W3C home > Mailing lists > Public > www-archive@w3.org > June 2003

RDF Programming Language

From: Sean B. Palmer <sean@mysterylights.com>
Date: Tue, 1 Jul 2003 03:33:20 +0100
Message-ID: <01a101c33f79$24fc19c0$1d540150@localhost>
To: <www-archive@w3.org>

Pat Hayes was right [1]; the Semantic Web needs a programming
language--probably sometime soon. It needs its Python, of course.

So why not just use Python?

It's not Webized. It's never going to have a <type 'graph'> or be able
to do anything without requiring some external RDF API. RDF API
designs are currently horrible. I've seen some good ideas out there,
but they're all working in a vacuum. I've taken the approach that I'm
not going to write an API until I've got things that I want to do with
that API. And I have got things, so I am working on an API, but it
turns out that I'd really like for Python to undergo some syntax
changes to help. Basically, I need some facilities offered by

So why not just use Notation3?

The builtins facility is broken: you shouldn't mix constraints and
query data. It's reliant upon CWM, which is frequently broken or
buggy, and the builtins packages are non-extensible without source
munging. I've asked for that to change, but it hasn't. It's a
declarative language, and people think in processes. It's a
long-winded way to program; iterating through a list in Notation3 is a
LISPish nightmare. It's not efficient. Notation3 is, in some ways, the
Java of the Semantic Web.

But Notation3 is also very very good. It's the best Semantic Web tool
for processing data that anyone's developed that I've seen so far. It
leans heavily upon the Python stdlib. Because it was developed by
TimBL, it gets things Right with respect to the Semantic Web, more
than anyone else could ever follow the vision correctly.

But this rant is derived from a use case, namely data-use. I've got a
few projects where all I want to do is handle structured data as
smoothly as I would handle text data with regexps. I need bits of
Python, and I need bits of Notation3.

I need that language that some guy proposed ages ago... hmm. I need to
find that.

I should publish my ideas for the Python/N3 hybrid that I've got, but
I'm afraid that someone will actually implement it.


Sean B. Palmer, <http://purl.org/net/sbp/>
"phenomicity by the bucketful" - http://miscoranda.com/
Received on Monday, 30 June 2003 22:33:29 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:42:25 UTC