April 7, 2005 06:05 | WebBlogging / WebTech

Yahoo! Term extraction for MT

So Jonas has gotten this working for WordPress, but I have some ideas of how to use it, somewhat differently, with Movable Type. I couldn't code Perl (or anything else really) to save my life so here is just the idea.

In MT admin > entry screen.
a) above the keywords field, add a checkbox to enable "Yahoo! Terms keyword generation"
(We don't wanna start creating new columns/rows/tables in the MT database so let's use Keywords, cause, well, that's what they are.)
b) when an entry is posted, after MT has done everything else it already does, and before we get the "operation completed" message, submit the title, entry and extended entry, as well as the generated permalink (why not... this will notify Yahoo! of a new entry, give it the URL and the text so it doesn't have to come and crawl for it...), get the XML reply of generated "Terms" and save them in the keywords field.

From here, there's all kindsa stuff you can do. Use the keywords for (yawn) header meta-keywords tags, Technorati entry tags or, with a bit of template wizardry (and either some yet to exists plugin, or some CRON and PHP code), generate a weighted list, which you can display as an overall "zeitgeist" (I prefer "what my blog is thinking") affair, as I made for Joi's sidebar. (I generate Joi's global keywords list using the MTKeywords plugin to create and array which I then parse with some PHP.)

Ken McClosky released MTYahooSearch which talks to Yahoo!'s API, but it seems to only pull stuff into templates. Cool, but I want to capture and save the Terms results along with the entry. Why? Because it's relevant?

For fun, someone could also write something that takes the generated keywords and feeds them to Flickr's "tags.getRelated" API call, creating a view of "what the web thinks about what I think". (Nod)

Of course, as someone very smart recently said to me:

"With natural language processing, you end up twiddling a lot of knobs to get what you want."

... so whether you trust Yahoo! to generate relevant keywords is a matter of personal faith.

(Two days in a row staying up till dawn, and I still have an entry on folksonomies in Drupal keeping me awake.)


I haven't checked with the 3.0 code, but in 2.x to do something similar (from the MT standpoint) I had to alter the MT code itself, in addition to the edit post template of course. I couldn't just use a plugin or external library and make user-template changes.

So I'm sure this could be done, but it was annoying to keep adding my changes whenever I upgraded MT, and I doubted any "normal" users would want to go through that so I quit keeping up with it. Maybe they added an ability to hook into the posting process better in 3.x but I don't remember seeing anything like that when I scanned the change list. If 6A would make doing things like this easier I think it would greatly enhance the ability to do customized content management.

Hi! What you do today by evening?