terminological clarification


After all the recent discussions and quite a lot more not CCd to this 
list, I believe there may be a systematic misunderstanding between 
two senses of some of the vocabulary used in the TAG architecture 
document, RFC 2396 and Roy's thesis, and that at least part of the 
fuss has arisen from my falling into this misunderstanding.  Since 
this is likely to happen again with other readers, I would suggest 
that some clarifying prose be added to the document to prevent or 
ward off the possible confusion.

The words in question are "resource" (of course), "representation of" 
and "semantics".

Let me illustrate the point with a simple example. If you click on 
your web browser will show you a picture of Yosemite valley with some 
surrounding text which asserts, not unreasonably, that the web page 
you are looking at is about Yosemite. As we all know, how this 
happened is a complicated process which can be summarized as follows: 
your machine took that URI, analyzed its prefix, asked the Web to use 
the HTTP protocols to prod my server www.ihmc.us with the URI body 
users/phayes/Yosemite.html, which in response emitted some HTML which 
was transmitted back to your browser which then chose some suitable 
way to render the text and image on your screen.  (I know its more 
complicated than that, but never mind.)

Now, there are two ways we could use the above vocabulary to talk about this.

First story (based on my understanding of REST). The "resource"  is 
an idealized abstraction of this page on my server, thought of as a 
kind of idealized Platonic document-in-the-sky (since this particular 
resource is static) and the act of accessing it caused it to emit a 
representation of this idealized platonic document which was then 
used by your browser to create a rendering on your screen. The thing 
on your screen isn't the actual resource, in part because the browser 
itself made some of the decisions about how to render it (eg the 
choice of font for the text), and chiefly because if I were to alter 
the web page (but not the URI) then a refresh would change what was 
on your screen but it would still in some sense be of the 'same' 
resource (so a better idealization of a resource would be a 
time-series of platonic information sources, and this example is a 
rather special case). "Representation" here means a representation 
*of the information*, in the sense that HTML source can be said to be 
a representation of the document one sees when the HTML is rendered. 
"Semantics" here refers to the relationship between representations 
and their network sources. Yosemite valley isn't even mentioned; what 
the thing on your screen is 'about' is irrelevant.

Second story (based on a logical semantics). The "resource" is 
Yosemite valley; the representation is either the HTML source or the 
thing you see on the screen - it doesn't really matter, in this story 
- and the representation refers to, or denotes, the resource. 
"Representation" here is the relationship between the document (in 
whatever form one cares to render it, since at this level they are 
all essentially the same document) and what the document refers to, 
what it is 'about', what it describes. "Semantics" refers to the 
relationship between representations and what they describe. My 
server isn't even mentioned; network protocols, sources of 
information, proxies and so on are irrelevant.

I suspect that the entire corpus of REST-inspired architectural 
documents is written with the intended meanings used in the first 
story. Those words are often read as though they had meanings along 
the lines of the second story, however. Hence, I suspect, much of the 
confusion and debate. If I am close to right, then if there were any 
way to rule out the un-intended meanings, I think that the effort to 
provide that clarification would be very well spent. If there is any 
way I could be of help, I could try to craft suitable wordings.

Thanks for your attention.

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
phayes@ihmc.us       http://www.ihmc.us/users/phayes

Received on Wednesday, 23 July 2003 01:34:53 UTC