[
Lists Home |
Date Index |
Thread Index
]
[Mike Plusch]
> The ConciseXML syntax can express things that XML 1.0
> syntax can not. For example, attributes that hold any type
> of object, not just a string.
>
But so what. And it is not right anyway. The atribute content is still a
string, right? Unless you are going to allow binary characters disallowed
by XML in there, and then you are really departing from XML.
> There is nothing that the XML 1.0 syntax can express that
> the ConciseXML syntax can not.
>
> Therefore, ConciseXML is a superset of XML 1.0
> (and XML 1.0 is a subset of ConciseXML)
>
No, a subset of Concise it may be isomorphic to XML, perhaps.
> SGML required domain knowledge (knowledge about
> tag meaning) to be able to parse a document. This is
> something that XML 1.0 fixed, and ConciseXML
> shares that same property.
>
> Here is the major problem with XML 1.0:
> If you ask 5 developers to create an XML 1.0 representation
> for a single, well-specific object (say a Java object), then
> you will likely get 5 _different_ XML 1.0 representations
> for that same object. This is a HUGE problem that leads
> to a lot of semantic ambiguity. ConciseXML has no
> such problem.
>
Hell, if you ask 5 developers to create a non-markup serialization for a
Java object you will get five different representations. It has nothing to
with XML-ness and everything to do with design and modeling choices.
Anyway, XML elements are not objects. If you want them to represent
objects, you have to use some conventions. And they will be different for
C++ and Java objects and Eiffel objects and VB objects, so where do you draw
the line? That is assuming that you though that XML's most important
purpose in life is to represent objects. You would get a lot of debate
about that from this list.
Cheers,
Tom P
|
