Re: Linda and the Web

(to www-archive)

BTW Mike, another way to look at this is via a hierarchy of
generalization;


     Object/resource (all things with identity)
        +
       / \
      /   \
     /     +---Containers (all objects that can "contain" other objects)
    /              +
    |             / \
    |            /   \
 Light-Bulb      |    ...
                 
Tuple spaces coordinate at the container level, since a space is the
finest granularity of "identifiable thing" it understands, and just a
container, not a good way to model things whose state you need to
explicitly manipulate (see my "REST Compared" presentation[1] to see
how difficult it is to turn on a boolean lightbulb with tuple spaces).

 [1] http://www.markbaker.ca/2002/08/Rest/ 

MB

On Thu, Dec 05, 2002 at 01:26:33PM -0500, Mark Baker wrote:
> 
> On Thu, Dec 05, 2002 at 10:30:19AM -0700, Champion, Mike wrote:
> > Once again, I relied on my fading memories rather than Google. Bad idea!
> > Sorry for the confusion.  (In my own defense, I was mainly interested in
> > getting discussion started, not in pontificating!)
> 
> No apologies required!
> 
> > Yes, I find Gelertner more intriguing than illuminating ... and AFAIK the
> > Linda papers are not on the Web anywhere, so what I know of it is mostly
> > secondhand. Still, I'm not clear on why GET / PUT / POST / DELETE of 
> > resource representations on the Web is RESTful but (using Javaspaces
> > terminology) read / write / take of Entries in a Space is not.
> 
> Mostly because there's a dichotomy in Linda between tuples and tuple
> spaces; tuples are not representations of tuple spaces like they are
> on the Web (which is why Linda doesn't have the equivalent of PUT).
> Another manifestation of this is that in the Linda protocol, only
> tuple spaces are identifiable, not tuples.  The Web, through the
> composite relationship between a resource/representation (tuple) and a
> container resource (space), defines a more general model.
> 
> There were a lot more design flaws with Linda that explained why it
> didn't become the Web, though.  But since this has little to do with
> reliability, perhaps we should take it offline?
> 
> MB
> -- 
> Mark Baker.   Ottawa, Ontario, CANADA.        http://www.markbaker.ca
> Web architecture consulting, technical reports, evaluation & analysis

-- 
Mark Baker.   Ottawa, Ontario, CANADA.        http://www.markbaker.ca
Web architecture consulting, technical reports, evaluation & analysis

Received on Thursday, 5 December 2002 18:42:34 UTC