Re: [xml-dev] "XML is just syntax" versus "Use semantic markup" (Is this

["James Fuller" <james.fuller.2007@gmail.com>]

On Mon, Mar 10, 2008 at 7:48 AM, Rick Jelliffe <rjelliffe@allette.com.au> wrote:
> All computers can do is play with symbols: labels and pointers.
>
>  XML is just a syntax for annotating strings with labels and pointers. It
>  may even be that successful systems try to defer or avoid or hide issues
>  of semantics as much as possible: some data with a highly specific
>  semantic is just sent as a string, or as a number. Rather than saying
>  "how can we clarify the semantics of this  data" sometimes you need to
>  say "how can we neutralize the semantic problems with this data?"

what I get from this is 'applying semantics late is best practice' is,
for me, a really wonderful (and easy to remember) insight .

somewhat related: I had a discussion with a math professor friend, who
found it amusing to embed meaning into symbol/tags (he has the same
bemusement to variable/constant names in programs).

Of course, he agrees that names/symbols/labels are useful as unique
identifiers and to  imply relationships (esp with operators and other
symbols); but he iterated that the language which describes his
mathematical theories should always be completely de-coupled from the
theory itself (whereby we went on discuss turing machines, and all
manner of problems). I argued that many times the boundaries of
physics and math were pushed forward when new 'languages' (with its
symbols) were developed to comprehend/explain phenomena.

I then went on to explain to him that most people are inextricably
linked to the language they speak .... the analogy I used was that a
lot of perl programmers would find it difficult to use the perl
Lingua::Romana::Perligata module, which oddly enough lets one program
perl in latin of all things;

http://search.cpan.org/~dconway/Lingua-Romana-Perligata-0.50/lib/Lingua/Romana/Perligata.pm

(note there are folks doing perl in klingon as well).

I then started to explain that he is right, as most coding 'anything'
at some point gets converted to a set of arcane symbols that no one
natively understands (or very few of us mere mortals) so in theory he
is correct.

The point that I got out of this discussion was that things like XML
(with the potential to embed 'tactical semantics' in what is
essentially just a syntax of angle brackets) was a *good* thing as it
fundamentally broadened the base of users interacting and creating
datasets; instead of leaving it to the math professors to have all the
fun ;)

cheers, Jim Fuller