Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!unmvax!deimos.cis.ksu.edu!rutgers!apple!voder!berlioz!rgo
From: rgo@berlioz (Ronald Olshausen)
Newsgroups: comp.software-eng
Subject: Re: software engineers
Message-ID: <219@picasso.nsc.com>
Date: 15 May 89 22:57:39 GMT
References: <4105@ficc.uu.net> <4750@goofy.megatest.UUCP> <8309@venera.isi.edu> <8338@venera.isi.edu>
Reply-To: rgo@logic.nsc.com (Ronald Olshausen)
Organization: National Semiconductor, Santa Clara
Lines: 25

In article <8338@venera.isi.edu> raveling@venera.isi.edu (Paul Raveling) writes:
>In article <4750@goofy.megatest.UUCP> djones@megatest.UUCP (Dave Jones) writes:
>>From article <4105@ficc.uu.net>, by cliff@ficc.uu.net (cliff click):
>>> I've produced 30000 to 35000 lines of code/year for the past 3 years that end 
>>> up in income-producing products.
>>
>>Wow. Fifteen lines an hour, eight hours a day, 250 days a year. Love those
>>keyboard macros! :-)
>
>	For my own numbers, median production of debugged product-quality
>	code is around 200-300 lines per day when the day has no meetings
>	and the project's learning curve is past history.  Hacking

       
Those figures sound very high. But really, statistics regarding
Source Lines/Hour mean nothing, unless they're put in the
context of what the code does. For example, writing 5000
lines of code to do real-time embedded applications could be a whole
lot different than writing 5000 lines for a user-friendly
terminal interface. Depending on the degree of difficulty involved,
productivity, as measured in SLOC/hour, could vary by a 
factor of ten, or more.

Ron Olshausen
NSC-Santa Clara