From: Eliezer S. Yudkowsky (email@example.com)
Date: Tue Jul 17 2001 - 16:09:48 MDT
Bernd Eilers wrote:
> Can you explain why RDF was not considered for Flare ?
RDF is something that you could easily do in Flare, by annotating the
class of a Flare element with the specific kind of information that RDF
calls "metadata". But Flare is not something that you could do in RDF.
The kind of metadata that RDF talks about is a very small subset of what
Flare is talking about.
Similarly for AspectJ; you can easily write programs in Flare that do what
AspectJ does, but you cannot write Flare programs in AspectJ.
There are many bits and pieces out there that allow annotation with one
specific kind of data, in one specific place, to accomplish one specific
useful thing. These individual patterns are not annotative programming
languages, even if they reach out toward annotation in some small way.
> The planar annotations described in
> http://intelligence.org/flare/idioms.html#planar look like a poor mans bad
> reinvention of the XML namespace mechanism that does not provide the same
> flexiblitiy and elegant design of XML namespaces to me.
Well, back when I was first writing my design notes on Flare, there were
no XML Schema. I need to investigate XML Schema in more detail, but my
preliminary investigations seem to show that they don't match the natural
idioms for Flare modules and Flare planes. It looks to me like XML Schema
are based around the idea that every <xyz:arms> tag means the same thing
everywhere in the document. In Flare, <arms> means "appendage" in the
<human> class and <arms> means "ammunition" in the <quakegunner> class,
which is the way that almost every object-oriented language works it.
-- -- -- -- --
Eliezer S. Yudkowsky http://intelligence.org/
Research Fellow, Singularity Institute for Artificial Intelligence
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:36 MDT