Keith Devens .com

Search term highlighting on my site

New feature: click through to my site from a search engine (try the first link on this search), and you'll get to see the new feature. It pretty much explains itself. You may have to refresh the page, however, if the page is stored in your disk cache.

I took one regular expression trick from Dean Allen (thanks Dean).

Two issues: I can't do real highlighting of search terms of more than one word. Say you have a phrase in which one word is emphasized. If you search for "one word is emphasized", there's no real way to highlight that with regular expressions, since you could have an arbitrary nesting of HTML. "one word is emphasized"... etc. Is there a trick I don't know about?

The problem could theoretically happen even within one word, since you can have HTML markup within a word, but that's very rare.

The other issue is that I'd probably like to somehow make you only see the highlighted words once, so if you're browsing back and forth around the site you don't keep seeing these ugly yellow words staring you in the face. Not sure whether this is a feature or a bug as it stands. Hmm... by the way, it's super easy to set different caching levels in my CMS. To turn caching off, for instance, all I have to say is $cms->setSetting('http.cache',0).

Lastly, I may eventually do multi-color highlighting like Google does. But plain old yellow will have to suffice for now.

Permalink | 7 Comments | Tags: This website | id 3822

00:39 (utc-4)

Comments

It didn't even dawn on me that this type of thing was even possible.

I got the idea from Dean Allen. I'd been meaning to try it out for a long time, and it worked pretty well. That's why I was looking for a list of search engine syntaxes yesterday. I didn't get to use any of that code from MT Ref or whatever, but it gave me an idea of how to structure mine.

So, right now this thing works with Google, Yahoo, MSN, AOL, AltaVista, Overture, Ask Jeeves, and AlltheWeb. As I get searches engine hits in my referrers I've been making sure they work, and adding things when they don't. I may do it for searches within my web site someday too.

there's no real way to highlight that with regular expressions, since you could have an arbitrary nesting of HTML

Then you have answered your own question. That's the way Google does it and I have yet to see it cause problems in the rendering of a page. Sure it's disgusting when you look at the source code, but on the surface everything seems to be situation normal:

http://72.14.203.104/search?q=cache:74P3...+for+a+list+of+search+engine%22&hl=en

Feel free to post a comment below. Please see my comment policy.

Formatting Rules (No HTML):

**bold**, *italic*, _underlined_, --strikeout--
"text"="url" creates a link, and URLs are auto-highlighted
Blockquote: Like e-mail, begin paragraph with > (greater-than sign)
Lists: begin paragraph with *,-, or + (unordered), or # (ordered)
Code block: ?!code:language=perl|php|sql|javascript|etc.{\n}...{\n}?!/code

Tuesday, April 29, 2003

Search term highlighting on my site

Comments

Recent comments