Forum Moderators: open
I'd rather create a directory structure that matches the logical structure of the pyramid, but I'm not sure if this might lower the PR of deeply nested pages (such as www.site.com/a/b/c/my_page.html).
Is there any evidence that this might hurt my PRs? Should I just dump all pages in the root directory instead?
Thanks
The physical place of a file has no bearing on your PR. It's the hyperlink structure that points towards those pages.
The best use of your directory structure is to include your keywords (theme) and to use it to make sense of all your files.
No need to lump all pages into the root folder.
You might want to check out this thread [webmasterworld.com], by ciml, one of the moderators here.
It basically covers all the things you may have in mind here.....
SN
As people say, the subdirectories have no effect on pagerank. There may still be a fragmentary benefit to be had from keywords as subdirectory names, but I've never really noticed it.
That's from a webmastering point of view, but I still, to this day, believe there's some advantage to using subdirectories. One thing to watch for is URLs getting too long - that's unwieldly and not always attractive. One possibility is to keep it one or two levels deep and just interconnect the right ones by use of the navigation structure itself, which is what controls the flow of PR.
Espeically since you'll probably want to skin/templateify your content pages anyways, andif you have a large site serve them out of a db, you shouldn't have many files in the first place.
For me "directory structure" as seen in the url is simply for easy readign and writing by humans. sorth them out in a mod_rewrite or script. How, and where you pout files should have NO BEARING on your URLs (read the W3C recommendation I posted a while back). This way you'll also never have problems with moving and reorganising content sine the URLs will stay the same (for ever!).
SN
killroy, they do have a relation to maintenance for a good number of sites, and they do have a relation to SEO, which is a significant concern for a great number of us.
Google is not the only search engine, and there can be a certain bearing on rankings with having keywords in the file path of the URL. One of the disadvantages of a data driven site can be lack of exact control over file locations and file names for SEO purposes, including the sequence/order/proximity of the keywords in the file path.
Also, while the visible anchor text is seen on the page itself, the contents of the <a href= > under the hood is read. Some of us would rather have keywords in there, as well as in the path to images.
This might not get readily linked to:
[example.com...]
These are more likely to:
[example.com...]
[example.com...]
If I'm selling stuff for men and stuff for women, that's how I want my URLs to look, if possible.
It's for users, too. There's a recognition factor if people look at a URL that matches what they're looking for.
Even worse are URLs that tell the server what technology to use (script, querystring, variable-value pairs). In my opinion (and those of the W3C) those should be deprecated.
Your URLs should STRICTLY represent the resource. Don't make te visitor tell your server HOW to get the information, jsut what information they want.
So URLS should always be in the form of:
domain.tld/category/sub-category/contentlabel
no extensions, your server should know what type each info is) and no script related fluff, all unnecessary.
And of course this labeling provides the most efficient space for placing your keywords not only in a SEO friendly manner but also human friendly. Which is really what the W3C and Google want.
Does this make more sense?
SN
PS: Sticky me if you want a sample of what I'm talking about...
SN
