Re: CG: Semantics, XML, and XQuery
Danny,
Your summary is a fair and realistic assessment of the
situation. I am just frustrated at seeing one more case
where syntax wins out over semantics.
I had been teaching logic to IBM programmers since the
late 1960s. In the early 1980s, I was teaching Prolog
and arguing against the tidal wave of expert systems
implemented in ad hoc notations.
As a result of my preaching, a bright young programmer
in IBM Paris implemented a world-class version of Prolog
that ran on IBM mainframes (and later on PCs). But IBM
threw away several million dollars on pushing two pieces
of garbage -- an Emycin look-alike and an OPS5 look-alike
-- both of which were far slower, far less powerful, and
far less general than the equivalents implemented on top
of IBM Prolog. I showed them how to write a simple Prolog
parser that would give the users a friendly syntactic
sugar over Prolog, but the US consultants assured IBM
management that Prolog was "outside the mainstream"
(at least in the US).
More recently, I have been trying to show people the
common logic underpinnings of UML, SQL, and every other
declarative notation on planet earth. But the syntactic
tidal wave sloshed over everything and capped the little
wavelets in angle brackets.
DA> Download time isn't really a problem - gzip (immediately
> available on most web servers) can drastically reduce bloat,
> and hardware is improving all the time.
I've been hearing arguments like that since the 1960s, and
they have led people astray every time. I've complained
about them every time, and the answer has always been
"But this time is different."
DA> On the other hand, to get a new syntax widely adopted
> would require a major feat of social engineering. I think
> it's easier just to accept that XML isn't perfect, but
> at least it's out there. Which leaves time for more
> interesting problems above the syntax ;-)
I don't believe that most programmers will ever see anything
above the syntax. Right now, the only syntax I'm pushing is
controlled natural language. My goal is to appeal to a higher
authority -- the end users.
John