Expert developers

[luckykaa]

I keep coming across articles about things like why good developers are worth 10 times as much as a mere mortal.

Every time I read an article by a developer for the layman it seems to come across as totally up himself and congratulating himself on how good he is at what is actually quite a small amount of the job.  Nobody seems to be quite so self congratulatory if they can come up with an adequate solution to a problem that's easily maintainable and well documented.  90% of the time we can cope with inefficiency. 

Clearly I'm lacking the arrogance needed to be a really great programmer.  I should take tips from Mel Kaye. 

Comments

[raygungothic.livejournal.com]

That's a really silly article. I wouldn't hire that person.

[masterofapath.livejournal.com]

Interesting you both seem to (from my perspective) be misreading the intent of the article.

Some reasonable points are made about good developers:
- cross language ability (although this neglects the
complexity of some APIs frameworks etc which can take a
disgustingly long time to learn).
- The importance of writting readable, maintainable, and
even documented code.
- The importance of choosing good tools, and caring about
ones work.
- That throwing more people at a problem doesn't always
mean that proportionally more work gets done.

The article makes the mistake of mentioning specific languages showing the authors personal bias.

I think where the article really fails though is in the use of the term 'expert developers' which makes them sound like an elitist group whose ranks are beyond the dreams of mere mortals. In fact the best developers are always keen to share information and advice, and help a team be more than the sum of its parts.

I've worked with some really good developers and some really poor ones. What it largely boils down to is a mixture of experience, and attitude. I would happily take 1 of the best over 10 of the worst in a team. Is 1 of the best better than 10 average ones? that's one I wouldn't like to try and answer.

The underlying point I think they're trying to make is that a lot of companies fail to recruit new developers as effectively as they could, and sometimes don't do enough to hang on to the good ones they have.

[raygungothic.livejournal.com]

I didn't read it as being about the good points you mentioned, because the author refers to all those points as if they're standard wisdom anyway (and most are) - he seemed to be much more about "I and my kind are leet! and tower over those puny mortals who don't use my favourite language". If the article seemed to be about hiring people who document their code, that would be another matter...

[luckykaa.livejournal.com]

Hire people who write self documenting code instead. The bit I'm working on right now is a delight.

Is there a function that gets the position of a menu item? Why, yes there is; Menu::GetItemPosition(int Index) seems to do that.

[raygungothic.livejournal.com]

Yeah, I like working that way - but I don't think it's enough.

[luckykaa.livejournal.com]

The basic advice is okay. I just find the tone a bit irritating.

I just don't believe that an expert programmer is 10 times more efficient. More efficient that 10 programmers on the same task, certainly but then we get to the whole "digging a ditch" argument from Mythical Man Month. A lot of the time you really want to match the developer to the task. the good ones do the hard stuff, the not so good ones do the simple stuff.

There's an attitude of "Your product is nothing without me, worship me!". True, but the same applies to sales and distribution as well.

It just all seems a bit prima-donna-ish

[raygungothic.livejournal.com]

There are so many different kinds of "good". Some tasks are conceptually simple, but need heavy wizardry to make the computer do it - only experience will do. Others are all in the concept and not terribly difficult to implement - raw brainpower is much more important than experience here. Simple one-size-fits-all rules probably don't work. (Though "the MMM" seems to have the whole "division of labour" aspect pretty much sewn up, from my cursory examination)

The Cult of Sales is still much more dangerous than that attitude amongst programmers, though.

[ingaborg.livejournal.com]

A really good, effective programmer IS worth 10 ineffective programmers, purely measured by how long it takes to get a usable end product. The trick is how to find this out before you hire them.