Need help from developers

[viliusz]

Hi,

I'm doing a research for my diploma and I need developers to fill in a questionnaire about how they are using math in their daily jobs. Fill it here - http://bit.ly/be4Uz3

If there will be some1 interested in the results - will be able to share them later.

Thanks a mil

--
V.

[jwenting]

nope. All such "surveys" are completely useless. And noone in his right mind would click on some obscure link created by a first time poster, most such links give you malware infections.

[viliusz]

Well, then please explain how it is possible for first timer to get his survey answered? thanks.

[jwenting]

any such online "survey" has no value whatsoever as there's no control group and no way to know the demographics of respondents.
So getting it answered is not something you should worry about, as the results would be pointless anyway.

Rather get a decent course on statistics to learn why that is so, then maybe hire a professional firm to create a poll for you and get a relevant base of respondents if the thing is worth the money that costs.

[viliusz]

These are carefully chosen and approved questions. Control group - developers. Demographics - experience. This survey is showing very interesting results so far. So please stop mocking, thanks.

You should understand that it is not some company requested research (nor it can be sponsored enough to get decent stats company). Developers are that group that is very conservative and cannot be correctly surveyed by some stats company.

[jwenting]

You don't know if anyone who responds is what he says he is...
That's flaw #1.

[viliusz]

Nor you will know if it is true even when you will pay money for the answers :) Under this assumption all researches are failed ones. In this case - I believe that people are initially honest and results at the moment show that I'm right about that. Will see how it goes along the way.

[griswolf]

Interesting survey, apparently perfectly safe. Odd that the questions are in English, but the navigation is in (Finnish?). Also: There is at least one question that is probably misstated (the weighing problem does not state the kind of scale)

However, no 'self selected' survey has good statistical value

[viliusz]

Griswolf, thanks for the remarks. In weighting task - it is balanced scales are used. Navigation is in Lithuanian.

Yes it will not show good statistical results, but I'm trying to prove a main point that math helps developers and should be used during job interview.

[griswolf]

Well, I have a math degree (ancient), and I disagree that math questions are important for a programming job. In fact I have a very different interview philosophy. I think:

1. Make sure the person can do some of the task (not all, for a new hire). This is about being able to program, not do math. Logic is important, and being able to see all the cases.
2. Then have all the team interview this person to see if they are easy to like
3. If they are a good personality fit, and they pass step one, hire them.
4. If any team member thinks they are hard to get along with, don't hire them.

That's all.

You can make it more elaborate, but in the end, I think it is mostly about how the team feels... after being sure the person is at least an ok programmer. One thing that might be good, if your shop has a 'death march' mentality, is to be sure that the new hire can handle pressure with sufficient grace. But better to fix the process so that 'death march' doesn't usually happen in your shop.

[viliusz]

I agree with you completely, actually I myself am applying Joel Spolsky methods for hiring and interviewing. The difference is that he points out that there should be no math questions (he refers to them as "a-ha!" questions, sort of ones that can be answered once and if a person have solved one of these sometime ago - he will not think about solution, but will know it instead, thus leaving you with nothing).

Still what I'm trying to prove is that there should be at least 1-2 math tasks, just to understand the way a candidate is thinking or approaching the task, i.e. the process. E.g. few days ago I was in one of interviews where I have had to construct few functions - one for Fibonacci sequence and one for log() function use (that was used in some verification algorithms).

BTW this is also being used on Google side as well (puzzles where they are interested in a way you think and approach a problem).

You see, math, as a fundamental science - gives you tools to solve problems and it is up to you to demonstrate if you can use tools that were given you or you just trying to bang the wall until it falls (or not).

I have read so many comments where a software developers call companies stupid when during interview they have to solve math (and a job really involved lots of math, e.g. 3d development, etc).

[griswolf]

Agree that if the job actually involves math, then the process should check that the candidate can do it. Same for cooking or washing dishes... But... I think the process is really in two parts:

First part: Assure that the candidate has the education/skill/background to do the work. That is not part of what I think is the 'interview', but is done by taking a test of whatever kind(s) make sense for the job. Any HR person or contract test giver does this, maybe a phone interview or online if that is reasonable (probably not for cooking!). Also a very quick personality-fit check on the phone. These are 'cheap filter' and are used to decide who gets to the ...

Second part: The much more expensive interview process. This is almost entirely about team/personality fit, and almost nothing about skill. It is more expensive because it takes developer time and may involve travel for the candidate. And it is done face to face because team members are going to interact face to face, and because we 'read' people much better when we see them. Also a final chance to filter out someone who is much better on the phone than in person.

But this has become not at all about computer science, even though it is interesting...