Of course, following up on yesterday’s post about my keyword gatherer, what I forgot to factor in was how long it actually takes to create a large list. Anything up to 1000 words or thereabouts is manageable within a few minutes, more than that and consecutive sessions start to overlap, and my server starts to sweat.

So, I’ve built in a lock so that only one keyword list building process is running a time, and I’ll see how that goes.

Technorati

Remember a few days ago I wrote about related keywords, and how I wanted related keywords to not hav the same words (eg, “texas holdem” is semantically related to “online poker”, but not the same words)?

Well, today’s the acid test.

I’ve just upload my DB with 702 root words (eg, “online poker”, “nlp”, “car tires”, etc). And I’ve set my cron to call my keyword expansion script every 5 minutes. This script first finds the related words and adds them to the DB. Then drills down on the root word (”online poker”) in the normal manner (ie, find all phrases that begin with “online poke”).

One of two things is going to happen:

1. My server will blow up.

2. In 702 * 5 minutes (58 hours) I’ll have a set of 702 very large related lists.

I’ll keep you updated.

Technorati

- Next »