News for hackers. Stuff that matters.
Re:On Wages by kurtosis on Thursday January 21, 2010 @02:44PM (#1068137)
I've heard the story where there was a 10x difference in the time it took CS students to complete a series of system programming assignments, but are there other anecdotes out there? I'm very suspicious of this story, and think that a lot of this fabled 10x multiple can be explained by extrinsic circumstances or luck. At least that's what I've observed in science where easy experiments in a field that is new and fashionable can yield lots of interesting papers, where taking on more difficult projects in established fields can require much more work to achieve incremental results.
Re:On Wages by rw140 on Thursday January 21, 2010 @05:35PM (#1068542)
Some of it might be a matter of experience.

It's just taken me 30 days to add some new CLI functionality to a piece of code (my first ever time working with this bit of the code stack - I'm not familiar with the underlying code, the debug tools, the various ways the build can fail). I know that the standard time for an experienced CLI developer to implement the equivalent extension of functionality on this section of the code is 3 days.

My second piece of CLI functionality is on track for 15 days... so by the time that's finished I will only be 5 times worse. :)

Re:On Wages by fhars on Thursday January 21, 2010 @03:07PM (#1068184)
That page gives an overview over the (dead tree) literature on the topic: http://forums.construx.com/blogs/stevemcc/archive/2008/03/27...

An interesting data point he cites is the comparision between Lotus 123 version 3 and Excel 3.0: "Excel took 50 staff years to produce 649,000 lines of code. Lotus 123 took 260 staff years to produce 400,000 lines of code. [...] The difference in productivity between the two teams was more than a factor of 8." But then it will probably not be too difficult to find people who will tell you that 123 was still the better product.

[Edit: Another important point is that the these factors are between the best and the worst. The difference between the best and the median can be considerably smaller, so it is not that surprising not to see any programmers paid ten times the average.]

Re:On Wages by ytinas on Friday January 22, 2010 @04:17AM (#1069375)
The Lotus vs. Excel thing is more likely showing the fact that smaller teams can outproduce larger teams, since with smaller teams there is less coordination needed.
Re:On Wages by earl on Thursday January 21, 2010 @03:04PM (#1068179)
There is some empirical evidence in a book called Peopleware.

http://www.amazon.com/Peopleware-Productive-Projects-Teams-S...

I would paste some quotes but I lost my copy.