Re: How to measure SIP?

Kent Beck wrote:

> I "only" have an aesthetic/intellectual argument for when to start the 
> SIP clock, not any experience. You have some capital to invest. At 
> some point, you start applying it towards a particular feature, so you 
> can't apply towards any other feature (opportunity cost). At some 
> future point, you receive revenue, so you have capital again. Symmetry 
> suggests to me that we should start measuring when the capital becomes 
> unavailable and stop measuring when it becomes available again. This 
> suggests B to me, since writing a story on a card doesn't really 
> prevent me from writing 100 other stories on other cards.


I agree with you in general, but in our case and from a pure customer's 
point of view, capital is budgeted before iteration planning (ie: at the 
beginning of the month of work), but really expended only when the 
software is validated in pre-production, not before (actually, it's even 
worse: typical invoices in Italy have payment conditions such that you 
receive money 30, 60 or 90 days after issueing the invoice). 
Technically, if the software is not validated, because it is too buggy, 
or we messed up the relationship with the customer or ... the customer 
doesn't spend anything and capital may be diverted to another project.

I don't think that my comment would affect how SIP is calculated, but I 
think it would be worth mentioning it.

> Does that answer your question?


Yes, sort of. :-)
Thanks. I'll report back again as I experiment more.

The real problem that I perceive is customer's idle times: it really is 
out of our control and it depends on organizational and political 
factors that have nothing to deal with our project and the relationship 
with the customer.

Ciao, Giuliano