Re: 3vl 2vl and NULL

dawn wrote:
> <snip>
>

>>>IBM U2 (IBM's Pick products) suggest significant annual growth, but I
>>>don't know if it is from "the outside" or from ports from other Pick
>>>products.
>>
>>Pollies answer follows - "I can neither confirm or deny".

>>HEY! Isn't that a real world case of 3VL?

For the moment yes but its the real world! Of you get to toss them out in a few years time if it is a material matter and you can replace them with another NULL which is not the same as the previous NULL!

>>>And I probably write a lot because as one of my mentors said (who a
>>>decade later become my employee) "You sure do type fast, but girls do.
>>>That's why they make good secretaries."  (He was explaining why I wrote
>>>code faster than he did.)  Cheers!  --dawn
>>
>>All true - but my measure is the same as Brooks et al - that is working
>>(or debugged) lines! ;-)

Yep, lines! Sure less can be more, but it is not axiomatic in the IT world. Viz the 600 tables - I script all my DDL, makefiles, sprocs and datafixes - so its still lines of code.

> I haven't read Brooks in a while and don't recall what he said about
> lines of code (I loved his MMM book and even bought the revised version
> too a few years ago, even though I have never put in a single man-month
> in my career).

I was referring rather obliquely based on my hazy memories of his quality metrics in a large project.

> I much prefer measuring debugged features completed with regression
> test cases developed compared to required features for a defined phase
> of a project. Even if some features are big and some are small, this
> measure combined with % complete estimates works for me.

Yep me too when working on the surface. When I'm down the hole its cubic metres cut.

Frank. Received on Thu Dec 08 2005 - 12:58:14 CET