W3C home > Mailing lists > Public > www-ws-arch@w3.org > July 2003

RE: Machine-To-Machine

From: <michael.mahan@nokia.com>
Date: Sat, 26 Jul 2003 21:42:12 -0400
Message-ID: <5C76D29CD0FA3143896D08BB1743296A0101CF57@bsebe001.americas.nokia.com>
To: <RogerCutler@chevrontexaco.com>, <www-ws-arch@w3.org>

I think the intention of machine-machine and app-app verbiage is to emphasis 
the absence of presentation data encoded into the message- things like 
<head> and <p> and <table><tr><td> .... hence the 'direct human' phrasing

Tweaking Paul's definition, only changing the first line, I like the rest, 
for the reasons he described:

<suggestion>
A Web service is a distributed software system designed for the exchange 
of messages encoded with functional rather than presentation data. Web 
services use URIs for identifiers, and have interfaces described using XML 
(typically WSDL). Agents interact with the Web service in a manner 
prescribed by its description, using XML-based messages typically conveyed 
using HTTP, SOAP and other Web-related standards.
</suggestion>


-----Original Message-----
From: ext Cutler, Roger (RogerCutler) [mailto:RogerCutler@chevrontexaco.com]
Sent: July 26, 2003 05:45 PM
To: Cutler, Roger (RogerCutler); www-ws-arch@w3.org
Subject: RE: Machine-To-Machine


One more thought.  I believe that if you think about it you will find that the wordings that emphasize the absence of live human beings actually exclude some rather important usage scenarios where the Web services are being invoked from applications being used by people.  Since a number of the member companies of the W3C are putting big bucks into devising and marketing such implementations, this would seem to me to be ... unwise.

-----Original Message-----
From: Cutler, Roger (RogerCutler) 
Sent: Saturday, July 26, 2003 4:28 PM
To: www-ws-arch@w3.org
Subject: Machine-To-Machine


I am in favor of using the phrase machine-to-machine, possibly with a footnote or glossary entry explaining more fully what we mean and reassuring people that we do not intend to exclude implementations on one machine but simply to point out the thrust of what Web services are intended for.  My reasons are as follows:
1 - This was the phrase used by D Booth in his survey which formed the basis of the recent discussion and move-forward on the subject.  That phrasing found strong support and energy.  Nobody seemed to have trouble understanding what it meant.  Why mess with success?
2 - The phrase, to my understanding, is commonly used to mean pretty much exactly what we want to use it for.  Why mess with common usage?
There are some reasonable alternatives that have been proposed.  I think that they all have weaknesses.  Specifically: 
"Application-to-application": 
1 - Somewhat embarrassingly, browsers and Web servers are applications. 
2 - Some might think that this term unduly emphasizes the RPC style of Web services. 
"interactions between agents (as defined by ws-arch definition of agent), where the agents are not directly controlled by humans":
1 - Wordy and not oriented toward common usage of terms. 
2 - "Not directly controlled by humans" might be interpreted completely differently than intended.  "Directly" might reasonably be interpreted as "without intermediary".  I think that the word David might have been reaching for might have been "synchronously", but that word has problems with us at the moment.  We are currently deprecating its use.  At any rate, I think that this kind of detailed verbiage belongs in an explanation of what we mean by "machine-to-machine", not up front.
"designed to support interactions that do not require human involvement at runtime 
between agents over a network" 
1 - Grammatically fuzzy (antecedants are unclear), but I suppose it can be cleaned up.  Is it "interaction between agents" or "involvement between agents"?  I actually don't know.  Hmmm, probably interaction.  How about, "designed to support interactions between agents over a network that do not require human involvement at runtime"?   Ugh, that's a syntactic mess too.  Does "require" refer to "network"?
2 - Why not just "machine-to-machine".  This seems to be an attempt to explain in a lot of words what that means.  Flesh it out into several sentences so that what modifies what is clear and this seems like another good start at a glossary entry for machine-to-machine.  But fleshed out into several sentences is probably too much for the context in which we are talking about using it.
Received on Saturday, 26 July 2003 21:42:47 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:21 GMT