Programming Gotcha's
Where to examine and change source code, and the gotcha's thereof, is very interesting. But I think there are some much bigger gotcha's:
Suppose you don't have the source code
Suppose you "have" it, or you think you do - but it doesn't quite match what's running in production.
Suppose you find and resolve any Y2k issues according to your chosen
approach - but the program won't clean compile.
Suppose you do get it to compile - but it won't execute.
Suppose you do get it to compile, the program will execute - but the results are disastrously incorrect
Suppose you do get it to compile, the program will execute - but the results are incorrect in subtle ways that no-one notices until 9/1/2000.
Each of these situations is going to require some attention from someone who knows what they are doing, and depending on many things, could be very expensive to resolve. Which is why I think the use of "$ per LOC" when trying to assess a project, a contract or a y2k stock should be treated as wild guesswork, not a precise estimation.
Of course, I'm not suggesting I have a better metric - all FBN Consulting clients are naturally given a price based on LOC, but only after they have come up with the required upfront "LOC SWAG evaluation" charges, "good faith" fees, "resource retention and escalation" premiums and just plain old moolah.)
Excerpt from a post written by: KEVIN WATSON |
