More visualizers is good.  An idea for a visualizer I'd had was to have a 
vertical bar graph, with each bar representing a different thread.  As each 
sample comes through, the visualizer determines which thread it is, matches 
it with the appropriate bar, and the height of the bar is changed to reflect 
the sample time.  The value in this is that a user can see that ALL threads 
are working.  The can visualize the parallel nature of the test better, In 
think, and they can see how each simulated user is being responded to - maybe 
one user gets quick responses, and another gets slow.

the graph could be configured to either show just the lates sample time (and 
then the bars would be constantly jumping about), or they could represent 
cumulative times, with a colored line to represent the average response time 
for that thread (so, the bars would only grow, and a red-line would show 
total time/number of samples that that thread had done).

Also, I think you should consider making your changes below part of the 
regular Graph Visualizer - maybe as options in the UI.  I just think it would 
be easier for a user to be presented with "Graph Visualizer" as a choice, and 
then configure it to look like they want, and give the info they want, than 
to be presented with "Graph type A", "Graph type B",.... and have to choose 
what they want that way.  Does that make sense?

-Mike

On Friday 17 August 2001 03:40, neth6 atozasia.com wrote:
> I'm gonna work on a few new visualizers :
> 1.  Graph of response time vs sample no. for each
> 'Web Testing' components.  This includes the response time for each
> components (images, applets, etc) and the averages.  Hopefully, this graph
> will give a good view of the components' download times at one glance. e.g.
> ^
>
> |*
> |  *   * *
> |    *
>
> +----------------->
>  1 2 3 4 5   sample no.
>
> 2.  Graph of response time vs number of threads(clients).  Allows u to set
> the upper limit for threads or response time i.e. the logic controller will
> stop sampling once either of the upper limit is reached.  This will give an
> idea of what is the max number of clients tt can hit the site before it
> response time deteriorates to the lowest acceptable one.
>
> ^
>
> |  *****
> | *     *
> |*      *
>
> +------------>
>  1  50  100  Number of clients
>
> Is anybody doing anything similar?  If so let me know so tt I can work on
> something else.  If not, pls give your views on whether the graphs will be
> useful.  Opinions are really welcomed!!!
>
> Thanks.

-- 
Mike Stover
[EMAIL PROTECTED]

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

Reply via email to