Re: The State FAQ

Just a thought:

System theory uses the term "state" in a behaviourial sense.  The state 
represents the "results" of all previous interactions with
tha thing, with respect to how it will behave in a subsequent interaction.

If a thing, which may be accessed thru a URI, has behaviour which 
depends on its previous interactions, then the concept of state
comes into play.

A print server has no relevant state for a print job, but it has state 
which is relevant to the system admin who has to replace
paper and cartridges.  Thus state is relevant to the particular 
interaction which is being considered.

Tom Rutt
Fujitsu

Hao.He@thomson.com wrote:

>Hi, Mark,
>
>So if a request contains a list of URIs and the server needs to deference
>all the URIs in order to process this request, would you consider this
>stateful?
>
>Hao
>
>-----Original Message-----
>From: Mark Baker [mailto:distobj@acm.org]
>Sent: Saturday, 8 January 2005 2:29
>To: Robert Mark Bram
>Cc: www-ws@w3.org
>Subject: Re: The State FAQ
>
>
>
>On Sat, Jan 08, 2005 at 02:09:22AM +1100, Robert Mark Bram wrote:
>  
>
>>Hi Mark,
>>
>>Well done on the FAQ!
>>    
>>
>
>Thanks, Robert.
>
>  
>
>>==
>>1.1 What does it mean for something to be stateful or stateless?    (4ZT)
>>In general it just means that the thing encapsulates state (or not, for 
>>stateless).
>>==
>>
>>I believe the definition should include something describing what state 
>>is. For example:
>>
>>In general it just means that the thing encapsulates state (or not, for 
>>stateless), where state can be thought of as a set of data or attributes 
>>or properties belonging to the stateful entity.
>>
>>It may seem obvious, but I think it is easy to confuse state as a set of 
>>data with a behavioral state as expected in a state transition diagram for
>>    
>>
>
>  
>
>>example.
>>    
>>
>
>I originally had a "What is state" question there, but the best answer I
>could come up with was "data", so I removed it.  I note that yours says
>much the same thing 8-)... though perhaps carries some redundancy with
>"belonging to the stateful entity").
>
>But please, feel free to edit it yourself.  That's why I put it on a
>Wiki!
>
>Mark.
>  
>

-- 
----------------------------------------------------
Tom Rutt	email: tom@coastin.com; trutt@us.fujitsu.com
Tel: +1 732 801 5744          Fax: +1 732 774 5133

Received on Monday, 10 January 2005 05:58:37 UTC