Re: XSL completeness

["Michael Kay" <M.H.Kay@xxxxxxxxxxxxx>]

Paul Prescod wrote:
>The transformation abilities of the style language are complete: any
>transformation that is possible can be specified.

That's a remarkably strong claim. Do you mean Turing-complete, relationally
complete, or something else? And does the claim refer to XSL itself, or to
XSL extended with ECMAScript?

In a practical sense, I find it hard to regard a transformation language as
"complete" when it lacks string manipulation operators, sorting, totalling,
relational operators such as join, computation and conditional logic, when
it explicitly prohibits global variables, and when it doesn't exhibit
closure (you cannot generate output that you could process as input).

Presumably the claim relies on the language's functional origins, but these
have been so well hidden that without a DSSSL background I for one can't
spot them! Please enlighten me...

Mike Kay


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list