views:

83

answers:

7

Let me explain what I am trying to ask

If we assume that one programmer has 20 years programming experience in general (means in many languages) and he has 5 years experience in Java language, of course, these 20 years are including 5 years. Additionally, he has 8 years as designer of design architecture & system as well as he has 5 years experience as tester. So the question is that can I mathematically evaluate the overall experience of this programmer in term of all these parameters (experience of general programming, experience in particular language, experience of design architecture, and testing experience) I hope that the question is clear

A: 

How many lines of code has he written?

How many lines of code has he debugged and fixed?

Loadmaster
How do you define lines of code? (Don't answer, that's a rhetorical question)
Andy Lester
Well, that's easy: #define lines_of_code
Santiago Lezica
I'm sorry, you didn't ask what language it was to be written in. As it turns out, I was using Perl, where you have the `constant` pragma. Don't feel bad. My programmer test is a difficult one. I wish you luck in your job search. Let me show you the door. :-)
Andy Lester
A: 

How can evaluate the overall experience of a cook? He's studied two types of cuisines, worked for 5 years in a restaurant, then moved on to another restaurant for another 10 years. He knows 176 different types of ingredients, and 21 kitchen tools.

I don't think you can measure that with numbers.

Santiago Lezica
+1  A: 

Have a look at the Joel's How do You Compensate Programmers

He provides a method of evaluating and grading programmers based on a number of different factors.

Paul Alexander
+3  A: 

Can I mathematically evaluate the overall experience of this programmer in term of all these parameters

Absolutely not.

Further, you don't want to, because you will likely use it as a crutch and blind yourself to all the other factors that go into hiring a good programmer.

If hiring programmers was as simple as plugging numbers into a formula, we could let trained monkeys and/or the Human Resources department do it. It's not, and you shouldn't.

Andy Lester
A: 

No. You can write a million lines of crappy code or a thousand lines of good code. You can spend years in the industry, but do a poor job, or spend little time in it and do a great job. So, this is a poor measure. The best way to evaluate someone for a position is to see how they tackle a variety of problems that they've never seen before, as well as obtaining feedback from prior colleagues and other references.

Michael Aaron Safyan
A: 

Major practical metrics of development skills and software quality include

  • Time (ability to deliver on schedule)
  • Cost (ability to deliver within budget)
  • Quality (ability to deliver a worthwhile product)
  • Size/complexity (ability to architect systems with sane power-to-weight ratios)

Focusing on any one to the exclusion of the others is a danger. Code metrics (and regular reviews!) can help establish meaningful baselines for measuring software quality.

Joe
A: 

From a stats point of view you could come out with something that would probably give you a good indication if you had a large enough sample with the right data.

MykC