Thomas Calivera wrote:
>
> Vincent et al
>
> I attempted to post this earlier, but it apparently did not make the list.
> At this stage in the game, with work on V2 starting, feel free to
> respond, "Sorry, too late."
>
I too am coming in to the thread late- have been having mail problems
over the weekend. Comments below.
> --
>
> complicate matters further. Does anyone disagree with the normative
> statement of "simple, modular testing frameworks are better than integrated,
> complex ones?"
>
I disagree, sort of... ish.
I agree that simple frameworks are better than complex ones.
I don't agree that integrated frameworks are necessarily more
complicated than modular ones. What i think is important is the level of
documentation provided by a framework, whether modular or integrated.
Good documentation is IMHO most important in getting people to use a
project. There are a number of ways of testing server side code, so why
did I recently try out cactus? because of the level of documentation on
the site - it made me believe I could use it, and got me started
quickly.
> Another persuasive argument made by others is that the J2EE interfaces and
> contracts of those interfaces are the central reference for both mock
> objects and containers. In theory, containers run their *own* tests against
> the specified contracts of the interfaces such that an application tested
> against well-designed mock objects will work perfectly in a container so
> tested.
It's the real world evidence that containers are far from perfect that
is an issue, and the "deployment" issues that testing can help with,
like having a team developing on a mixture of NT and Linux, but then
deploying the finished app against Linux only, or other config changes
like using Resin to develop, then Apache and Resin to deploy.
Jari
>
>
<snip>
- comments on container testing removed - i think you've partly resolved
that issue by pointing out watchdog for container comnpliance testing.
--
Jari Worsley
Senior Programmer
Hyperlink Interactive Ltd
