By golly, you’re right. And that’s something a style might want –
superscripting plus boldface. So maybe what the processor does right
now is more correct.**
The current setup works like this:
(a) On ordinary rendering elements, applying formatting attributes
first, then affixes, in the following order, with jiggery-pokery for
punctuation and quotes:
text-case
text-decoration
font-style
font-weight
font-variant
*font-family
(pre-quote punctuation from suffix)
quotes
vertical-alignOn Thu, Jul 9, 2009 at 2:46 PM, Rintze Zelle<@Rintze_Zelle> wrote:
On Thu, Jul 9, 2009 at 12:26 AM, Frank Bennett <@Frank_Bennett> > wrote:
but I think a uniform order of application can safely be used
everywhere, if it’s
made a bit more fine-grained than I now have it. From innermost to
outermost:
text-case
text-decoration
font-style, font-weight, font-variant, *font-family
(pre-quote punctuation from suffix)
quotes
affixes (less pre-quote punctuation)
vertical-align
That’s what I had done. 
I’m saying that for any rendering
element, applying the formatting attributes to its content in the
order I’ve suggested will produce correct output in all cases.
Doh!
OK, so you’re suggesting I put this in the spec more-or-less as is, as
an ordered list?
Yes – it hasn’t been implemented or tested, but if neither you nor
Rintze pick up any problems with it, I’d like to give it a try.
If I recall correctly, in my lifelong experience as CSL style author (:P) I
only ever had trouble with a lack of affix-superscripting in citation
clusters. Any rare cases where affixes should receive all the same
formatting as the main output can always be handled with a conditional and
separate text element for the affixes, e.g.:
>
Alternatively the output can be put into a group, specifying the markup:
>
Label output doesn’t even need the conditional. The only limitation of
Frank’s suggestion I can come up with is formatting of the affixes of the
layout element in citation. You wouldn’t be able to specify (Doe 1999, 2000;
Johson 2000) with everything in bold, right?
affixes (less pre-quote punctuation)
(b) On layout, apply affixes first, then formatting, as follows:
affixes
+
text-case
text-decoration
font-style
font-weight
font-variant
*font-family
quotes
vertical-align
** “Correct” is looking more elastic to me with every passing day,
pretty soon it will have achieved the amorphous transcendence of terms
like “cool” and “legal in Nevada”.
Frank
E.g. by using ‘<layout
font-weight=“bold” prefix=“(” suffix=“) delimiter=”; ">’. Now that I think
of it: Frank, does your CSL processor currently format delimiters as
specified in the calling rendering element, as is the case in the current
Zotero CSL processor*? I think that’s desired behavior.
*With I
get:
(Baranowska et al., 2009; Fayed, 2008; Gerstein, 2008) [i.e. the
delimiters show up as italics, but the affixes don’t]
Yes, delimiters would be boldfaced there. To avoid that, the bold
would go a group enclosing the full content of the layout. The group
would apply to each cite, while the layout applies to the result of
rendering all of the cites.
I’m saying that for any rendering