Canonical tag for variable urls?

[realmaverick]

I'm currently redesigning a file search engine and improving the architecture of the site.

There is a top searches section, which links to the friendly URL version of the query i.e site.com/f/widget/. When the user searches, they get a non friendly version of the URL i.e site.com/search.php?q=widget, though there is a canonical tag, linking back to the friendly URL.

On the search page, there are various sorting options, for changing file size, source, file type etc. Resulting in site.com/search.php?q=widget&select=all etc

When a filter is applied, my instinct is to use the Canonical tag, linking back to site.com/f/widget/

Though here is a possible issue; when a filter is applied the number of files and files displayed, can change, making the page technically different.

I've taken a look at the leading file search engine and they're not using canonical tags at all, let alone in this scenario. They're not using any method at all, to consolidate the friendly and non-friendly urls or the sorting options.

So, would it be a good idea to use the canonical tag here?

[deadsea]

For search results, I use robots.txt.

1) There are so many search terms that robots can waste tons of time and your site resources crawling in there.
2) When you have sort options and such, it just multiplies the problem
3) Google has said that they don't like search results in their index. I've seen evidence that when a page looks like a search result, it loses rankings.

Put /search.php in robot.txt and be done with it.

[tedster]

First, make sure that these "top searches" URLs are doing you some good. As deadsea accurately says, as a general rule Google does not like search results in their index.

Even if those top searches are good for traffic, I would definitely not allow the urls for various "sorts" to be indexed. So even if you do decide to allow a few top searches to be index, I'd say use robots.txt to block the sort parameters and use the canonical link tag as a secondary failsafe.

[realmaverick]

Hi guys thanks for your replies.

If you search for [a 2-word phrase] in Google, second result is [a competitor] with a 126 Alexa rank. It's usually their search results pages, that ranks before their actual info pages.

So it's essential we do index our results. But good idea about using robots.txt as another way to keep them out.

[edited by: tedster at 3:35 am (utc) on Aug 6, 2011]
[edit reason] no specifics, please [/edit]