DMOZ and Collaborative Filtering.
Sep. 16th, 2004 11:03 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
swestrupI was just reading an article about The economics of open source hijacking and was only sort-of agreeing with what it had to say.Then it came to a discussion of DMOZ, the Open Source Web Directory. DMOZ was proposed as an alternative to Yahoo, and it has failed. You can read the article if you want more of the details.
This is one of those cases where I don't think the creators of DMOZ can be blamed for the failure. The project was and is a good idea. Its just that today's web is a far different place than the system was designed to work in.
The idea of letting folks sign up and be editors was a good one. The idea of giving away the data was also good. The problem was that the organization that was decided upon does not scale, and its easily overwhelmed by spam. None of these things were easily predictable back then, but that doesn't mean we're stuck with those choices (although DMOZ acts as if it is).
To once again play one of my favorite tunes, I think the answer is in global collaboritive filtering and classification. What is needed is a useful search system that learns what things match what you are looking for, and improves with each use of the system. Strangely enough, I think that the new Search Engine from Amazon called A9, is a stab in that direction. Its not there yet, but could grow into it. The problem, of course, is that its a commercial system that doesn't give a flying fig about privacy concerns, but that's another topic.
This is one of those cases where I don't think the creators of DMOZ can be blamed for the failure. The project was and is a good idea. Its just that today's web is a far different place than the system was designed to work in.
The idea of letting folks sign up and be editors was a good one. The idea of giving away the data was also good. The problem was that the organization that was decided upon does not scale, and its easily overwhelmed by spam. None of these things were easily predictable back then, but that doesn't mean we're stuck with those choices (although DMOZ acts as if it is).
To once again play one of my favorite tunes, I think the answer is in global collaboritive filtering and classification. What is needed is a useful search system that learns what things match what you are looking for, and improves with each use of the system. Strangely enough, I think that the new Search Engine from Amazon called A9, is a stab in that direction. Its not there yet, but could grow into it. The problem, of course, is that its a commercial system that doesn't give a flying fig about privacy concerns, but that's another topic.
