[JMeter] Contribution to research

[Alberto Bacchelli <al...@usi.ch>]

Dear JMeter developers,

  I'm Alberto Bacchelli, a Ph.D. student in software engineering.

We want to help new developers who join a new software system, and
we believe that a good first impression would attract more contributors.

Imagine a new developer joining JMeter:
As a first step, he needs a high-level view of the system.
Then, and this is what we want to address, he needs to know
what the most important classes of the system are --the hotspots.


We'd like to find *automated* methods to suggest a newbie
which classes he should start to study/understand.


To find the best recommendation method, we must know
the important classes of the system, and you,
as the system developers, are the only ones who can
answer this question.

If you agree to do so (and I really hope so :) )
we will create a small questionnaire for you,
that will take less than 15 minutes to be completed.

Thank you very much for reading this e-mail.
Please, reply to this thread or send me an e-mail
if you want to participate, and/or give me feedback.

Cheers,
  Alberto


PS: Since we want our work to respect the free software philosophy,
if you agree, we would make all your answers public (anonymyzed if you 
wish) as a benchmark, so that other researchers can use your answers to 
propose even better techniques for this task.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: dev-help@jakarta.apache.org

Re: [JMeter] Contribution to research

[Alberto Bacchelli <al...@usi.ch>]

Hi Peter,

  Thank you for your feedback, that addresses one of the reasons
why we are doing this research :)

As you said tutorials (or documentation) are more useful than a list
of "hot" classes. Because they guide you and explain the rationale
behind all the steps. I completely agree with you.

However, the problem lies in the fact that these tutorials/documents
must be kept updated. In your example, you write about the *old*
tutorial, showing the fact that it is somehow outdated. So, we want
to find automatic ways to provide to the new developer the best
guidance that we can.
And we believe that having a list of hotspots, that is always updated,
can be useful. It won't be an excuse for not writing tutorials,
but it could be a good starting point when time is a problem.

The questionnaire will be ready in the next days.
It will be "automatic web interface stuff", an a couple of open
questions. I strongly hope that you can give your contribution.
It would be awesome.

Alberto

On 5/26/10 6:10 PM, Peter Lin wrote:
> several years back Mike Stover and I worked on a detailed tutorial on
> writing plugins for JMeter.
>
> Having a list of the classes is a good thing to have, but I personally
> find tutorials more useful.  The old tutorial is in the jmeter
> resources.
>
> peter


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: dev-help@jakarta.apache.org

Re: [JMeter] Contribution to research

[Peter Lin <wo...@gmail.com>]

several years back Mike Stover and I worked on a detailed tutorial on
writing plugins for JMeter.

Having a list of the classes is a good thing to have, but I personally
find tutorials more useful.  The old tutorial is in the jmeter
resources.

peter


On Wed, May 26, 2010 at 11:59 AM, Alberto Bacchelli
<al...@usi.ch> wrote:
> Dear JMeter developers,
>
>  I'm Alberto Bacchelli, a Ph.D. student in software engineering.
>
> We want to help new developers who join a new software system, and
> we believe that a good first impression would attract more contributors.
>
> Imagine a new developer joining JMeter:
> As a first step, he needs a high-level view of the system.
> Then, and this is what we want to address, he needs to know
> what the most important classes of the system are --the hotspots.
>
>
> We'd like to find *automated* methods to suggest a newbie
> which classes he should start to study/understand.
>
>
> To find the best recommendation method, we must know
> the important classes of the system, and you,
> as the system developers, are the only ones who can
> answer this question.
>
> If you agree to do so (and I really hope so :) )
> we will create a small questionnaire for you,
> that will take less than 15 minutes to be completed.
>
> Thank you very much for reading this e-mail.
> Please, reply to this thread or send me an e-mail
> if you want to participate, and/or give me feedback.
>
> Cheers,
>  Alberto
>
>
> PS: Since we want our work to respect the free software philosophy,
> if you agree, we would make all your answers public (anonymyzed if you wish)
> as a benchmark, so that other researchers can use your answers to propose
> even better techniques for this task.
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: dev-help@jakarta.apache.org
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: dev-help@jakarta.apache.org