Search engine indexing and '?'

TomGoodacre

Hello!

It's well known that many search engines will not index pages that contain a question mark in the URL.

Can anyone provide me with any advice about a better way to pass information between my pages without using a query string e.g. index.php?page=home. After searching the web a little a number of solutions have popped up - the most popular being 'PATH_INFO' (if using Apache).

I'm just not sure the best way to go with this. Any suggestions and advice most welcome.

Thanks in advance,

Tom

slindstr

You could probably get away with doing something like:

http://www.someURL.com/variable

and then just parse

$_SERVER["REQUEST_URI"]

I've never tried it before but I don't see why it wouldn't work.

If you do try it let me know - I might have to change a few things around on a site I'm working on 🙂

stonefish

Use post rather than using get

then on the next page grab the variables by using $variable = $_POST['nameofinput'];

MarkR

TomGoodacre wrote:
It's well known that many search engines will not index pages that contain a question mark in the URL.

It's also well known that that statement was a lot closer to being true in 1998 than it is now.

Search engines WILL index pages, even if they have a ? in the URL. What they won't do is trawl arbritarily deep into a site, especially if they detect multiple URLs for pages which appear similar (or identical).

They also generally try to avoid indexing URLs which contain session IDs - a good reason not to put session IDs in query strings.

Mark

TomGoodacre

MarkR wrote:
It's also well known that that statement was a lot closer to being true in 1998 than it is now.

So is this something that developers need to worry about or are sites now just as visible to search engines if a query string with question mark is used?

Megahertza

The only serach engine you should worry about are the major ones like google and i'm sure they index pages with ? marks. CNET and EBAY do it so i can't see why its a problem no. so go for it.

bokehman

MarkR wrote:
What they won't do is trawl arbritarily deep into a site, especially if they detect multiple URLs for pages which appear similar (or identical).

I recently did a job for a guy and it was to turn a calendar multi-lingual (4 languages) Calendar script. Anyway to cut a long story short I put it on my personal server for the guy to look at to see if he was happy with it. A few days later I noticed massive traffic in the apache log from Google. It was going through the calendar at some 3000-5000 pages a day.

MarkR

TomGoodacre wrote:
So is this something that developers need to worry about or are sites now just as visible to search engines if a query string with question mark is used?

It depends.

Although a search engine can spider a simple site which uses query strings quite easily, as it becomes more complex, you can overdo it, and the spider won't find things more than a certain distance from your home page (or its entry page).

It's better to have a simple query string rather than lots of mutually related variables. It's much better to have exactly one URL which points to a given page, otherwise a spider will waste its resources fetching the same page under many different URLs, and eventually give up before spidering the entire site.

Having an infinitely deep web site is a bad idea.

It's also a good idea to have your scripts generate redirects and errors correctly for expired, deleted or moved content. If you get an invalid value from the query string, it's better to send a 404 not found error than to either fail with a 400 or 500 error (which a spider such as Googlebot may assume is a temporary fault), or worse still, create a page which is not useful but still has a 200 status.

If a page has genuinely moved, it's best to send a 301 redirect (moved permanently) and then hopefully a search engine will apply existing rankings and links to the page's new location.

If you previously had more than one URL pointing to the same page, it's a good idea to have the non-preferred alternatives generate a 301 and redirect to the "preferred" form of the URL.

Mark