Re: Saxon VS XT

Subject: Re: Saxon VS XT
From: Paul Tchistopolskii <paul@xxxxxxx>
Date: Wed, 02 Aug 2000 03:19:59 -0700
----- Original Message ----- 
From: Sebastian Rahtz 

> Paul Tchistopolskii writes:
>  > I know about the key() ;-) I still want the *particular* usecase ;-)
>  > 
>  > For example,  I'm not sure that I'l abuse some part of pipe 
>  > with constant sorting / grouping if I can avoid this in principle. ;-)
> 
> http://users.ox.ac.uk/~rahtz/xsltest/test6.xsl is my simplistic
> example. what's your alternate formulation?

I'l look into it and will get back to you, many thanks.
 
>  > Supporting output encodings other than UTF8 is easy with 
>  > UTF8 - to - some filter. If l have any need in such a functionality 
>  > I'l  grab one of the output handlers from 4xt.org and hack one 
>  > of them. Should take not longer than 5 days. Is it a big deal ?
> 
> Put like that, yes, it sounds a big deal, and well beyond my
> capacities. 

Remember - your claim was : "how can somebody live with
processor which is not 100% conformant". I'm telling you 
that I can live and how do I live with that  not 100% 
conformant XT. I'm *not*  saying *you* should live with 
not 100% conformant processor. 
 
>  > To me - yes. They are not useless when / if you try to save 
>  > performance trying to use 'plain XSLT for everything'
> 
> the use of keys for grouping/sorting is weird, 

And I don't like weired things, because when 
something is weired it is usually 'bad'. I can not 
explain why I think so. 

> but the concept of lookup tables seems eminently sensible 
> to me. what about, for example,
> using a key table to build a back index of of links? so, if 5 elements
> have an IDREF link to an element X, we could list those 5 elements
> when we show X

Are you placing one-pass  as a hard requirment ?

I found ( even in UNIX with 'ordinary' pipes ) that splitting 
some complex processing into several steps improves the  
code and things in general.

Lookups are OK and XSLT allows you to perform lookups 
using readabale constructions == Xpath expressions 
with variables, for example.

Do you  think 'key()'  is *that* critical, if the requirment 
'do it all by one and only one stylesheet' will be removed ?

>  > On another hand  implementing 'too much of XSLT' 
>  > ( like SAXON does ) *also* provides some problems.
> 
> like what, exactly?

Like absolutely not portable 'XSLT stylesheets'. Maybe 
you are accurate and you are using XSLT carefully, 
encapsulating  weired places inside the templates. 
Most people will not,  because writing call-template 
is not a pleasure but it is very close to occupation.

>  > As I said in 4xt list - if somebody will really need key() 
>  > in XT - he'l implement that. 
> 
> I was using XT; I read about key in the spec; I wanted key in XT; I
> could not implement it; I switched to a processor that had it. Sounds
> a more likely scenario to me

For some developer - this is a scenario. For 
some other developers - sighlty hacking 
some good engine is a scenario. 

> when it comes down to it, your real point seems to be that XT is
> "elegant". its probably true. but I am not sure that many of us can be
> bothered either way, its just one bit of software on the desk, and its
> not worth making a fuss over. The `elegance' that makes a difference
> to me is the XSLT language itself (thats what I actually see in my
> emacs buffer), and while I can intellectually appreciate that XT may
> be a marvel of engineering, it doesnt impact on me in my daily life.

Well. You should agree that we are just using XSLT in a different 
way. Not only I'm not using key() ( apply-imports and al otr of other 
XSLT stuff ) , but I'm also not more writing the XSLT code itself, 
because  I wrote XSLScript and it is now much easeir for me to 
write the complex stylesheets. Because XT is 'elegant' is was 
1 day for me to make XT execute the XSLScript code.  That's 
why XT is good for me. You are doing some other things with 
XSLT. That's why XT may be not good for you.  I like consistent 
things and XT is consistent. Maybe you like writing call-template's  
or you found some way with cool emacs programming - 
I don't know. People are different. And taks are different.

Do you remember what was the original question ?

"What is better SAXON or f XT'? 

As I said long time ago ( and at that point of time 
there was some agreement on the list , I think ) : 
it depends on the tasks you are doing. 

If you are embedding XSLT engine - I suggest to take XT and 
enjoy. If you think you can solve real-life tasks ( in some areas ) 
with 'plain XSLT' - god bless you.  

Rgds.Paul.




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


Current Thread