Re: Metrics for automation ...
- From: Matthias Wolpers <matthias.wolpers@xxxxxxxxxx>
- Date: Wed, 26 Jul 2006 08:37:52 +0200
In article <1153855357.035477.106900@xxxxxxxxxxxxxxxxxxxxxxxxxxx>,
"Vivekanandan M" <vivekm@xxxxxxxxxxx> wrote:
Metrics is colected and analysed to improve the process/efficenfy ofI guess you all know the old joke about the drunk who searches his key
working. In this situation if the metric that you colect does not meet
your objective, you should not consider it at this point of time.
under the street lamp, saying: oh, i lost it over there, but it's dark
so i can't see anything.
In too many shops, too many desperate managers use merics like this.
misconceived / misunderstood (hence: abused!!) metrics are dangerous.
i believe those of us who understand this danger must work to mitigate
those risks. one fairly obvious method is to not randomly collect data
that happen to be flying around anyhow and pronounce them relevant
(like: yes we have a metric, namely the average of tester blood pressure
2 hours into the newest build. can i have my cmm level 3 stamp of
approval now?)
so collecing data is indeed easy (as easy as wasting time, i believe) ,
but knowing which data to collect and being sure of what - if anything -
they measure is highly non-trivial. let's not pretend otherwise, shall
we?
mats
.
- Follow-Ups:
- Re: Metrics for automation ...
- From: Vivekanandan M
- Re: Metrics for automation ...
- From: Shrinik
- Re: Metrics for automation ...
- References:
- Metrics for automation ...
- From: Shrinik
- Re: Metrics for automation ...
- From: Vivekanandan M
- Re: Metrics for automation ...
- From: Shrinik
- Re: Metrics for automation ...
- From: Vivekanandan M
- Metrics for automation ...
- Prev by Date: Re: Exception Handling..
- Next by Date: Testing multiple scenarios
- Previous by thread: Re: Metrics for automation ...
- Next by thread: Re: Metrics for automation ...
- Index(es):
Relevant Pages
|