s there yet myself. It still feels at the “hm, this is an interesting idea” stage to me.
One possible fix is to simply get tons of people to use it. The more users Findka has, the more data it’ll have and the better the recommendation algorithm will be. So more users might be all we need to make Findka great. I’ve had success with my tech blogging, so I think I could get steady growth if I made that my #1 priority.
On the other hand, there are lots of interesting additional features I could add to Findka. Perhaps with the right set of features, Findka could become great even without tons of data. That would make bootstrapping growth in this early stage a lot easier.
Some of the features I’m thinking of:
RSS subscriptions. Put RSS links into Findka and then it’ll mix their contents in with your other recommendations. Findka will also learn which subscriptions are more likely to have content you actually like. Compared to existing feed aggregators, Findka would be more adaptive/automatic.
This would be especially helpful with getting good articles into Findka. We have lots of music, books and movies, but few articles (they’re less convenient to import, and an individual article is usually less memorable than e.g. a book you’ve spent 8 hours reading or a song you’ve listened to 100 times).
User profiles. If you opt in, Findka could publish a simple profile that showcases the content you’ve liked. They’d probably look like a simplified version of my personal website. I like the idea of profiles because people have a natural inclination to share good things they’ve discovered with their friends.
There might also be some nice second-order benefits for Findka. For one thing, I noticed that publishing recommended content on my personal website made me much more motivated to go out and actually curate the content I like. If Findka users had simple profiles to share their content, maybe they’d start using Findka more. Besides that, if people share their profiles, that could help Findka get new users.
There are lots of social networking features that I could add on top of that. I could combine it with RSS subscriptions—your profile could have a subscribe button which lets people subscribe to your recommendations by email but also imports an RSS feed for your profile if they’re a Findka user. I’m a little afraid of getting sucked down a rabbit hole though.
So I’m trying to hone in on what the minimum feature set of Findka really should be. One question I’ve tried asking myself is “If I was just taking a sabbatical instead of doing a startup, what features would I want to complete before getting a job?” It’s been somewhat helpful. I’m hoping that letting my thoughts on Findka simmer this week while I work on Biff will help me figure it out.
Another question I ask myself sometimes is “WWPGD?” (What would Paul Graham do.) I think he’d say “talk to your users.” So I’m trying to do more of that too :). I have a couple people I talk to in-depth regularly, but it’d be nice to get some more perspectives.
Next week we’ll be leaving Utah and moving back to the Holy Land (i.e. Washington state, where I grew up). I’ve been here since I started at BYU almost five years ago.