[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: XML as a transition to s-expr

Dorai Sitaram wrote:

> Scott McKay writes
> >
> > Common Lisp already has a syntax for named parameters.  Lisp hacks
> > like this for generating HTML are common.
> >
> > (table (:cellpadding "0" :cellspacing "0")
> >    (tr ()
> >      (td (:width "50%")
> >        "Cell one")
> >      (td (:width "50%")
> >        "Cell two")))
> There are some unnecessary parens here.  I submit that
> s-expression is as much about ruthlessly avoiding
> unneeded parens as about using them when needed.
> (table :cellpadding "0" :cellspacing "0"
>        (tr
>          (td :width "50%"
>              "Cell one")
>          (td :width "50%"
>              "Cell two")))

I think the two of you are getting at the heart of the issue here.
Scott's example uses a syntax that is very easy for a Lisp program to
parse.  Thus it's easy to make a variety of tools which grovel over it
is various ways.  Dorai's example is less verbose (paren-bose??), but
requires a more complex program to parse.  It's more in the direction of
a syntax without parens.  If you wanted to make tools to grovel over it,
you'd invent a simpler intermediate representation.

That's the XML solution, of course.  Of course Lisp lists are an
intermediate representation for s-expressions, too, but the connection
is more direct.

I'm not going to try to argue which is better for programmers.  I will
point out that the s-expr mechanism allows you to choose different
syntaxes, within the range of syntaxes with parens and whitespace.