Re: A few tweaks for ReSpec

On Mar 25, 2006, at 1:19 AM, Maciej Stachowiak wrote:

>
> Hi everyone (and especially Robin),
>
> I made a few changes locally to ReSpec that I think are  
> improvements, and I'd like the ok to check them in.
>
> Here's what I did:
>
> * Implemented <termdfn> and <term> elements, to generate properly  
> styled definitions of terms in prose, and to allow uses of a  
> defined term to link to its point of definition.

* Anne convinced me this was lame, so I did something slightly  
different. Instead of having to write this:

   A <termdfn xml:id="dfn-flux-capacitor">flux capacitor</termdfn> is  
the part of user agents with the highest flux impedance.
   When the engines are damaged, user agents MUST calibrate the <term  
ref="dfn-flux-capacitor>flux capacitor</term>.

I just reused the dfn element for termdfn, and made the XSLT  
stylesheet automatically generate ids and hrefs based on text  
contents of <dfn> and <term> elements, like so:

   A <dfn>flux capacitor</dfn> is the part of user agents with the  
highest flux impedance.
   When the engines are damaged, user agents MUST calibrate the  
<term>flux capacitor</term>.

This generates html that looks like this:

   A <dfn id="dfn-flux-capacitor">flux capacitor</dfn> is the part of  
user agents with the highest flux impedance.
   When the engines are damaged, user agents MUST calibrate the <a  
class="term" href="#dfn-flux-capacitor">flux capacitor</a>.


* I also added a style on the <code> element to make it match normal  
text in size, since most browsers default to a 16-point default font,  
but a 13-point monospace font. (The only exception I could find is  
Opera, this makes <code> spans look too big, I am not sure how to  
address this in a cross-browser way. Any CSS wizards want to help?)

Someone (best of all Robin) please sanity check my XSLT / XPath hackery:

This still includes the other two changes I made:

> * Made xml:id get copied only as id and not as both xml:id and id  
> to passed-through xhtml elements. This is more consistent with  
> generated xhtml elements, less likely to cause validation issues,  
> and the only right thing in case of html output.
>
> * Changed copyright and registered trademark symbols to use numeric  
> entities instead of being included as raw unicode to avoid running  
> into charset issues.


Regards,
Maciej