Clean URLs revised

[beate_r]

In order to improve the SEO capabilities of my new site i went into the clean URL subject again. First of all i'd like to note that the rewrite rule i put into the wiki was buggy (.html had to be removed from the request string), and that i have corrected my entry.

Nevertheless i ran into a few issues i would prefer to discuss in the community first:

• sitemap and printing links are not converted. I am unsure wether this is desirable and thus worth fixing or if it is better left alone.
• Umlaute and similar special characters in headers are currently not usable. I consider this a severe limitation in languages using many special characters like french, spanish or german. IMO it is desirable to convert these characters into useful replacements in the links containing them. What Do You think?
• As soon as the URI separator is set to a "/", problems with relative links will occur.

The last one may be nasty because it affects links to style sheets - the head() function writes these links as relative links, such as

Code: Select all

<link rel="stylesheet" href="./templates/epi_beate/stylesheet.css" type="text/css">

. Images and downloads are affected as well. So changing the head() function alone as suggested in the wiki is probably not sufficient.
As soon as the URI separator is a / the link points to a nonexistent location. On the other hand, using relative links is desirable - the installation can be easily moved between different sites and subdirectories. And this is one of the features of CMSimple i like compared to other web apps (Drupal, even worse, SMF). Any suggestions, thoughts?

BTW, as a first workaround, i use the "." as the URI separator. This leads to valid albeit long file names.

Beate

[Till]

sitemap and printing links are not converted. I am unsure wether this is desirable and thus worth fixing or if it is better left alone.

I think this is somethong which can be neglected. This clean_url stuff has been put together, since there are still people who think these "?" in addresses is bad for listing in Google et al. (although this is not true anymore. All modern search engines are able to handle it). And print and sitemap is not that important for listing.

Umlaute and similar special characters in headers are currently not usable. I consider this a severe limitation in languages using many special characters like french, spanish or german. IMO it is desirable to convert these characters into useful replacements in the links containing them. What Do You think?

I believe to rember having seen a code fragment in the old forum which converted umlaute to normal letters. But I may be wrong. One should look there, though.

As soon as the URI separator is set to a "/", problems with relative links will occur.

This is something which I saw myself when I tested this clean_url stuff. But to solve this may not be so easy.

Till

[beate_r]

Probably these two threads?
http://www.cmsimple.dk/forum/viewtopic.php?t=3265
http://www.cmsimple.dk/forum/viewtopic.php?t=5076

Beate

BTW: i can imagine a few other situations where clean URLs are useful, e.g., caching or exporting into static pages (similar to drupal's burst mode). Somewhat exotic, but sometimes needed.

[Till]

Probably these two threads?

Yes, that's it. And I made even a suggestion and cann't remember any more. Wow, what a brain. It was better once, I know it. It may be time for a bullet

Till

[Connie]

no bullets please, take some coffee ;=)

[cmb]

If anybody stumbles upon this rather old thread: I've tried to tackle a clean URL solution for CMSimple_XH 1.5.x, see http://cmsimpleforum.com/viewtopic.php?f=4&t=5719.

sitemap and printing links are not converted.

The sitemap may already work with the new version (however, I have not tested that). But I don't think the print version shall be converted to a clean URL. IMO ?print is correct here: just a parameter which triggers a different view on the same information. However, to avoid DC a canonical link to the page (without ?print) seems to be appropriate.

Umlaute and similar special characters in headers are currently not usable.

This is probably impossible to fix. Eliminate those special characters as described in the mentionend thread.

As soon as the URI separator is set to a "/", problems with relative links will occur.

This should have been fixed in the new version.

This clean_url stuff has been put together, since there are still people who think these "?" in addresses is bad for listing in Google et al. (although this is not true anymore. All modern search engines are able to handle it).

Not long ago I thought so too; but I'm not sure anymore, if clean URLs may actually be a SEO improvement.