How to measure SIP?

I enjoyed reading Kent's paper a few months ago, but I have never felt
the need to measure SIP in my projects.

Now I have some time available to experiment therefore I have asked
myself which SIP we have in our current project and I have discovered
that I'm getting confused about the way to measure it.

First of all a brief background about the project. We have an optional
scope contract in place in which we deliver something-like 1 month of
work, chosen by the customer iteration after iteration. The customer
pays the work and then he decides if he wants to buy more (another
month) and so on. The contract is ran using XP.

The customer (a TLC company) has bought 1 month of work and renewed
another time so far, for a total effort of around 1000 "pomodori" (if
you don't know what it means consider 1 pomodoro = 30' actual work of
a pair ==> see for example thread
http://groups.yahoo.com/group/extremeprogramming/m
essages/68430?threaded=1)

Historically:
*) Tue 25/02/03 ==> first meeting with the customer
*) Fri 11/04/03 ==> deploy in pre-production what he has called
release 0.0; customer validates our work and gives permission for
payments
*) during April the customer was almost "sleeping" for his own
organization internal issues
*) Mon 05/05/03 ==> new features from the customer that will
constitute release 0.1 and release 0.2
*) Wed 14/05/03 ==> deploy in production of release 0.0, but really
never officially launched to end users by the customer (organizational
issues again)
*) Wed 21/05/03 ==> deploy in production of release 0.1
*) Thu 29/05/03 ==> official launch to the end users of release 0.1
*) Mon 09/06/03 ==> deploy in pre-production of release 0.2; customer
validates our work and gives permission for payments again

We are currently waiting to deploy in production release 0.2, but no
more functionalities were added since Mon 09/06/03 and all the
requested features have been implemented. Apparently the customer
continuously has these idle times that affect our rhythm.

Ok, if you are still awake I can show you some measures. All the
scenarios that I'm presenting here include customer sleep during
April.

SIP=93 days
This is the value that we get if we count only features that are
currently in use by end users and if we consider that roughly
everything was already at least mentioned during the first meeting
(but not detailed).

SIP=39 days
This value is calculated considering releases 0.0 and 0.1 only, taking
into account as delivery dates Fri 11/04/03 for release 0.0 stories
and Thu 29/05/03 for release 0.1

SIP=35 days
We get this value if we include release 0.2 into the above and if we
consider delivery date Mon 09/06/03 for release 0.2 stories

SIP=22 days
Like the "35 days" case, but considering as the starting date of each
user story the day in which the user story was really planned (every
iteration was defined in a planning game session, every iteration
lasted for about 1 week).

SIP=19 days
A slight modification of the "22 days" case: the starting date of each
user story is the first day in which the user story has been actually
worked. This is suggested in SIP paper too, because "the details are
specified when a programmer has a conversation with a customer just
before programming".

I think that the "19 days" case is a little bit too stretched, because
this particular customer was not on-site every day, sometimes there
were phone conversations, maybe in other occasions the developer
started working with less details, asking questions to the customer
the day after. In any case, I hope that it gives an idea of the
different possible ways in order to measure SIP.

So, what do you think? Which is the most appropriate way to calculate
SIP?

Cheers, Giuliano


------------------------ Yahoo! Groups Sponsor ---------------------~-->
Looking for the latest Free IT White Papers?
Visit SearchNetworking.com to access over 500 white papers.
Get instant access at SearchNetworking.com Today
http://us.click.yahoo.com/YyjvfD/OLNGAA/uitMAA/NhFolB/TM
---------------------------------------------------------------------~->

To unsubscribe from this group, send an email to:
[email protected]

 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/