URL Rewriting Policies

[haggul]

As part of an overall tidy up we are looking to implement rewriting using IIS URL Rewite Module 2 running on IIS7.

Our current urls are like:

/Product.aspx?PID=123456
/SubCategory.aspx?CID=1&SCID=99

We want to move them to :

/Product/Product-Name/123456
/SubCategory/Subcat-Name/99

The rewriting part is working well on my test site but I then got to reading about rules for trailing slashes and also enforcing lower case.

If I have these urls would they all be seen as different?

/Product/Product-Name/123456
/Product/Product-Name/123456/
/product/product-name/123456

Should I be adding rules to ensure trailing slashes for this style of url, and forcing everything to lower or is that going overboard?

Obviously if we are moving to a new structure we want to do it once and do it right first time.

I am setting canonical tag in each page, and also going to put a 301 in if the incoming URL doesn't match what I expect. Oh and 301s from old style URLs to new. Also in the rewrite module I can set a canonical rule for overall urls to always include the www or not.

Does that sound like a plan or a recipe for instant traffic massacre?!

Thanks.

[phranque]

everything after the hostname in a URL is case-sensitive and the trailing slash is significant and indicates directory structure or hierarchical structure in the path.
those are three different URLs.

on windows server, you must Configure Case Sensitivity for File and Folder Names:
http://technet.microsoft.com/en-us/library/cc725747.aspx [technet.microsoft.com]

i would strongly suggest going with all lowercase in the URL and internally rewrite all requests with uppercase in the URL to a redirect script.

make sure all non-canonical requests are redirected in one hop.

[incrediBILL]

everything after the hostname in a URL is case-sensitive

While that is generally true when working in the real world, I thought Windows boxes were configured case-insensitive by default?

[phranque]

it's the Windows Server OS that is case-insensitive by default, which is a perfect example of why you should look away from MS for proper implementation of HTTP protocol or best web practices.

so yes - "out of the box" it will happily serve 200 OKs for 2**N - 1 non-canonical casing combinations but that doesn't make it optimal or correct or recommended.
the search engines are where it gets "real world" here and they will index those as different URLs which means duplicate content.
your browser won't style your anchor text as "visited" unless the casing of the url matches.

windows server is not the "real world" - it's more like disneyworld.