Re: the term "identifies"

>  > Suggestion, though: don't
>>  say 'denotes'. Instead say something like "indicates",
>How about "identifies"?  :-)
>It seems to me like the word "identifies", as used in RFC 2396 and
>WebArch often bothers you, and I don't quite understand why.

Mostly, I can think of two rather different 
meanings for it, and I want to know which is 
intended, and I can't get an answer, which is 
very frustrating.

>Trying to think as a native English speaker (instead of someone who
>wastes his afternoons on www-tag), I'd say "X identifies Y" means that
>X is a some property uniquely held by Y, or at least held only by Y
>among some implied set of objects.  ("Her dark hair identifies my
>youngest daughter.")  That's a very different sense from how
>RFC2396/WebArch use the word, though.   Let's try a dictionary
>     (transitive senses)
>     1 a : to cause to be or become identical b : to conceive as united
>     (as in spirit, outlook, or principle) <groups that are identified
>     with conservation>
>                     [ No, that's not it. ]
>     2 a : to establish the identity of b : to determine the taxonomic
>     position of (a biological specimen)
>                     [ That's my sense, as in "dark hair" ]

Right, that makes intuitive English sense, sure. 
BUt that just doesn't make sense applied to URIs: 
they needn't do that. If I just make one up at 
random, it almost certainly won't do that.

>     (intransitive senses)
>     1 : to be or become the same
>                     [ Nope ]
>     2 : to practice psychological identification <identify with the
>     hero of a novel> 
>                     [ Nope ]
>So my native-speaker-sense agrees with the dictionary. 
>Putting my programmer/TAG-groupie hat back on.... hold on.  Ah, okay.
>That makes sense: "identifies" as found in RFC2396/WebArch is probably
>back-formed from "identifier" as the term is used by programmers.

That is my hypothesis, also.

>programming language grammars, "identifier" is the superclass of
>"variable", "constant", and maybe some other things.  That is, an
>"identifier" is a string of characters which are associated (in some
>scope) via the symbol table with some area of memory, a constant
>value, a class definition, or something like that.  The string of
>characters "identifies" the memory location, constant value, etc.
>This is kind of sense 2a above: in the context of the particular
>program, the string shows us (via the symbol table) a particular
>(identified) program structure.

Right, that (extended from a memory location to the entire Web) is one sense.

>Anyway, this makes "identifies" pretty much synonymous with "is a name
>for", or "denotes", right?

Nooooooo. That's the point, it doesn't. Denotes 
is the OTHER sense.  They are NOT the same, in 
fact they have almost nothing to do with one 
another, and I still don't know which one is 


IHMC	(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.	(850)202 4416   office
Pensacola			(850)202 4440   fax
FL 32501			(850)291 0667    cell

Received on Tuesday, 22 July 2003 01:57:15 UTC