[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: How to spell "No PSVI" in XSLT 2.0 ?

• To: xml-dev@l...
• Subject: Re: How to spell "No PSVI" in XSLT 2.0 ?
• From: Arjun Ray <aray@n...>
• Date: Mon, 20 May 2002 20:52:22 +0000
• In-reply-to: <8BD7226E07DDFF49AF5EF4030ACE0B7E06621C24@r...>
• References: <8BD7226E07DDFF49AF5EF4030ACE0B7E06621C24@r...>

"Dare Obasanjo" <dareo@m...> wrote:
| From: Arjun Ray [mailto:aray@n...] 
|>
 
|> What compatibility, other than the mere use of pointy brackets, is at  
|> issue?
| 
| As Paul later pointed out is that they'd have to define a mapping from
| their custom syntax to an XML infoset which in itself isn't hard. 

So then, what or where is the incompatibility?  I sense a shibboleth.

| You are only focusing on one tiny aspect of the implementation. 

Only to the extent that this has been the subject of this subthread.
Perhaps the lack of threading support in your software[*] has lost the
context for you?

| An XML parser that doesn't have to deal with DTDs is not difficult to 
| implement so the payoff there is small. 

The point is expressiveness of a simply parsed syntax (Len's suggestion of
delimited ASCII may have been tongue in cheek, but inability to abstract
from the presumed wonders of pointy brackets seems to be a very common
affliction.)  XML syntax has origins in markup for discursive text; a lot
of it is verbose baggage for representing fielded data sets.

| On the other hand, creating transformers from their custom syntax to XML 
| syntax

Not necessary.  Things like XSchema and XPath and XSLT start with the
infoset.

| or XML object models so that they get the benefits of XSD or XSLT 

What exactly, then, is a parser?

| is an unecessary step that adds runtime overhead

A pointy-bracket parser doesn't add runtime overhead?

| and would involve their users having to learn separate syntaxes

When was the last time you hand-composed a SOAP message for dispatch in
all its pointy bracket glory?

| and other problems.

| For instance, if a stylesheet or validation error occurs in their custom
| syntax should the offending data be shown as XML or in their custom
| syntax?

Is an XSLT or Xschema processor required to "show the offending data as
XML"?  

XML syntax has origins in markup for discursive text.  The practical
considerations for representing fielded data sets are not the same.
Unless, of course, one believes in Golden Hammers. 


[*] Bozoware seems to be the rage these days. 'X-Mailer: Novell GroupWise'
is now in my permanent killfile.

-- 
"The bottomline is that it is really difficult to solve a problem
 when the problem does not exist." - Masataka Ohta.

• References:
  • RE: How to spell "No PSVI" in XSLT 2.0 ?
    • From: "Dare Obasanjo" <dareo@m...>

• Prev by Date: Re: Penance for misspent attributes
• Next by Date: Re: ANN: REST Tutorial
• Previous by thread: RE: How to spell "No PSVI" in XSLT 2.0 ?
• Next by thread: RE: How to spell "No PSVI" in XSLT 2.0 ?
• Index(es):
  • Date
  • Thread