[Top] [All Lists]

Re: [ontolog-forum] OWL and lack of identifiers

To: Waclaw Kusnierczyk <Waclaw.Marcin.Kusnierczyk@xxxxxxxxxxx>
Cc: Ontolog Forum <ontolog-forum@xxxxxxxxxxxxxxxx>
From: Ed Barkmeyer <edbark@xxxxxxxx>
Date: Tue, 17 Apr 2007 18:01:09 -0400
Message-id: <462543A5.6010903@xxxxxxxx>
Wacek,    (01)

you wrote:    (02)

> I was actually curious whether you will refer to the XML example.
> The use of 'well-formed' and 'valid' here are synonymous in that both 
> are used to speak of a document's conformance with a grammar;      (03)

With that definition, the terms are equivalent, yes.  One can distinguish:
   well-formed/valid with respect the XML grammar
   well-formed/valid with respect to the DTD/schema
But XML chooses to use 'well-formed' for the first, and 'valid' for the second.    (04)

> equally 
> well it might be said that a document is valid as a XML document and 
> well-formed as an RDF document.  This is just a terminological convention.    (05)

Exactly.    (06)

> The issue with IANA is a bit different, in that a domain name may be 
> registered and unregistered, i.e., its status as valid or not may 
> change, though the rules of the language do not change, in general.      (07)

An interesting point.  This is a big difference between a registry of values 
and an enumerated list.  The validity of a symbol is time-dependent, and 
related to events out of the context of the usage.  So my earlier analogy with 
local symbol definitions, which are solely dependent on other elements of the 
context of use, is inappropriate.    (08)

> On the other hand, a document that is XML-well-formed and XXX-valid remains 
> well-formed and valid, unless the grammars are redefined.  (You might 
> say that a change in IANA registry is, analogously, a change in the 
> languages rules, if you insist.)    (09)

I agree that is stretching the point.  The time-dependency is, to me, a good 
reason for discarding the idea that "hostname validity", in the sense of being 
registered, is "syntactic".    (010)

> The issue is that in the case of XML the use of 'well-formed' for the 
> one and 'valid' for the other is well-defined and documented.  In the 
> case of URLs, I am not sure whether there is an officially established 
> convention of calling a URL 'well-formed' if it conforms to the RFC and 
> calling it 'valid' if it is registered;  maybe there is, but maybe this 
> is only wishful thinking.    (011)

Upon examination, RFC 1123 refers to "syntactically valid" hostnames as those 
conforming to the production rules.  Beyond that, the rest of the terminology 
is all wrapped up with the DNS protocols and failure modes.  There really 
isn't a concept of "valid hostname"; the concept is "DNS lookup succeeds". 
This is further complicated by the fact that the DNS folk believe their 
mechanism can work for anything that supports URI syntax (as long as the names 
are short enough), and therefore DNS is not limited to hostname lookups.    (012)

So I have to admit that there is nothing "syntactic" about the validity of a 
hostname beyond its lexical rules.  After that, there is a process that 
succeeds or fails in one of several ways.  The function of the hostname in a 
URL is to identify one or more Internet hosts that MAY be able to provide the 
service that maps the URL to a resource.  And the mapping of hostname to host 
is properly seen as just part of the complex process of performing the 
URL-to-resource mapping.    (013)

I also have to admit that until just now, I hadn't looked at RFC 1123 for 
years, and I clearly should have, before creating a concept not actually 
supported by the standards.  My apologies to all.    (014)

-Ed    (015)

P.S. I realize that this is one of those areas in which the "grey hair" is not 
useful.  A certain implementation, with which I used to be intimately familiar 
(15 years ago), talks about "valid host names", but the standards don't.  More 
importantly, that experience predates HTTP as the dominant Internet protocol; 
  the concept URL was just coming into existence.  And in that time, one's 
tool did the hostname lookup to find the IP address, and then inaugurated the 
application-specific protocol.  So "hostname validity" was a concept in its 
own right.  But the URL/URI has made the "hostname" issue just a technical 
element of the "valid URL" concept.  And in this "identifier" context, 
separating out the "hostname" is inappropriate -- it is now a "lower-level 
concern".  I just didn't realize until now that I'm still carrying old baggage 
in this area.    (016)

Edward J. Barkmeyer                        Email: edbark@xxxxxxxx
National Institute of Standards & Technology
Manufacturing Systems Integration Division
100 Bureau Drive, Stop 8263                Tel: +1 301-975-3528
Gaithersburg, MD 20899-8263                FAX: +1 301-975-4694    (017)

"The opinions expressed above do not reflect consensus of NIST,
  and have not been reviewed by any Government authority."    (018)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (019)

<Prev in Thread] Current Thread [Next in Thread>