On 1/17/07, Gregg Wonderly <[EMAIL PROTECTED]> wrote:
> The last time that I monitored the exchange of a web document retrieved from
> java code invoking;
>
>         URL url = ...
>         InputStream data = url.openStream();

Assuming that's meant to be java.net, right.  I didn't know anybody
used that anymore though.  Any of the 2 HttpClient libs are far
superior.

> There was no use of persistent connections.  An application would have to
> perform some very specific actions, as I said to make this work.  It doesn't
> happen out of the box from a clients perspective as it will with JERI after 
> the
> server is deployed with such a configuration.

It happens out of the box for the vast majority of users.

Seriously, it's been mainstream since about 1999.

> If an application simply opens a socket and transports data, then it is
> ultimately responsible for latency, and the ideas and concepts outlined in 817
> are certainly important issues.
>
> If an application is written using a toolkit or platform or standard which
> doesn't provide a level of control that allows the developer/architect to deal
> with performance at the appropriate level, then that is where the problem 
> lies.

I'd suggest re-reading the layering section.  The tradeoffs there are
complex and important.

>
> >  > WS, REST and IIOP, because they lack mobile code
> >
> > FYI, REST includes a mobile code constraint;
> >
> > http://roy.gbiv.com/pubs/dissertation/rest_arch_style.htm#sec_5_1_7
> > <http://roy.gbiv.com/pubs/dissertation/rest_arch_style.htm#sec_5_1_7>
>
> REST can transport mobile code.  It doesn't use it.  That mobile code doesn't
> allow the transport layer or the interface to "change".

Sure it does.  Have you seen what the WHAT WG is specifying with
regards to non-HTTP protocol use?  Not that I'd recommend it for most
people mind you, but it's available.  See;

http://www.whatwg.org/specs/web-apps/current-work/#network

The caveat is that subsequent interactions with this code will not be
RESTful if the usual constraints aren't followed, but REST most
certainly provides that hook through mobile code.

>  One might use REST/HTTP
> to load a webpage with an applet or javascript which uses NO HTTP to talk to 
> the
> server.  But, I don't consider that a "feature of REST" as much as a "feature 
> of
> Java" or a "feature of javascript" (the feature is the ability to vary
> transport/transfer techonology based on need).

It's a feature of REST, by definition.

Mark.
-- 
Mark Baker.  Ottawa, Ontario, CANADA.         http://www.markbaker.ca
Coactus; Web-inspired integration strategies  http://www.coactus.com

Reply via email to