Hello,
I used the W3 validator to check a multi-lingual SPIP website that uses <multi> in the rubriques. It validates with warnings. I'm not too concerned, as I've used Transitional, and that's adequate for now.
But the warnings that came back are curious.
They tell me that the URLs generated by SPIP are non-conformant:
----------------
Warning Line 111 column 36: reference not terminated by REFC delimiter.
<h2><a href="spip.php?rubrique1&lang=fr">Link Title</a></h2>
If you meant to include an entity that starts with "&", then you should terminate it with ";". Another reason for this error message is that you inadvertently created an entity by failing to escape an "&" character just before this text.
----------------
This was generated by the following SPIP code:
<a href="[(#URL_RUBRIQUE|parametre_url{'lang',#ENV{lang}})]"[ class="(#EXPOSE)"]><span>#TITRE</span></a>
The same problem occurs with #LANG instead, so it's not the #ENV{lang} causing the warning.
It's interesting that the "source code" of the page shows this:
<a href="spip.php?rubrique1&lang=fr">Link Title</a>
That's correct, like the Validator suggests, but it's processed somewhere before hitting the validator.
So it looks like the "generic" syntax of an SPIP-generated URL with the ampersand such as:
http://www.website.com/spip.php?rubrique14&lang=fr
is non-conformant.
Anyone have any experience/views/resolutions to this ?
Does it matter ? 
Best regards,
Mark