- From: Xiaoshu Wang <wangxiao@musc.edu>
- Date: Tue, 25 Jul 2006 18:21:23 -0400
- To: "'Golda Velez'" <gvelez@webglimpse.org>
- Cc: "'Reto Bachmann-Gmür'" <reto@gmuer.ch>, <public-semweb-lifesci@w3.org>, "'Semantic Web'" <semantic-web@w3.org>
-- Golda > Of course it can - > > http://anydomain.com/some/directory > > may be handled by a script handler, and it can make decisions > based on any environment variable, including browser version, > referring page, etc, what to return. > > Its not HTTP that is partitioning, but the intelligent server > process on the other side... I never say it can. In principle, you can do anything with HTTP. For instance, as I have given the example to put Xpath in the header too. In fact, you can do a lot of "cute" things with the HTTP protocols. But what you CAN do does not suggest that you SHOULD do. Use the right tool to do the right thing and carry it out at the right places. One of the primary principles in engineering is "separation of concerns" because it helps the division of labor, which in turn improves productivity. If you are talking a specific application logic, do whatever you feel comfortable. But if you are proposing an HTTP extension and standard practice, you have to think much beyond if you CAN. Cheers, Xiaoshu
Received on Tuesday, 25 July 2006 22:28:55 UTC