W3C home > Mailing lists > Public > www-rdf-rules@w3.org > March 2003

comment on /2001/11/13-RDF-Query-Rules/ re 'single arc languages'

From: Dan Brickley <danbri@w3.org>
Date: Tue, 18 Mar 2003 16:31:20 -0500
To: eric@w3.org
Cc: www-rdf-rules@w3.org
Message-ID: <20030318213119.GA21440@tux.w3.org>

Eric,

I'm reviewing http://www.w3.org/2001/11/13-RDF-Query-Rules/

Rather than try to prepare a single detailed review, I'm sending 
separate msgs on particular bits as and when I grab time to do this. If you'd
rather I tried to make a more unified review doc, I could do that. The 
current approach seemed a quicker way to make gradual progress...


So first thing I wanted to ask about:

Under "Goal Characteristics", you distinguish...
[[
graph or arc
    Some languages express a single arc, others an open subgraph. 
    No observed single arc languages support variables. This 
    leaves them unable to answer the specific query 
    represents(?x ?x) "What lawyers represent themselves?" but 
    instead the more general question represents(?x ?y) "What 
    lawyers represent anybody?". At this point, all single 
    arc query languages are outside the scope of this survey.
]]

This paragraph makes me curious. You rule discussion of these languages
out of scope, and don't cite any. Are they a theoretical rather than 
actual possibility? In my experience 'single arc' functionality is only
exposed via RDF APIs, rather than textually represented query languages.
I don't know of any actual 'single arc' query syntaxes, though perhaps 
some of the path-oriented efforts might fall under this heading?

Your claim that "No observed single arc languages support variables" would 
carry more oomph if you listed some observed single arc languages. Can you
give examples? Even if for the purposes of ruling them out of scope for 
detailed analysis. If there aren't any, this is probably worth noting.

Dan

ps. you also have "some languages" in the first entry, and "some query engines"
in the second, under "Goal Characteristics". Are you comparing query 
languages or software designs / implementations? (these have tended to be 
1:1 in recent history, I guess...). My assumption is that the focus here is
on language comparison.
Received on Tuesday, 18 March 2003 16:31:20 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:53:10 GMT