Re: [xml-dev] What things have achieved universal acceptance acrossthe e

[David Carlisle <davidc@nag.co.uk>]

On 10/01/2011 16:48, Costello, Roger L. wrote:
> Hi Folks,
>
> What things have humans universally agreed to?
>
> I suspect that there is nothing that humans have universally agreed
> to. However, these math symbols:
>
> 1, 2, 3, ..., 9, +, -, /, x
>
> are pretty darn close to being universally accepted.

You haven't defined your terms. It depends what you mean by "accepted" 
or "these symbols".  Few if any systems for example will let you use x 
for multiplication (if that is what you meant) and there is wide 
variation on the meaning of / (especially in the case of integer 
division) not to mention its meaning in xpath.


>
>
> We are 10 years into the XML experiment. In that time span what has
> become universally accepted?
>
> Here are my thoughts on this:
>
> (a) The XML Schema 1.0 datatypes--string, integer, date, time,
> boolean, etc--are used in many XML technologies. For example, they
> are used in XML Schema, XSLT, Schematron, and XQuery.

The other systems you mention define datatypes by reference to xsd so 
it's not that they are sharing datatypes just that they include XSD by 
reference.

>  I don't see any
> XML technologies abandoning those data types in favor of some other
> set of data types.

Others are routinely used with (eg) relax ng, which has a pluggable 
datatype system, either using custom java code or a declarative datatype 
definition language such as dttl

>
> I think that the XML Schema 1.0 data types are universally accepted
> by the XML community.

why do you think that? Often the advice is to avoid xsd datatypes unless 
you are already tied to XSD schema. For example if you want a boolean 
type with two values it's best to avoid xsd:boolean which builds in 
coercion of 1/0 to true/false which is often not intended and just 
define your own datatype that takes true/false (or yes/no or whatever).

>
>
> (b) The XML namespace is built into every XML application.

every xml application that supports namespaces, which isn't quite the 
same thing.

> The XML
> namespace has 4 attributes:
>
> 1. xml:lang 2. xml:space 3. xml:base 4. xml:id
>
> The xml:lang and xml:space attributes are defined in the XML REC.
> They are supported by the XML parser and therefore are available for
> use with any XML application. The other two are not.
>
> I think that the XML namespace, xml:lang, and xml:space are
> universally accepted by the XML community.

presumably by definition the xml community supports things defined in 
the xml spec.
>
>
> That's it. I can't think of anything else that has achieved
> acceptance across all the XML technologies.

>
>
> QUESTION #1
>
> Am I missed anything?


>
>
> QUESTION #2
>
> Why were we (the XML community) able to achieve universal acceptance
> of a set of data types?

we didn't, why should we?

 > What is it about data types that make them
> easy to get universal agreement? Are data types inherently easy to
> standardize? What is it that makes them inherently easy to
> standardize?
>
> Consider a community with many diverse members. The community wants
> to create community-wide standards. Are data types a good target for
> standardizing? For example, the community may create a standard
> Family Name data type.
>
>
> QUESTION #3
>
> Why has there been such a dismal ability to create universally
> accepted markup items? In 10 years we (the XML community) have been
> able to only agree on two markup items: xml:id and xml:space. Why?

the whole point of xml is that it is a syntax definition which allows 
language designers to design a language using a shared syntax.

> For example, why haven't we achieved universal acceptance on
> a<location>  element? What is it about markup items that make them so
> hard to get universal agreement?

It's not hard, it is just not something that xml attempts to do having 
reserved names (such as "location" is totally against the spirit of xml. 
A small (and not particularly successful) exception was made with the 
names in the xml namespace, but they should be seen as the exception 
rather than the rule.
>
> Again, consider a diverse community that is trying to create
> community-wide standards. Are markup items just so notoriously
> difficult to achieve universal acceptance that it is simply a waste
> of time and money to try?

There are many xml languages that have wide acceptance in their 
particular fields,
>
> /Roger

________________________________________________________________________
The Numerical Algorithms Group Ltd is a company registered in England
and Wales with company number 1249803. The registered office is:
Wilkinson House, Jordan Hill Road, Oxford OX2 8DR, United Kingdom.

This e-mail has been scanned for all viruses by Star. The service is
powered by MessageLabs. 
________________________________________________________________________