On 8/25/07, Thomas Hruska <[EMAIL PROTECTED]> wrote: > Paul Herring wrote: > > Out of interest, how many companies (seriously) use LOC as a > > performance metric for their programmers, or as a billing metric for > > their customers? I'm sure I could, without thinking too hard, double > > the LOC in a 'serious' program. > > > > In fact, thinking about it, I do it now - though some of the code is > > autogenerated from existing code using sed.... (No, we don't use LOC > > as a metric.) > > AFAIK, most of that portion of the industry relies on "billable hours" > not metrics. But I sit around spinning my thumbs for hours thinking > about a problem and not write a single line of code. So does that work > any better? That sort of thing makes more money, that's for sure.
This is probably peculiar to my industry but what happens at work is a) client submits requirements b) tech spec drawn up, price agreed there and then (well sorta) c) programming happens on the basis of specs from (b) d) client sees result of (c) and argument ensues about missing/'broken' functionality. Variations to (b) result in increase of price. None of this relies on LOC/hours etc. Just estimates to get things done (we lose if it takes longer) or price rises on stuff that's missing from the spec. -- PJH "Statistics are like a bikini. What they reveal is suggestive, but what they conceal is vital" -- Aaron Levenstein
