The Joel on Software Discussion Group (CLOSED)

A place to discuss Joel on Software. Now closed.

This community works best when people use their real names. Please register for a free account.

Other Groups:
Joel on Software
Business of Software
Design of Software (CLOSED)
.NET Questions (CLOSED)
TechInterview.org
CityDesk
FogBugz
Fog Creek Copilot


The Old Forum


Your hosts:
Albert D. Kallal
Li-Fan Chen
Stephen Jones

95% of Programmers... (a joke)

Linus Torvalds: "95% of Programmers consider themselves in the top 5%".

Shlomi Fish' Corollary: "95% of Programmers consider 95% of the code they did not write, in the bottom 5%."

The other 5% have read Joel... ;-)
Shlomi Fish Send private email
Friday, May 13, 2005
 
 
Personally I don't see how you'd measure how good somebody is at coding. Past a certain point everything about being a 'good coder' seems to become subjective.
Colm O'Connor Send private email
Friday, May 13, 2005
 
 
Added to which, I doubt even Linus Torvalds has seen a representative cross-section of all coders.
Colm O'Connor Send private email
Friday, May 13, 2005
 
 
Nothing like a good dose of literalness to suck the life out of a joke...
Kero Send private email
Friday, May 13, 2005
 
 
Well, it's been shown that, generally, the weakest people tend to overestimate their skills the most, with the best often underestimating: http://www.phule.net/mirrors/unskilled-and-unaware.html
Scott McMurray Send private email
Friday, May 13, 2005
 
 
Personally I don't see how you'd measure how good somebody is at coding. Past a certain point everything about being a 'good coder' seems to become subjective.

=========

To steal a line from Chevy Chase in CaddyShack, "By Height."
Mr. Obvious
Friday, May 13, 2005
 
 
i agree with the rest "read joel".

i think good people who care about what they do, or are interested in something, some here.

the rest go to slashdot :)

el
Lemon Obrien III Send private email
Friday, May 13, 2005
 
 
Scott: what's dangerous about your (correct) observation is that using it as a heuristic completely screws up exceptions to the rule.

If you assume that people who are superconfident about their code are actually not great coders, but just tend to overestimate their abilities, then you end up punishing the best group of all: the group of great coders who *accurately* estimate their abilities.

Similarly, if you assume that humble coders are actually good programmers with low self-esteem, then you end up rewarding a bad group: the poor coders who can (at least) see their own lack of talent.

Not that you were advocating using it as a heuristic... it's just an observation about your point.
Tail of the "g"
Friday, May 13, 2005
 
 
what about those of use who are in the bottom 5% but compensate by charging ridiculously high rates so others think we are in the top 5%.
Tom Vu
Friday, May 13, 2005
 
 
"what about those of use who are in the bottom 5% but compensate by charging ridiculously high rates so others think we are in the top 5%."

That's called genius :-)
Bart Park
Friday, May 13, 2005
 
 
In my book Linus Torvalds is a pretty avergage programmer.
Zyzyphus
Saturday, May 14, 2005
 
 
In his book you're a pretty avergage speller. :-)
Aristides
Saturday, May 14, 2005
 
 
Linus Torvalds may not be a very good programmer. (I don't know, I haven't looked at his code) But he's one of the best software managers on the planet, and also proved the inspiration for the highly claimed "Bazzar style of Software Management".

I think what he says is basically right. Basically, most programmers really think their code flies. I no longer consider myself in the top 5% of programmers, but I think I'm in the top 50% (I've seen some pretty bad programmers in my days).

As for code, I no longer see a code as bad as long as it's more or less working, but just as "not good enough". I used to refactor other people's code quite a lot, and it was quite enjoyable.

Another thing is that I thought that I seem to be getting a certain kind of sixth sense for code, that I can tell how good or bad it, just by looking at it. Then I heard a friend's account, on how some good that she initially believed to be bad, turned out to be very maintainable, while code that seemed very elegant, was an unmaintainable mess of templates and classes. (C++... %-)).

Ben Collins-Sussman said in his interview with me, that his code doesn't fly: it's not bad either, but it's not particularly good. Thing is the code of Subversion, where he works on (except the working copy library, which is a mess, but for a good reason) is very nice. Maybe he underestimates himself.

Adrian Ettlinger OTOH (yet another one with an interview I conducted) wrote a code with some bad design decisions which I have to workaround then, but when I told him that "I was not too impressed from the quality of your code". He admitted he was a not-so-good coder.

One thing I've noticed is that some people are better programmers after 1 year, while not having finished elementary school yet, than others who've been programming for 10 years and have a Ph.D. in Computer Science. I can't really explain it, but have some theories. OTOH, some people actually get better in time, and do their best to learn from their mistakes.
Shlomi Fish Send private email
Saturday, May 14, 2005
 
 
>One thing I've noticed is that some people are better programmers after 1 year, while not having finished elementary school yet, than others who've been programming for 10 years and have a Ph.D. in Computer Science.


After getting Master Degree and years of experience, I tend to believe that Comp Sci education has nothing to do with becoming a good software developer.
The only two benefits I got from there are meeting other people who whant to become software developers and getting right checkbox on education section of my resume.
Eugene Send private email
Sunday, May 15, 2005
 
 
Eugene, I share a lot of your sentiments about computer science education. However, I believe that a computer science and/or experience can help good programmers become better. I am a much better programmer today than I was when I started the Technion, and then I was still pretty good. A lot of it is due to open-source experience, etc. but school also helped a lot.

A few things that bother me about computer science education:

1. There should be a few mandatory courses in the history of computers. I talked with an M.Sc. student who was only familiar with Dijkstra because of his graph algorithm, and did not know who he was.

2. I believe CS students should learn Digital Design and Computer Architectures. This is not only so they'll be able to work in VHDL, Verilog, etc. but also so they'll understand the tradeoffs involved in designing efficient code (in any language), how threads and processes work, etc.

Many CS graduates think that programs are written for Pie-in-the-sky Turing Machines, and don't understand the iron below it.

3. There should be necessary courses that cover all modern and useful programming paradigms (OSes, RDBMSes, scripting languages, OOP, FP, etc.). Many people who decide to become theoretical computer scientists, then write sub-optimal code due to lack of knowledge in modern paradigms. (Re-inventing a square wheel).
Shlomi Fish Send private email
Monday, May 16, 2005
 
 

This topic is archived. No further replies will be accepted.

Other recent topics Other recent topics
 
Powered by FogBugz