Re: [ssax-sxml] The DSSSL/Scheme engine of Askemos / STX

[Kirill L. <lis...@ac...>]

Hello!

On 28 Nov 2002, Joerg F. Wittenberger wrote:

> I would also recommend to restrain from magically bound variables like
> the one which seems to be called 'node' and instead use DSSSL
> primitives like (current-node).  This is done in the Askemos code for
> the sake of documentation.  DSSSL is an well documented ISO standard
> and quite a lot of code is available.  Why create an incompatible,
> while highly similar framework?
> 
'Transformer' of an STX template is a function 

(lambda (node stx:templates stx:root $)
  ...)

  where:
  node - current node 
  stx:templates - list of STX templates
  stx:root -  root of transformed XML document
  $ - environment (XSLT variables)

thus 'node' is its parameter rather than 'magically bound variable'
and I'm in doubts about wrapping it in a function. I do believe
that an extra pair of parentheses is not a bonus, especially for 
a people with XSLT background.

OTH, I'm not proud of my current naming, and I do agree that DSSSL-compatible 
names for 'node' and 'stx:root' ('current-node' and 'current-root' respectively)
will (hopefully) decrease the entropy. 
After all, the name 'current-node' is more descriptive even for people without 
SDQL knowledge, isn't it?

I'm finalizing new edition of STX just now, so it's a suitable time 
for a refactoring. 
I think that the alpha-conversion
[node/current-node]
[stx:root/current-root]
may be a reasonable solution.

What do you think about such a compromise?

Best regards,
         Kirill.