Syntactic sugar for options: a failed experiment?

Yesterday, we agreed to allow

  <px:step optname="value"/>

as syntactic sugar for

  <px:step>
    <p:option name="optname" value="value"/>
  </px:step>

All well and good, except:

  <px:step ex:optname="foo"/>

doesn't specify the ex:optname option, ex:optname is an extension
attribute. For a different reason,

  <px:step name="foo"/>

doesn't specify the name option, it names the step.

  <px:step name="foo">
    <p:option name="name" value="value"/>
  </px:step>

does specify the name option, but

  <px:step optname="value">
    <p:option name="optname" value="value"/>
  </px:step>

is a syntax error.

It's clearly true that is in some sense the shortcuts are easier to
read, and consequently easier to understand, than the long versions.

But there are a lot of exceptions. Will users find it too confusing?
I fear they might.

I also fear that we're opening a door here to a lot of requests for
syntactic sugar. Why not:

  <p:source port="stylesheet" href="someURI"/>

instead of

  <p:source port="stylesheet">
    <p:document href="someURI"/>
  </p:sourcE>

Allowing all of the different but obvious shortcuts would make a mess of
things, I think.

Abort. Abort. Abort. ? :-)

                                        Be seeing you,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com> | Truth lies within a little uncertain
http://nwalsh.com/            | compass, but error is immense.