[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: Generality of HTTP

• To: Mike Champion <mc@x...>
• Subject: Re: Generality of HTTP
• From: Dennis Sosnoski <dms@s...>
• Date: Mon, 21 Jan 2002 18:02:18 -0800
• Cc: xml-dev@l...
• References: <1ZXW944832104KJ06VUWRKG831TWTOJ.3c4cbea5@MChamp>
• User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.5) Gecko/20011012

Mike Champion wrote:

>I guess my question is, "If HTTP is so general, why are vendors such 
>as BEA, Tibco, Progress, etc. moving to other protocols to "ensure 
>delivery of XML documents" via web services?  It's easy to agree with 
>Mark Baker (especially after browsing the REST Wiki) that in 
>principle HTTP is a 99/1 solution for internet communications, so why 
>is the world backing away from this in practice?  
>
I'd suggest that HTTP is fine for most synchronous RPC applications. 
IMHO these are likely to be a small and relatively uninteresting subset 
of XML ("web") services in the long run, though. The more interesting 
ones are services designed to be hooked together and composed into 
complex distributed systems, and for these synchronous operation is 
unlikely.

Of course, you can still use HTTP for asynchronous messaging - it's 
sometimes convenient to avoid problems with firewalls and such - but 
it's a design misuse that adds overhead with no appreciable benefits.

  - Dennis

• References:
  • RE: Generality of HTTP
    • From: Mike Champion <mc@x...>

• Prev by Date: RE: Generality of HTTP
• Next by Date: Re: Generality of HTTP
• Previous by thread: Re: Generality of HTTP
• Next by thread: Re: Generality of HTTP
• Index(es):
  • Date
  • Thread