Re: AW: PPIG discuss: software estimating and partitioning

[Jorge Aranda]

Gerold,

Yes, you got the gist of the experiment right.

Regarding Allen's question, we also allowed participants to choose any 
estimation method they desired, to find out if the anchoring effect was 
stronger for some techniques. It seemed that participants were equally 
biased by the effect, no matter which method they chose.



paper at
http://www.cs.toronto.edu/~jaranda/pubs/MScThesis-JorgeAranda.pdf


That's my M.Sc. thesis based on the same topic. For a far more exciting 
read, you can instead find the ESEC/FSE paper here:


http://www.cs.toronto.edu/~jaranda/pubs/AnchoringAdjustment.pdf

Finally, Magne Jorgensen, at the Simula Research Lab, has been doing a 
lot of interesting research on expert-based estimation that should be 
relevant to Allen's question too.


Thanks,
Jorge

 
presentation at

http://www.cs.toronto.edu/~jaranda/pubs/Presentation-AnchoringAdjustment-Feb05.pdf
 
to the best of my knowledge the idea of the experiment was to
include some minor expected duration information (an "anchor") into a 
document

that was the basis for an estimation.
 
in the experiment the estimators got heavily biased by this anchor.
 
best regards,
 
gerold


-Ursprüngliche Nachricht-
*Von:* [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
Auftrag von *janice singer
*Gesendet:* Montag, 22. Januar 2007 01:24
*An:* [EMAIL PROTECTED]; discuss@ppig.org
*Betreff:* Re: PPIG discuss: software estimating and partitioning

Jorge Aranda, a grad student at Utoronto did an excellent study on
software estimation.

J. Aranda and S. M. Easterbrook (2005) Anchoring and Adjustment in
Software Estimation. European Software Engineering Conference / ACM
SIGSOFT Symposium on the Foundations of Software Engineering
(ESEC/FSE'05), Lisbon, Portugal, Sept 5-9, 2005.

Janice


On 1/21/07 5:23 PM, "[EMAIL PROTECTED]" <[EMAIL PROTECTED]> wrote:

A key aspect of programming in practice is the reliable
estimation of size, time and effort.  It seems like most people
that are good at estimating do so by partitioning the problem
into smaller pieces that can be handled more easily.  Then,
final estimates are accomplished by combining the pieces.  This
procedure is certainly what engineering approaches teach and I
think other approaches as well.
 
But I haven't been able to find much empirical data suggesting

that software estimation done by partitioning is superior to
that done more "wholistically".  I assume that I am missing
something huge and obvious since partitioning is such an
important cognitive tool (and has been for such a long time).
But, I haven't found empirical references yet
 
Can anybody direct me to references on this topic.  
Thanks very much
 
Dr. Allen Milewski

Department of Software Engineering
Monmouth University
[EMAIL PROTECTED]
--




-- 


Janice Singer, PhD

NRC Institute for Information Technology | Institut de technologie
de l'information du CNRC

Tel/Tél: (613) 993-7760| Facsimile/télécopieur: (613) 952-7151

[EMAIL PROTECTED]

http://iit-iti.nrc-cnrc.gc.ca  

National Research Council Canada | M50, 1200 Montreal Rd., Ottawa,
ON K1A 0R6

Conseil national de recherches Canada | M50, 1200 chemin Montréal,
Ottawa

(Ont) K1A 0R6

Government of Canada | Gouvernement du Canada



--
PPIG Discuss List (discuss@ppig.org)
Discuss admin: http://limitlessmail.net/mailman/listinfo/discuss
Announce admin: http://limitlessmail.net/mailman/listinfo/announce
PPIG Discuss archive: http://www.mail-archive.com/discuss%40ppig.org/

AW: PPIG discuss: software estimating and partitioning

[Gerold Keefer]

Re: PPIG discuss: software estimating and partitioningpaper at
http://www.cs.toronto.edu/~jaranda/pubs/MScThesis-JorgeAranda.pdf

presentation at
http://www.cs.toronto.edu/~jaranda/pubs/Presentation-AnchoringAdjustment-Feb
05.pdf

to the best of my knowledge the idea of the experiment was to
include some minor expected duration information (an "anchor") into a
document
that was the basis for an estimation.

in the experiment the estimators got heavily biased by this anchor.

best regards,

gerold
  -Ursprüngliche Nachricht-
  Von: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Auftrag von
janice singer
  Gesendet: Montag, 22. Januar 2007 01:24
  An: [EMAIL PROTECTED]; discuss@ppig.org
  Betreff: Re: PPIG discuss: software estimating and partitioning


  Jorge Aranda, a grad student at Utoronto did an excellent study on
software estimation.

  J. Aranda and S. M. Easterbrook (2005) Anchoring and Adjustment in
Software Estimation. European Software Engineering Conference / ACM SIGSOFT
Symposium on the Foundations of Software Engineering (ESEC/FSE'05), Lisbon,
Portugal, Sept 5-9, 2005.

  Janice


  On 1/21/07 5:23 PM, "[EMAIL PROTECTED]" <[EMAIL PROTECTED]> wrote:


A key aspect of programming in practice is the reliable estimation of
size, time and effort.  It seems like most people that are good at
estimating do so by partitioning the problem into smaller pieces that can be
handled more easily.  Then, final estimates are accomplished by combining
the pieces.  This procedure is certainly what engineering approaches teach
and I think other approaches as well.

But I haven't been able to find much empirical data suggesting that
software estimation done by partitioning is superior to that done more
"wholistically".  I assume that I am missing something huge and obvious
since partitioning is such an important cognitive tool (and has been for
such a long time). But, I haven't found empirical references yet

Can anybody direct me to references on this topic.
Thanks very much

Dr. Allen Milewski
Department of Software Engineering
Monmouth University
[EMAIL PROTECTED]
--





  --

  Janice Singer, PhD

  NRC Institute for Information Technology | Institut de technologie de
l'information du CNRC

  Tel/Tél: (613) 993-7760| Facsimile/télécopieur: (613) 952-7151

  [EMAIL PROTECTED]

  http://iit-iti.nrc-cnrc.gc.ca  

  National Research Council Canada | M50, 1200 Montreal Rd., Ottawa, ON K1A
0R6

  Conseil national de recherches Canada | M50, 1200 chemin Montréal, Ottawa

  (Ont) K1A 0R6

  Government of Canada | Gouvernement du Canada

AW: PPIG discuss: software estimating and partitioning

[Gerold Keefer]

> Let me point out that asking an individual developer
> to give an estimate is probably the least accurate and
> effective method, partitioning or no. Books on estimation
> flame about it.

i think it really depends on "how" the developer is asked and
what the purpose of the estimate is.
i see some advantage that the developer also holds the responsibility
for the estimate because this will encourage more commitment.
there is some evidence of this reportedn in a1992 lederer paper.

best regards,

gerold

  -Ursprungliche Nachricht-
  Von: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Auftrag von
Ruven E Brooks
  Gesendet: Montag, 22. Januar 2007 15:01
  An: discuss@ppig.org
  Betreff: Re: PPIG discuss: software estimating and partitioning



  Let me point out that asking an individual developer to give an estimate
is probably the least accurate and effective method,
  partitioning or no. Books on estimation flame about it.

  If you've got no historical data, a much better bet is to get a group of
people to give estimates.   As well as code writers,
  this group should include application analysts, testers, and documentation
developers.  If you want to get fancy about things,
  you can hold formal meetings using the wideband Delphi methodology.  Why
does this work?  For the same reason, any kind
  of group review works; it reminds people of the things they've forgotten
in their estimate.  Although I don't know of any research
  that shows this, getting a full list of tasks probably contributes more to
accuracy than the accuracy of the individual pieces.

  If you've got historical data, then there are a large range of techniques
available, ranging from estimation by analogy up to
  models like COCOMO that take into account a whole range of factors,
including experience of the team and memory constraints
  on the processor(s) on which the application will be run.  There are also
commercial tools such as KnowledgePlan and QSM.

  It's also worth pointing out that on any non-trivial problem, effort
estimation is not the same as schedule estimation.  To do schedule
  estimation you need to understand both intrinsic constraints - "this has
to get written before that" - and resource constraints - "these
  people won't be available until this other project finishes."

  People who do estimation also talk about the "estimation convergence."
Early estimates have very wide error bands; as the project
  proceeds, the estimates get narrower and more accurate.

  Ruven

AW: PPIG discuss: software estimating and partitioning

[Gerold Keefer]

allen,

what you are referring to is commonly called (wholistic) top-down estimation
and (partitioned) bottom-up estimation.

this paper bei jorgensen should give some references:
http://www.simula.no/departments/engineering/.artifacts/jorgensen_shepperd_b
estreview_tse.pdf

a common practice in the environments i am currently working is that
in the bidding process a rough top-down estimate is made by some
senior people mostly by comparison with accomplished projects. in many cases
the difference between the estimate and the bidding price (estimate +
margin)
is not thoroughly addressed.
during the execution of the projects bottom-up estimates are made in order
to
determine the schedule.
at this stage partitioning is more or less natural as tasks of reasonable
size
have to be assigned. in this regard  "partitioning" does not generate any
overhead but greatly improves chances of reasonable estimates and
avoidance of unidentified tasks - and consequently a reliable project
schedule.

best regards,

gerold
  -Ursprüngliche Nachricht-
  Von: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Auftrag von
[EMAIL PROTECTED]
  Gesendet: Sonntag, 21. Januar 2007 23:23
  An: discuss@ppig.org
  Betreff: PPIG discuss: software estimating and partitioning


  A key aspect of programming in practice is the reliable estimation of
size, time and effort.  It seems like most people that are good at
estimating do so by partitioning the problem into smaller pieces that can be
handled more easily.  Then, final estimates are accomplished by combining
the pieces.  This procedure is certainly what engineering approaches teach
and I think other approaches as well.

  But I haven't been able to find much empirical data suggesting that
software estimation done by partitioning is superior to that done more
"wholistically".  I assume that I am missing something huge and obvious
since partitioning is such an important cognitive tool (and has been for
such a long time). But, I haven't found empirical references yet

  Can anybody direct me to references on this topic.
  Thanks very much

  Dr. Allen Milewski
  Department of Software Engineering
  Monmouth University
  [EMAIL PROTECTED]
  --