I missed the answer for this topic.
Usually client ask this question:
What is the difference between testing client server and web based application?
Any professional answer?
Re: web/client server
Technically Web systems are client/server systems. I usually make the breakdown by referring to traditional client/server and Web-based client/server.
The testing is more in the type of client application and what you are looking for on the back-end. Testing with a Web-based will, most likely, be in a browser. Traditional client/server will, most likely be based on a client GUI application. Obviously this will have some differences on the type of functional and usability testing you will do.
In the realm of performance testing, as far as traditional client/server you are usually dealing with systems that have well-defined workloads and are not necessarily available to the entire world and, if they are, it is a limited type of availability. Web sites (unless you are dealing with an Intranet or even, to some degree, an Extranet) will more generally be potentially viewable by great swaths of people. Thus performance metrics can be very different.
There are also generally protocol differences that you may or may not have to be worried about depending on the nature of your testing.
A lot of the differences you mention would more stem from a distinction between how "white-box" or "black-box" your testing effort is. But the actual processes for testing really do not change, per se. In other words, the "best practices" for traditional client/server are still "best practices" for Web-based client/server.
Keep in mind that this is not talking about Web-enabled applications which are somewhat a different topic but nothing I said above contradicts that.
[This message has been edited by JeffNyman (edited 11-14-2000).]