[ih] Origin of the term 'locator'

[Noel Chiappa]

As a result of recent discussion on another mailing list, I went looking for
the origin of the term 'locator', which was invented by the Nimrod group
(Frank Kastenholz in particular), and used to denote an 'address' which wasn't
carried in packets (since people seemed to have their heads cemented around
the concept that 'addresses' had to be in packets).

I _thought_ it had been created on a public email list (perhaps the Nimrod WG
mailing list), but it turns out to have been in private email, a fork from a
thread on the Big-Internet mailing list (started by me at Mon, 16 Aug 93
23:09:05 -0400, Subject: Terminology problems). So, for the future
historians, I'm posting (with Frank's permission) that private email exchange
to the list, so it's here in the historical record.

	Noel


--------------


 Date: Tue, 17 Aug 93 09:00:06 -0400
 Message-Id: <9308171300.AA04707 at ftp.com>
 To: jnc at ginger.lcs.mit.edu
 Subject: Re: Terminology problems
 From: kasten at ftp.com  (Frank Kastenholz)

Pretty much works for me...

I'd probably use the term "network address" rather than "address" -- 
just to keep it clear that we are talking about them. The term "address"
by itself is used so often that it I fear it will always lead to
confusion. By having the qualifier "network" (or "topological") it
will be a little clearer.

> "interface name", "i-name" - Alternative name for the above, if we don't agree
>         to call it an "address".

No. "Name" does not have the implication of the structure that you
desire. To me, "interface name" would be synonymous with "interface
identifier" (ala endpoint identifier).

Also, using two names for the same thing will lead to confusion.

...


-------


 Date: Tue, 17 Aug 93 12:45:33 -0400
 From: jnc at ginger.lcs.mit.edu (Noel Chiappa)
 Message-Id: <9308171645.AA15614 at ginger.lcs.mit.edu>
 To: kasten at ftp.com
 Subject: Re: Terminology problems

    > I'd probably use the term "network address" rather than "address" -- 
    > just to keep it clear that we are talking about them. The term "address"
    > by itself is used so often that it I fear it will always lead to
    > confusion. By having the qualifier "network" (or "topological") it
    > will be a little clearer.

Yessss, but... It's more of a mouthful, more characters to type, etc, though.
I mean, every message here is about networking, so we only need to qualify
street addresses very occasionally, yes? I use the term "internetwork address"
to differentiate from the physical subnetwork address...

    > No. "Name" does not have the implication of the structure that you
    > desire. To me, "interface name" would be synonymous with "interface
    > identifier" (ala endpoint identifier).

Yes, I just couldn't think of a crisp term (other than "toplogically
structured interface name") to use if Deering won't let us get away with
snarfing "address"! Any suggestions for an alternative?

    > Also, using two names for the same thing will lead to confusion.

I was proposing a term in case we can't use address; if we get it, this will
go away.

...


-------


Date: Tue, 17 Aug 93 13:15:39 -0400
Message-Id: <9308171715.AA15387 at ftp.com>
To: jnc at ginger.lcs.mit.edu
Subject: Re: Terminology problems
From: kasten at ftp.com  (Frank Kastenholz)

> Yessss, but... It's more of a mouthful, more characters to type, etc, though.
> I mean, every message here is about networking, so we only need to qualify
> street addresses very occasionally, yes? I use the term "internetwork address"
> to differentiate from the physical subnetwork address...

I meant "Network" as in "Network Layer" and as opposed to "MAC/Physical".
I know that the 48-bit number that is called an 802 Address is not an
address in the sense that we mean -- but by explicitly prefixing
something to "address" it would be clearer (and would reinforce the
difference to those just learning the terms).

Actually, the ideal term would be "IP Address" but that would probably
lead to a bit of confusion :-)

 >     No. "Name" does not have the implication of the structure that you
 >     desire. To me, "interface name" would be synonymous with "interface
 >     identifier" (ala endpoint identifier).
 > 
 > Yes, I just couldn't think of a crisp term (other than "toplogically
 > structured interface name") to use if Deering won't let us get away with
 > snarfing "address"! Any suggestions for an alternative?

To try to indicate that we are talking about a different set of concepts
than are commonly associated with "address", it might be good to _never_
use the term address, regardless of any qualifiers. If we never use the
term address, then we will not run the risk of "which address"
confusion.

How about locater or location?


-------


Date: Tue, 17 Aug 93 14:14:50 -0400
From: jnc at ginger.lcs.mit.edu (Noel Chiappa)
Message-Id: <9308171814.AA16361 at ginger.lcs.mit.edu>
To: kasten at ftp.com
Subject: Re: Terminology problems

    To try to indicate that we are talking about a different set of concepts
    than are commonly associated with "address", it might be good to _never_
    use the term address, regardless of any qualifiers. If we never use the
    term address, then we will not run the risk of "which address"
    confusion.

As the original message indicated, I'm almost to the point of doing this. It
seems a shame to throw away such a good word, though!

    How about locater or location?

Hmm, pretty good. "Interface locator", maybe?


-------


Date: Tue, 17 Aug 93 14:47:55 -0400
Message-Id: <9308171847.AA20886 at ftp.com>
To: jnc at ginger.lcs.mit.edu
Subject: Re: Terminology problems
From: kasten at ftp.com  (Frank Kastenholz)


 > As the original message indicated, I'm almost to the point of doing this. It
 > seems a shame to throw away such a good word, though!

It must be a good idea -- it was independently arrived at by >1 person
:-)

 >     How about locater or location?
 > 
 > Hmm, pretty good. "Interface locator", maybe?

Sure

Maybe I should copyright it and demand royalties :-)