Re: [xsl] RE: Are there things missing in XSLT which force people to use, say, Java to process XML?

[ac <ac@xxxxxxxxxxxxx>]

Hi Roger, All,

In all fairness, I need to add that for most of my XSLT work, Java is my 
favorite I/O environment and does support most forms of I/O, well 
enough, although some are missing that are becoming more and more 
important, forcing me to turn to other environments, and one of them is 
social-network interactive-3D virtual worlds.

Regards,
Andre




Hi Roger, All,

This is interesting.

For both minds and computers, everything is knowledge and information. 
There is really nothing else.

There are only two types of things that can be done with knowledge: 
sharing (I/O) and computing (typically to generate new knowledge).

The best tool (e.g. language) that we found, so far, to represent 
knowledge (and information, and data, and meta-data), for both computers 
and minds (e.g. humans), is XML.

XSLT is a computing tool which has just enough I/O (sharing abilities) 
(e.g. reading and writing XML) to connect with other systems where many 
of those systems have very specialized sharing (I/O) abilities.

Playing MIDI, displaying graphics, networking, database ACCESS, data 
acquisition, and so on, are all sharing (I/O) operations, each typically 
very specialized.

Whatever sharing channels are used, some computing is required somewhere 
along the line. That is where I use XSLT.

For example, XSLT is great for music, generating and managing complex 
structures (graphs, compositions).  Written to XML they can easily be 
shared and understood by performance systems (e.g. MIDI).

The (e.g. music) things that I can do in XSLT (e.g. computing) I would 
not dare try them directly in something like MIDI systems (e.g. I/O).

Transforming knowledge structures from XML to MIDI performance, for 
example, is comparatively trivial and can easily be automated, often 
even in hardware.

It is good that some tools manage fancy I/O and it is great that other 
tools optimize computation and knowledge management.  We need both.  In 
fact it is probably why we have both: complex senses and minds.  The key 
though is to have a common knowledge representation format (notation), 
with meta-data, between them (and us), hence XML.

In summary, for specialized sharing, I connect to optimized (I/O) tools, 
but for managing and computing knowledge, I really appreciate XSLT.

As a note, Java is interesting and it tries to do it all, but it is not 
very good or specialized at any.  It can do some MIDI and it can compute 
to some degree.  It is mostly very good at managing memory allocation, 
especially on the heap.  As for processing knowledge structures, XSLT 
(with functional programming, streaming, parallelism, and XML) can teach 
it a few tricks.

Regards,
ac




> Hi Folks,
>
> I recently saw the following assertions. Can you help me refute them please:
>
> (1) XSLT is a complete programming language, but doesn't support
>      most things most developers need to do.  (Graphics, networking,
>      relational database access, parsing HTTP headers, generating RSS
>      feeds, peer-to-peer networking, memory management&  caching, thread
>      management, MIDI programming, the list goes on and on and on).
>
> (2) Java (and others) were also designed to be enterprise-class programming
>      languages.  This means the assumption that many programmers will collaborate
>      around a large project.  Encapsulation and complexity hiding are very important.
>      A strongly typed, compiled language (not interpreted) is also important.
>      In short, XSLT wasn't designed for "programming in the large".
>
> /Roger