On 11 Jul 2002 at 11:48, Jordi Salvat i Alabart wrote:

> Hi.
> 
> Some opinions and a proposal:
> 
> - I'm with Ilia and Berin that the JMeter development team shouldn't use 
> their limited time to reinvent the wheel.
> 
> - However I'm with Mike that the learning curve to use JMeter (or at 
> least to start using it) should be kept as low as possible. Actually, 
> when evaluating load-testing tools, I dismissed push-to-test's TestMaker 
> in favour of JMeter because TestMaker required me to learn Python. I've 
> been a programmer for many years and I've learned and used probably over 
> 50 different programming languages, so it's not that this scares me: 
> it's just that I'm too lazy. Should be even more of a barrier for 
> non-programmers.
> 
> - Paul's proposal (Bean Scripting Framework) sounds like a good idea -- 
> at least it would provide everyone the opportunity to reuse his/hers 
> programming knowledge. IMO, it would be very sound to use this as a new 
> Protocol or a new Thread type -- in any case, not in a way that 
> *requires* using one of those languages to start using JMeter.
> 
> - I disagree with the idea of adding full functional testing 
> capabilities to JMeter -- even though I agree that there's a need for a 
> functional testing tool out there. All load testing tools I know who 
> have attempted to cover both functional and performance testing have 
> failed in one of the two areas. Mercury covers both, but it's actually 
> by two separate tools for these two problems. Of course it would be 
> really great to have both in the same tool, but the risk of screwing up 
> the whole thing seems too high.

If ever what I do messes up stress testing - I hope people let me know.  My intention 
is to add functional testing capabilities as needed without degrading the stress 
testing 
capabilities.  Thus, I am willing to come up short in functional testing, but not in 
stress 
testing.

> 
> Still, we need some scripting features to be able to test highly dynamic 
> sites. The main requirements for me are:
>    - Being able to compute some variable values programmatically, 
> possibly from the results of a previous sampling.
>    - Being able to use those variable values in a few specific places: 
> form field value, URL query-string parameters, and little more.

Bingo - there are more possibilities, but this is a big chunk of what I'm aiming for, 
and you can see these are relatively simple things.  I'm not convinced a whole 
scripting language needs to be imported to accomplish these things.  

> 
> This could be accomplished by having some Configuration element running 
> a script (maybe a BSF one?) to compute those variable values, then 
> having a variable replacement mechanism in those samplers who wish to 
> implement it -- or maybe some generic mechanism that can be used everywhere.
> 
> Mike: does this approach what you're proposing?

Yes, it sounds like what I'm doing.  I'm creating some Variable and Function 
interfaces that can be implemented very easily to provide built-in variables and 
functions that can work on samplers and other test elements.  The test elements won't 
need to be changed at all.  

Regarding re-inventing the wheel:  As I said previously, it was a mistake to refer to 
what I'm thinking of as "scripting".  It took me one day to build the complex built-in 
function I described in the first email.  

-Mike

> 
> Salut,
> 
> Jordi.
> 
> 
> --
> To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
> 



--
Michael Stover
[EMAIL PROTECTED]
Yahoo IM: mstover_ya
ICQ: 152975688

--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to