Hacker News new | ask | show | jobs
by jrockway 5745 days ago
Oh, I agree 100%. The software industry barely understands software (see also: offshoring), so I simply don't expect a non-software company to even be in the ballpark.

Big companies assume software projects will be failures, and treat their employees accordingly. Then they get what they expect.

(Hire a bunch of good programmers with good management, and you can get amazingly reliable software from a team of two. Hire a bunch of bad programmers, though, and a team of 100 produces something worse than most kids' intro-to-java app. But nobody but the best programmers understand this, and it looks better to the business to pay 10 people each $60,000 a year instead of paying 2 people $200,000 a year. When a big company "gets this", their software becomes a lot better.)
1 comments
docgnome 5745 days ago
It's obvious that a small group of great programmers beats a large group of bad ones. This is pointed out constantly on HN and in other programming communities. I'm curious though if anyone has any hard data to support this, beyond anecdotal evidence. (I'm not disagreeing with the sentiment, I agree completely. I'm just interested in seeing any studies done on the subject. I imagine this applies to many other fields as well.)
link
gaius 5745 days ago
It's obvious that a small group of great programmers beats a large group of bad ones

The question is, tho', "at what"? Oracle's product offering is VAST. And most of it is technically quite simple; it just does an awful lot. You can't use clever Lisp macros to do stuff like this: http://c2.com/cgi/wiki?WhyIsPayrollHard You just have to chuck people at it.

link
tamersalama 5745 days ago
Data are hard to come by in the industry (compared to, say, manufacturing where science branches were spawned based on analysis, observations).

From where I stand, the cost of poor quality software goes beyond the production. It's more in maintenance, licensing, lost opportunities, and poor adaptability to changing business needs.

link
docgnome 5744 days ago
Right. Which is why we need a long term study. To follow a product from conception to production to end of life.
link