Re: Directed vs. Generic syntax reprise (Was: Re: Syntax noodling)

Jeni Tennison wrote:
> 
> I'd like to avoid defining a directed-syntax shorthand in this version
> of XProc, because it's more work and because I'm worried about
> specifying a language that different users may use completely differently.
> 
> I suggest that we put in place the restrictions that make a directed
> syntax possible, particularly ensuring that inputs, outputs and 
> parameters have to have different names. As you say, it should be a 
> fairly straight-fo XSLT transformation from directed to generic syntax, 
> so it would be easy to put together a toolset that supports authors 
> writing in a directed syntax without us having to specify it now.

I just want to register that I'm conflicted:

    * I believe, based on those around me using smallx, that a directed
      syntax for steps is *easier* to use.  Normal, non-XML-super-geeks,
      seem to find it much easier to understand how to use the step.

    * In smallx, this is way simpler because I have only one input and
      one output.  So, the above "user experience" doesn't map completely
      to a more general situation.

    * Having a directed syntax makes it harder on tool providers to
      understand an extension step in that they need to know a lot about
      the directed syntax (e.g. the element structure mapping to inputs,
      outputs, and parameters) to get the flow graph correct.

    * But, I still really want this... :)

I think we need to decide whether we will have step definitions in
our language.  If so, I think this problem becomes tractable.

I think whether or not we have step definitions is the first question we
need to answer.


--Alex Milowski

Received on Friday, 12 May 2006 15:53:20 UTC