RE: Summing up on visibility(?)

Firewalls have been doing content filtering for quite some time. What's
amazing is how fast hardware-based devices like routers are catching up on
that technology.

Interesting to note that content filtering was originally used for HTTP GET,
not HTTP POST. The fact that GET supposedly has no side-effect is all nice
and dandy and really appreciated. But firewalls don't care much for good
intentions and good development style. They're kind of paranoid. They like
to look at the worst case scenario. And the worst case scenario is about the
same for all past, present and future HTTP methods.

arkin

> > -----Original Message-----
> > From: Mark Baker [mailto:distobj@acm.org]
> > Sent: Thursday, January 09, 2003 1:17 PM
> > To: Champion, Mike
> > Cc: www-ws-arch@w3.org
> > Subject: Re: Summing up on visibility(?)
> >
> >
> > I'd like to phrase it this way, but I doubt the group would
> > ever agree,
> > because I'd have to say things like "If you want to cross firewalls,
> > don't put methods in the body", which clearly people disagree with.
> > That's why I decoupled the description/identification of
> > visilibity from
> > the conclusions about its role and value.
>
>
> Hmm, now I recall some previous threads. I understood the visibility issue
> when it was presented concretely (e.g., "firewalls can't filter messages
> unless the information needed to filter is in the IP or HTTP
> headers").  As
> you imply, I disagree -- <flamebait>saying that web services have
> to respect
> 1996-vintage firewall technology reminds me of the laws that require
> "horseless carriage" owners to have a man with a red flag walking 100 feet
> ahead to warn pedestrians to stay out of the way :-)
> </flamebait>. Firewall
> vendors are developting technology that understands the contents
> of XML and
> SOAP, such is the way of the world.  I can understand why one who doesn't
> have an XML-aware firewall would want to stay away from all SOAP messages
> with methods encoded in the body, but I don't think this is a good
> architectural principle.
>
> I guess I need to know why I should care about visibility other than as an
> abstract property that REST has that other approaches don't in
> order to care
> about it.
>

Received on Thursday, 9 January 2003 16:25:57 UTC