Rules for a better interview process

[BlasterBates]

His PhD project was allegedly programmatically generating java code. His result on the test was much worse than random chance, and it was a basic syntax multiple choice. Something didn't add up. Perhaps he answered all the questions wrongly as a protest.

But you´ve no idea whether he wouldn´t be a good programmer.

The guy has a PhD, well you can check that up. I´ve never ever met a PhD in Engineering/Maths who isn´t effing good. I do not believe that someone who is capable of understanding and coming with new Engineering ideas, solving partial differential equations won´t be able to write a decent program in Java. Did anyone check his spelling skills?

I used to work for a Software House and they sent me to do a Fortran project, in them days you didn´t have technical tests and althogh I´d played around with it and hadn´t really any real experience, but I didn´t have any problem doing the project.

I remember as a contractor being asked by the PM to develop the website, I hadn´t much a clue the new fangled internet stuff , because I´s never done anything with the internet before. but I just sorted out for him.

I usually do poorly on technical tests, but I consider it pretty anal really, and you can get into any new technology or language in a week or two.

[jmo21]

20 years work history does not mean you'll get someone with 20 years experience if they have been doing the same old crap year on year and not learning/progressing.

[bobspud]

I ask what technical books they have read recently. If they can't even name one, that's a bad sign IMO.

So the best bit of advice I will ever give you is this:

Some people aim to hire guys that read books...
What they should be doing is hiring the people that can write them instead...

There is a big difference in having the ability to follow instruction as opposed to improvise the application of freely learned knowledge.

[BlasterBates]

Over the last 20 years everyone I´ve worked with with very few exceptions are good. Programming is actually the simplest task a developer has to do.

Where I see problems it´s not their understanding of the technology, the key one is not testing their code, perhaps hurrying too much and another one is not understanding the requirements, i.e. going round in circles. I see tremendous variations in coding styles and practices but no correlation between being hip and being good, i.e. I have no problem with a C programmer who goes for a simple old fashioned style. In fact in most cases it is probably more appropriate. Indeed studies have not shown object oriented approaches to be any more productive than convention Cobol development. One of the best programmers I worked with just did everything in C. But the point was he understood the requirement immediately, the changes were in quick, and it worked

What PM´s need to think about who go overboard on patterns or exhaustive technical testing of new "talent" is why in spite of this, all the development is going off to India, China and Eastern Europe where they don´t do this kind of thing.

[doodab]

So the best bit of advice I will ever give you is this:

Some people aim to hire guys that read books...
What they should be doing is hiring the people that can write them instead...

There is a big difference in having the ability to follow instruction as opposed to improvise the application of freely learned knowledge.

Agreed. I have the same attitude to frameworks and tools. Lots of people know how to use them to do the stuff they already do, some people get into the internals of them and develop a deep understanding, extend them and so on.

[doodab]

But you´ve no idea whether he wouldn´t be a good programmer

So why take the risk? Hiring the wrong person can have a negative effect on the whole project. Much better to wait for someone who can convince you they are right than hire someone who isn't.

A technical test is just a piece of the puzzle, not the sole decision making criteria. The trouble is that most people don't have the flexibility you describe, especially younger guys who've studied to get into the industry cos it pays well but have no real passion for it as opposed to the older generation who are often largely self taught and got into it cos they love computers. Identifying the ones who do is quite a challenge.

[BrilloPad]

It was better 25 years ago. 5 minute chat - no HR. If you are no good then sad after 2 weeks.

[TheFaQQer]

Asking about forums, blogs etc is just to check the person is likely to be interested in what they are doing.

I have no interest in what I do for work outside work. It's dull and tedious the vast majority of the time, and I have enough to do during the day.

[DimPrawn]

I have no interest in what I do for work outside work. It's dull and tedious the vast majority of the time, and I have enough to do during the day.

Also, if you hire a team of people who live their lives on the latest and greatest cutting edge technologies, spend every waking hour on this blog and that forum, you end up with a team that don't deliver what the business requires, they just do what they want to do in the way they want to do it.

The only interest I have is just keeping up enough to keep working. Beyond that's it's an excrutiatingly dull profession to be in.

[d000hg]

The PhD Java programmer would probably have needed a week or two to get into it but I very much suspect he would producing very well structure code and very well tested code, because PhD´s on the whole are very thorough indeed at what they do.

If you're a scientist who needs to use programming as a means to an end, but you never had any education in programming, there is a big chance your code is terrible and a bigger chance you have no idea how to structure a project, use version control, etc. Being super smart doesn't imply you're a good programmer - probably they could be trained so that's a long-term hiring decision.