Re: Picking the Tools -- Marrying processing models to data models

[Al Snell <alaric@alaric-snell.com>]

On Wed, 23 May 2001, Simon St.Laurent wrote:

> Please keep in mind that there are a lot of people who find XML useful
> precisely because information is maintained in a deliberately
> human-accessible format.  

Yup, but I'm trying to isolate another attribute somebody is trying to
express here, so am laying that aside for a moment :-)

> Calling numbers "no less expressive" may be true if you think in turns
> of numbers, but is cold comfort for employees facing lists of
> unidentified SKUs, programmers attempting to figure out what their
> long-gone predecessors were up to, and documentation writers who have
> the pleasure of describing that format for the ages.

Indeed, indeed, which is why the namespace URI should point to something
meaningful, IMHO. A BER-encoded ASN.1 value has enough structural
information to be read as a tree structure full of integers and strings
and real numbers and all that, but unless you can look up the associated
OID (=namespace URI) to find the ASN.1 module declaration, you can't
assign names to the nodes of the tree... this seems to be a better
tradeoff IMHO; the names can be mechanistically extracted for the rare
cases when humans take a peek, but don't get in the way of normal machine
processing, where numeric tags are used instead.

> I still remember that peek(-16384) was the means of reading keystrokes
> on an Apple II, but I kept a reference card around for all the rest.

Modern programming languages would read:

#define KEYSTROKE_IOMAP_ADDR -16384

int getkey() {
	return peek(KEYSTROKE_IOMAP_ADDR);
}

for human consumption, but would actually store:

pushl (-16384)

for execution :-)

ABS

-- 
                               Alaric B. Snell
 http://www.alaric-snell.com/  http://RFC.net/  http://www.warhead.org.uk/
   Any sufficiently advanced technology can be emulated in software