"Don't you mean "LAST NAME (Given name)"?"
Of course, sorry for my really bad English!

"I don't think there is the for this situation required support in CSL to explicitly call either given names or last names."
I was afraid to read this kind of answer... Because of the possibility to manage the display of the given name's initial letter (option "initialize-with" for the <name> block), I still hope to find a solution.

Could anyone please help me?

I'm the author of CSL. This is tricky, but I think it probably makes sense to add a new option to names; maybe something like "family-name-format" with option of "capitalize" as a first step.

So there's no solution now, but hopefully there will be before too long.

If you add new option to "names", it will be fine to add prefix ans suffix options (given-name-prefix and given-name-suffix) for givenname part because sometime givennames have to be written with braces and in small caps.
To be able to write LastName part in bold font and FirstName part in normal (roman) font should be also nice.

felwert, I agree with you. Your proposal is simple and seems backward compatible.

And maybe could be compatible with an other request: In french citations sometimes firstnames should be "initalized" not with the first letter but with the first "digram" or "trigram" (Ch. for Charles, Chr. for Christian, Th. for Thomas, Cl. for Claude, ...).I believe it's a usage but not a standard. I don't know how to specifies this.

@bdarcus: I'm not sure I totally understand your concern. Of course sorting and display rules are important when dealing with names, but I don't think this should prevent the implementation of specific requirements from publishers.

If it clarifies the semantics, take "family name" and "given name" instead of forename and surname.

Let's take "Richard Strauss" and "Takeda Shingen". In sort order, they should probably be written as "Strauss, Richard" and "Takeda Shingen", since the family name is usually taken into account for sorting. Given the rules mentioned above, this would display as "STRAUSS (Richard)" and "TAKEDA (Shingen)". The interesting question is how this should be displayed when not in sort order.

I can think of two approaches:

a) Only take the locale into account. So write "Richard Strauss" and "Shingen Takeda" in a European locale. This is the way quotes are handled, when even English titles are quoted using German quotes in a German locale.
b) Take the order of the locale of the person into account, whose name is concerned. So write "Richard Strauss" (as he was German) and "Takeda Shingen" (as he was Japanese).

But none of these should prevent the styling of names. So for example, in many Euro-Japanese co-production movies you would see names in the form of "Richard STRAUSS" and "TAKEDA Shingen".

To be honest, I don't know how Zotero handles the non-sorting order of names. But anyway, style authors should have the possibility to style family name and given name independently. And there are real-world requirements that encourage such a feature.

Also, now that CSL is implemented in a number of different projects, with over a thousand styles, backward-incompatible changes are painful.

Just to be clear: I don't dispute the use cases. Am just leery of these kinds of particular implementation changes.

My concern is that the CSL that would result from your suggestion would either:

be very complex, and very verbose, or ...

not be international-friendly

I have the impression that we're talking of totally different things. So please let me clarify what I really meant.

My suggestion doesn't touch order issues at all. And, as zebulon lined out concerning the initialization, Zotero/CSL does already handle family name and given name differently. E.g., initialize-with does initialize the given name, not the family name ("R. Strauss"). Your suggestion ("family-name-format") also has to distinguish family name and given name.

How, for example, would you encode the rules in your hypothetical example, using the changes you'd like to see in CSL?

E.g. maybe I'm just being dim, but I don't see how a processor is supposed to know how to get "Richard STRAUSS" and "TAKEDA Shingen."

I wouldn't know, either. But I think the problem is display order here, not capitalization.

The MODS standard encodes names like this:

<name type="personal">
    <namePart type="given">Richard</namePart>
    <namePart type="family">Strauss</namePart>
</name>

All I am asking for is different formatting rules for namePart[@type=given] and namePart[@type=family]. The CSL processor can use whatever ordering rules it used to, I don't request any change here.

Also, now that CSL is implemented in a number of different projects, with over a thousand styles, backward-incompatible changes are painful.

I don't think that the proposed changes are in fact backwards incompatible. Quite the contrary.

<names variable="author">
  <name name-as-sort-order="all" sort-separator=" " delimiter=", ">
    <given-name prefix="(" suffix=")"/>
    <family-name text-case="uppercase"/>
  </name>
</names>

The given-name and family-name elements are purely optional. So all the existing styles would be perfectly valid, since they offer all the information needed. New styles could make use of these elements to specify formatting rules that differ between given name and family name. All other information (like ordering) is taken from the name element, just as it used to.

I don't think this syntax is very verbose. It only adds two optional elements which can carry the standard formatting attributes. I think this is a cleaner solution than adding specialized attributes to control formatting of different parts of the name.

I hope I could clarify some points I might have messed up in my previous post.

As I said, "Or, alternately, we'd need some rule that says order per se is unimportant with the cs:names element.." That's what you're suggesting.

But I also suggested that the change might cause other problems. I'll raise the issue with the xbib group (where there are a couple of different projects working on CSL implementations). If they like the idea, then we can add it.

PS - forums maintainers, quotes aren't properly rendered.

PLEASE DON'T POST REPEATEDLY ABOUT THE SAME TOPIC. WE CAN HEAR YOU.

AS IT HAPPENS, WE DID DO SOMETHING FOR EUROPE, AND MUCH OF THE WORK WAS DONE BY EUROPEAN CONTRIBUTORS. ZOTERO 2.1 USES CSL 1.0, WHICH IS DESCRIBED HERE:

http://citationstyles.org/downloads/specification.html#name-part-formatting

THANKS FRANK! EUROPE LOVES YOU!