[sc34wg3] CTM: Realistic use cases or toy examples?

[Steve Pepper]

Hi Robert,

You wrote:

| On Wed, Jan 30, 2008 at 02:12:41PM +0100, Steve Pepper wrote:
| > These people, not the programmers, are our primary audience
| > for CTM, and their needs - optimal readability - should be
| > our major concern. That, at least, was the Working Group's
| > position at the Kyoto meeting.
| 
| Here is why your argumentation completely derails me:

Knowing exactly what it is that completely derails you is very
useful, so I will respond at length (even at the risk of
repeating myself).

| I quite agree with you that our target audience should be not
| such much the Java|Lisp|Prolog|Python|Perl programmers, but
| more the computer savvy knowledge worker.

Bravo! That is a step in the right direction. We *almost* agree
on who the target audience is. (I would insert the adverb "more
or less" before "computer savvy". Some of these knowledge
workers are not particularly computer savvy...)

| But all syntax _you_ propose (curly brackets, semicolons
| around every corner, commas, ...) seems all to originate
| from a classical programming world.

Let's start with commas, since you just used three of them in
the preceding paragraph. They clearly do not "originate from a
classical programming world" (although they are admittedly
sometimes used there). They are in fact near-universal artefacts
of most written communication for the last couple of thousand
years and more.

The same goes for semi-colons; despite the fact that they are
less frequent, everyone knows what they mean.

The syntax arrived at by the Working Group in Kyoto uses both
commas and semi-colons in pretty much the same way they are used
in most written language. They are immediately understandable by
any knowledge worker, computer savvy or not.

Curly brackets are (perhaps) a different matter, although in one
sense they can be viewed as variants of the brackets you used in
the paragraph quoted above. (Brackets contain inserted
information that adds to what has already been said and can be
removed without affecting the grammaticality.) As such they too
originate in ordinary written communication and not the
classical programming world. Consider the following:

	puccini {
	  - "Giacomo Puccini";
	  isa composer;
	  born-in lucca;
	  birth-date 1858-12-22;
	  composer-of tosca, butterfly, turandot;
	}

Everything in brackets is "additional information" to the
fundamental assertion. It's like saying

	There exists this subject (and this is what-all we
	know about it...)

I won't belabour this because, as I said, it's just one sense in
which you can look at the curly brackets.

The other sense is as block delimiters familiar from various
kinds of formal languages (including programming languages,
style sheet languages, and more). Precisely because they *are*
familiar -- even to non-programmers -- they are not a bad choice
for delimiting a block of statements.

Personally I am not as wedded to the curly brackets as I am to
the semi-colons and commas. As far as I am concerned, the jury
is still out on the matter of Curlies vs. Period. The balance of
opinion in Kyoto favoured curlies and I will respect that until
such time as a new draft appears and the WG has the opportunity
to assess the consequences.

| If I look again at
| 
|    http://www.semagia.com/tmp/ctm-comparison.html
| 
| then the left column is clearly more closely to natural
| language. And you yourself mentioned in an earlier post
| that you like to see this move.

I'm not sure which post you are referring to and it probably
doesn't matter because I've said quite a lot of contradictory
things in the course of the last two years ;-) At one point
(Montreal 2006) I was interested in exploring how close we could
come to natural language, but I quickly realized that we
couldn't get very close without introducing unacceptable
cultural bias.

What I *would* like to see is a syntax that feels as "natural"
as possible for the target audience, whatever language it
speaks. Naturalness leads to readability, and that's the primary
goal. Semi-colons and commas (and perhaps periods as topic block
terminators) are important in that respect.

Steve
 
--
Conference Chair, Topic Maps 2008
Oslo, April 2-4 2008
www.topicmaps.com