Managing Software Engineers, by Philip Greenspun (philg@mit.edu)


anteriorpróximo

I think I have to agree with most of the comments about long hours and such. I am currently a junior computer science major at Texas A&M University, and I have had 2 internships, and a research grant (at TAMU). My experience so far has been that I like to work longer than 8 hours when I don't really have anything else better to do, but 70+ hours wouldn't be good. MAYBE 55. As a young single guy who loves computers and technology and electronics and math, etc., I like to accomplish things at work. However, I also like to go back to my place and learn new stuff, so I can accomplish more. If I don't get home until 7-8 pm, I don't really have adequate time to spend with my family or friends, or doing anything that I love to do outside of work. That kind of thing would just make me resent my job, and look for another one. And remember: all this from the perspective of a guy who the author says would be happy to work 70+ hours.

-- Ben Collins, November 7, 2000
The best way to alienate women from IT is to continue spreading this myth about the 70 hour weeks. The only people who truly believe that they produce better results by working more are people that Women don't usually fall into either of these categories! Women generally have a much more balanced life, understand the need to do basic things in life, i.e. grocery shopping, cooking, dishes, hanging out with friends, hobbies and yes, raising their kids.

The funny thing about a balanced lifestyle is that although one works less, the work one does is better. And because one is able to go home in the evening (at 6:00!), one has time to reflect upon the work that one has done with some perspective. And consider what goals one has for the next day. And if the programming from that day was worth anything.

I suspect that the reason that I encounter so much buggy, lame code in the ACS is that the code was developed exactly the way you are suggesting in your article.

-- Sarah Arnold, November 7, 2000

This article is seriously flawed.

Written in a goading style, it ignores much of the ongoing research into software productivity, (See Capers Jones, Barry Boehm, Steve Mc Connell, et al.). The recurring truth is that in successful software projects, a mere 20% of the effort is spent in coding. This article pretends that coding is the only thing that matters.

This is very flattering to those who code, but ignores reality. Coders may read this and say, "Yeah, finally someone is speaking for us!" But anyone who has occasionally looked up from the monitor to take a look around the rest of the office knows that this view is extremely myopic. Programmers aren't the only people in the organization. And any software company that cannot make use of the average programmer while slavishly worshiping and abetting the whims of the often overrated and inconsistent superstar is bound to fail in the long run. The superstar will sometimes do something brilliant that saves everyone's ass. But planning for this, and staking the company on such erratic behavior is just plain stupid.

Rather, the key to consistent (rather than haphazard) software success is planning, coordination, communication, and perseverance. To suggest otherwise is special pleading and vanity of the sort that leads to hubris. And the gods always punish hubris.

Fred Williams

-- Fred Williams, November 7, 2000

I see a lot of complaining about the 70+ hour work week issue. I think most of the people here miss the point of what "great" programmer means. You can be one of the best writers in the world but if you write only a paragraph a day you'll never be "great". You may technically be one of the best artists of your era but if you don't immerse yourself in it day and night you'll never be "great".

Sorry, you may be a genius problem solver and coder but part of being a "great" is giving yourself over to the art. It DOES mean giving up your free time, it DOES mean maybe not having children, it DOES mean maybe burning out, it DOES mean sleepless nights telnet'ing in from home to work on the problem you just left. Being a programmer at the "great" level, building a major system with the purity of vision you need to be "great" means no distractions and getting the core of the code down before the "Man(ager) from Porlock" comes along to snatch your vision away.

Putting it more succintly, arsdigita gratia artdigitas =)

-- Mark Mills, November 7, 2000


anteriorpróximo