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

Re: Why XML?

• To: <xml-dev@l...>
• Subject: Re: Why XML?
• From: Arjun Ray <aray@n...>
• Date: Fri, 16 May 2003 00:56:57 +0000
• In-reply-to: <OF1AE715F3.D4D85349-ONCA256D28.0001D5B4@f...>
• References: <OF1AE715F3.D4D85349-ONCA256D28.0001D5B4@f...>

Matthew.Bennett@f... wrote:

| I don't think [XML] has much of a role in data exchange. 

You may be right.  XML is better suited to the exchange of information
whose structure does not lend itself to a flat dataset representation.

| Two things make me wonder about the competence of those who designed XML.

RTFM is the recommended cure for speculating on competence from a position
of ignorance.

| 1. If you wanted to keep it comprehensible, why invent nonsense like
| attributes?

XML did not innovate attributes.  SGML had them.

| What can they do that nested elements can't? 

Carry information without misrepresentation of analytic structure.  The
element hierarchy is basically a set of nested "consists of" relations in
a direct parallel of parse trees (text at the leaves are "terminals", and
elements in the hierarchy are non-terminals structuring the terminals into
a coherent whole).  Ancillary or "meta" information is accomodated by the
attribute facility.

If you think of attributes and child elements as "equals" or substitutes,
chances are that you'll talk yourself into a poor design. 

| W3C has violated a first-order principle of language design; that there 
| should only be one way of doing something, 

Nonsense.  Orthogonality is for academic dreamers.

| And 2. If a start tag must have a matching end tag, what purpose is served 
| by that ridiculous slash in an end-tag?

Presumably you've never encountered a nested structure in your life.  Can
you decipher this structure according to your apparent solution?

  <section>
  <title>...<title>
  <para>...<para>
  <section>
  <title>...<title>
  <para>...<para>
  <section>
  <section>

(Indenting removed for "ease of comprehension").

That said, you missed the real critique.  GIs in endtags are redundant
unless tag omission is allowed (which is so in SGML, but not in XML.)

• References:
  • Why XML?
    • From: Matthew.Bennett@f...

• Prev by Date: RE: Why XML?
• Next by Date: Re: What are the characteristics of a good type systemfor XML?
• Previous by thread: Re: Why XML?
• Next by thread: re: Why XML?
• Index(es):
  • Date
  • Thread