Measuring the quality of code

[NotAllThere]

Definitive standard: http://www.osnews.com/images/comics/wtfm.jpg (SFW)

[Churchill]

[minestrone]

Is there any point in writing good code? Nobody else does it, why should I.

[NickFitz]

Definitive standard: http://www.osnews.com/images/comics/wtfm.jpg (SFW)

Not so far fetched:

In my BillG review meeting, the whole reporting hierarchy was there, along with their cousins, sisters, and aunts, and a person who came along from my team whose whole job during the meeting was to keep an accurate count of how many times Bill said the F word. The lower the f***-count, the better.

- http://www.joelonsoftware.com/items/2006/06/16.html

[NotAllThere]

Is there any point in writing good code? Nobody else does it, why should I.

Broken window syndrome in action.

New developments, I write good code. When maintaining code that's well written, I write good code. After all, the next person along to maintain it may be me again.

When maintaining code that's badly written, I really can't be to make the effort.

[BrilloPad]

I am sure in Japan in late eighties that coders wwre paid by line of code? Or did I imagine it?

[minestrone]

You cannot polish tulip as they say. I have to admit I have not seen a well written bit of software for years. Some of the stuff I see is so bad i'm surprised the writers can actually get it to compile.

[NickFitz]

I am sure in Japan in late eighties that coders wwre paid by line of code? Or did I imagine it?

LOC has been used as a metric by idiotic managers. I don't think anybody is stupid enough to do that anymore, but I wouldn't be too surprised.

I knew one team whose manager had decided that he could measure their performance by the number of bugs fixed per month. Towards the end of the month, they would enter a load of bug reports describing various manifestations of a single bug for which they had a fix. Then they'd mark all those bugs as duplicates of the original bug, fix that, and hey presto! two dozen bugs showed up as "Fixed" on the monthly report

[BrilloPad]

LOC has been used as a metric by idiotic managers. I don't think anybody is stupid enough to do that anymore, but I wouldn't be too surprised.

I knew one team whose manager had decided that he could measure their performance by the number of bugs fixed per month. Towards the end of the month, they would enter a load of bug reports describing various manifestations of a single bug for which they had a fix. Then they'd mark all those bugs as duplicates of the original bug, fix that, and hey presto! two dozen bugs showed up as "Fixed" on the monthly report

REminds me of a story I heard from Digital - coder fixed a bug and introduced 4 other bugs while doing it - patched them all. Then got pulled into Manager's office and congratulated for solving 5 bugs in 1 day!

[minestrone]

LOC has been used as a metric by idiotic managers. I don't think anybody is stupid enough to do that anymore, but I wouldn't be too surprised.

Big US investment bank I was working for there brought it in last year. They claimed that it was just for research but there would be emails asking you to justify your output if it was not the same as the previous month. They said that it was to count the lines modified or added before and after a release but they factored in the teams cost to give a cost for line of code added. Everyone ended up writing code to frig the results.

System.out.println( ";;;;;;;;;;;;;;;" ) ;;;;;;;;;;;;;;;;;;;;;

or

int a = 0 ;
int b = 0 ;
int c = 0 ;
int d = 0 ;

then the next month all of the lines would be deleted and you would claim that you were refactoring dodgy code.