patch to make the "alias url" look better (readability / search engines)

Wednesday 18 August 2004 3:13:26 pm

Hello,
I noticed that when eZ publish builds an "alias url", it stripped all special characters including accents to replace them with a "_".
I'm french and accents are really <b>common usage</b> here.

Nowadays, search engines (especially Google) give a considerable importance to the url of crawled pages and thats why i wanted to change this default behaviour.

I made a patch against eZURLAlias:: so that it does remove accents and convert ligatures to their 2-letters equivalent.

An url like :

index.php/site/pr_sentation_des_l_ments_de_mon_uvre

Where the title was "Présentation des éléments de mon ¼uvre" (some characters may have been dropped by your browser) would now be translated to :

index.php/site/presentation_de_mon_oeuvre

Witch can make a significant difference regarding to readability and search engines.

I can send the patch to anyone interrested. What is the official way to send a patch to the core team ?

Regards,
Ronnie Garcia.

Wednesday 18 August 2004 10:17:39 pm

There is actually functionality for this almost ready in the 3.5 development version. 3.5 will have support for this, the translation is configurable for the different locales.

--bård

Documentation: http://ez.no/doc

Thursday 19 August 2004 5:59:35 am

Because this discussion is important, i would add some informations about search engines.
Several forum [1] seem to confirm that search engines (at least Google's based ones) dont treat "_" as a word separator.

Then i think it would be worth, while you're at it for 3.5, to change the default behaviour and replace any non-valid character into an hyphen (-) instead of an underscore (_).

[1] :
http://forums.digitalpoint.com/showthread.php?t=210
http://www.markcarey.com/googleguy-says/archives/discuss-googleguy-confirms-that-underscores-are-not-word-separators.html

Regards,
Ronnie Garcia.