DOM::Document::getElementsByTagName bug ?

[Shift]

Hi,

I have few questions about "NodeList DOM::Document::getElementsByTagName 
(const DOMString & tagname)".

Is case important for this function ?
Few days ago when I used getElementsByTagName ("LINK") in 
http://bugs.kde.org/show_bug.cgi?id=57803 I retrieved 8 elements and now 0.
So I decided to use getElementsByTagName ("link") an then I retrieved 8 
elements.
Same test by modifying the HTML source and using "LINK" instead of "link", 
same results.

So my questions are :
Should the argument of getElementsByTagName  always in lowercase ? 
Is it a DOM specification ? (When I look at the DOM specifications 
http://www.w3.org/TR/REC-DOM-Level-1/level-one-core.html I can't find an 
answer)
Is it a bug ?

Thanks,

Franck