Brainbench

[Daniel]

My employer has asked me to take the C exam from brainbench.com.

Does anybody have an idea of what kinda questions do they have ? How
difficult it is ?

Any help would be appreciated.

Thanks

 

[Christopher Layne]

My employer has asked me to take the C exam from brainbench.com.

Does anybody have an idea of what kinda questions do they have ? How
difficult it is ?

Any help would be appreciated.

Thanks

Yes. It's actually Richard Healthfield, Keith Thompson, and Chris Torek on the
other side issuing the questions as you take the test, live.

But no, I've never taken it. If there's no time limit on the answers, I don't
see why one couldn't ace the test.

 

[Robert Gamble]

My employer has asked me to take the C exam from brainbench.com.

Does anybody have an idea of what kinda questions do they have ? How
difficult it is ?

I haven't taken the actual test (and don't plan to) but you can take
the "Sample Test" as many times as you want and judge for yourself.
The sample test contains 5 random questions from what appears to be a
pool of about 10 questions. For the most part the questions were
good, clear, and well-balanced although a couple of them them assume
knowledge of non-standard functions and compiler extensions. For
example, one question assumed familiarity with the strdup function (a
POSIX function, not ISO C, although available on many
implementations), while another one required you to assume that
pointer arithmetic on a void pointer behaved the same as on a char
pointer (which is a common extension but not legal in ISO C). While
it is obviously MUCH better than the Expert Rating C exam, and seems
to have potential, it obviously isn't limited to ISO C, a fact which
should be spelled out in the test description.

Robert Gamble

 

[fdmfdmfdm]

I haven't taken the actual test (and don't plan to) but you can take
the "Sample Test" as many times as you want and judge for yourself.
The sample test contains 5 random questions from what appears to be a
pool of about 10 questions. For the most part the questions were
good, clear, and well-balanced although a couple of them them assume
knowledge of non-standard functions and compiler extensions. For
example, one question assumed familiarity with the strdup function (a
POSIX function, not ISO C, although available on many
implementations), while another one required you to assume that
pointer arithmetic on a void pointer behaved the same as on a char
pointer (which is a common extension but not legal in ISO C). While
it is obviously MUCH better than the Expert Rating C exam, and seems
to have potential, it obviously isn't limited to ISO C, a fact which
should be spelled out in the test description.

Robert Gamble

I took it once. Very stupid. Nothing really helpful for daily work.

 

[Ernie Wright]

My employer has asked me to take the C exam from brainbench.com.

Does anybody have an idea of what kinda questions do they have ? How
difficult it is ?

I took it a couple of months ago. I did pretty well, although not as
well as some here reportedly have.

http://www.brainbench.com/xml/bb/transcript/public/viewtranscript.xml?pid=6417683

The test consists of 40 multiple-choice questions, with a 3-minute time
limit for each. If you need to take a break, you can pause the test
*between* questions (not during a question--there's a "pause after this
question" button).

Unless you're a guru, the test will seem fairly difficult, regardless of
your level of knowledge. The difficulty level of the questions depends
on how well you've answered previous questions. The test rises to your
level of incompetence.

There are several questions of the form "What is the output of the
following?", with a 10- or 15-line program or code fragment. You might
be tempted to try to copy-paste-compile-run to find out, but the text of
the page isn't copyable.

There's no real math, and I don't recall anything from outside the
language standard, so nothing about Win32 or POSIX or networking or
makefiles, or other extracurricular topics.

Before taking the test, I re-read my copies of K&R2 and Peter van der
Linden's Expert C Programming,

http://www.amazon.com/C-Programming-Language-2nd/dp/0131103628
http://www.amazon.com/Expert-Programming-Peter-van-Linden/dp/0131774298

and the C FAQ,

http://c-faq.com/

I also took the sample test. In addition to giving you the flavor of
the real thing, this gives you a chance to test your Web browser and
Internet connection.

- Ernie http://home.comcast.net/~erniew

 

[Richard Heathfield]

Daniel said:

My employer has asked me to take the C exam from brainbench.com.

Does anybody have an idea of what kinda questions do they have ? How
difficult it is ?

Of more concern to me is its quality.

A few years ago, a number of experts here in comp.lang.c took the test
and compared notes afterwards (back in the days when it was called
Tekmetrics, or something like that). We discovered an inverse
correlation between one's score on the test and one's level of C
knowledge as recognised by others here in comp.lang.c. For example, if
I am not mistaken I scored higher than LK and CT - which is patently
ludicrous.

It is possible that things have improved somewhat, as (I vaguely recall
that) Martin Ambuhl worked with Brainbench on fixing the quiz's grosser
stupidities, but we have not again conducted such a wide-ranging review
to the best of my knowledge, so the question remains open.

 

[Beej]

Of more concern to me is its quality.

This whole thing has me thinking...

Do you suppose it would be useful to have an open C test written by
the community that anyone could access? (With answers, of course.)

I'm not thinking of this as a competitor to this brainbench* thing;
more of as a learning tool.

Does such a thing already exist?

-Beej

 

[Martin Ambuhl]

It is possible that things have improved somewhat, as (I vaguely recall
that) Martin Ambuhl worked with Brainbench on fixing the quiz's grosser
stupidities, but we have not again conducted such a wide-ranging review
to the best of my knowledge, so the question remains open.

Yes, indeed, they did pay me to modify questions and answers so they
would conform to standard C, not rely on undefined or implementation
defined behavior, and actually relate to C as it might actually be used.
Unfortunately for the OP's question, I have never looked at the use to
which my changes were put.

I shudder to consider the consequences of your claim that

A few years ago, a number of experts here in comp.lang.c took the test
and compared notes afterwards (back in the days when it was called
Tekmetrics, or something like that). We discovered an inverse
correlation between one's score on the test and one's level of C
knowledge as recognised by others here in comp.lang.c.

My invitation to recast the questions and answers was based in part on
the fact that Brainbench (with a different name then) claimed that mine
was the highest score registered on the test. That pretty much puts me
out of the running as someone to answer questions here.

 

[Richard Heathfield]

Beej said:

This whole thing has me thinking...

Do you suppose it would be useful to have an open C test written by
the community that anyone could access? (With answers, of course.)

I'm not thinking of this as a competitor to this brainbench* thing;
more of as a learning tool.

Does such a thing already exist?

How about the comp.lang.c FAQ at http://c-faq.com ?

 

[Richard Heathfield]

Martin Ambuhl said:

Yes, indeed, they did pay me to modify questions and answers so they
would conform to standard C, not rely on undefined or implementation
defined behavior, and actually relate to C as it might actually be
used.

Thank you for confirming that, Martin. Now then...

Unfortunately for the OP's question, I have never looked at the use
to which my changes were put.

That is also interesting. Perhaps we should give them another round (if
it's still free - I see no reason why we should pay them for the honour
of doing their testing for them).

My invitation to recast the questions and answers was based in part on
the fact that Brainbench (with a different name then) claimed that
mine
was the highest score registered on the test. That pretty much puts
me out of the running as someone to answer questions here.

Well, Martin, dem's da breaks.

Seriously, I have two replies to that:

1) an inverse correlation doesn't mean an inverse law.
2) if everyone taking the test was saying how broken it was, and if they
accepted that sufficiently that they were prepared to hire someone to
fix it, why would they think that a high test score was meaningful?

 

[Christopher Layne]

Yes, indeed, they did pay me to modify questions and answers so they
would conform to standard C, not rely on undefined or implementation
defined behavior, and actually relate to C as it might actually be used.
Unfortunately for the OP's question, I have never looked at the use to
which my changes were put.

This is actually cool to see corporate companies embracing correctness rather
than mediocrity.

 

[Richard Heathfield]

Christopher Layne said:

This is actually cool to see corporate companies embracing correctness
rather than mediocrity.

Well, what's actually cool is that they entered into the debate, and
were prepared to recognise that they were wrong and fork out to get
some remedial work done. We have yet to determine whether they embraced
correctness to the extent of putting up a set of decent questions with
correct answers.

 

[Oliver Wong]

But no, I've never taken it. If there's no time limit on the answers, I
don't
see why one couldn't ace the test.

There is a time limit.

- Oliver

 

[Beej]

How about the comp.lang.c FAQ athttp://c-faq.com?

Well, the C faq is great, but it doesn't really present the material
in the same way as a test does. Peoples' brains operate in different
ways reading one or the other.

-Beej

 

[Daniel]

I wrote the exam and it was okay types and not extremely difficult but
the images took some time to load up and I lost some time.


Do you think a poor brainbench score means a bad programmer and vice
versa.

 

[Ian Collins]

I wrote the exam and it was okay types and not extremely difficult but
the images took some time to load up and I lost some time.

That's not very considerate of them.

Do you think a poor brainbench score means a bad programmer and vice
versa.

There's a lot more to being a good programmer than being able to score
well it tests.

 

[Richard Heathfield]

Daniel said:

I wrote the exam and it was okay types and not extremely difficult but
the images took some time to load up and I lost some time.

Draw your own conclusions about how much thought has gone into the test.

Do you think a poor brainbench score means a bad programmer and vice
versa.

No. I think there is little or no relation between your score on that
test and your skill in the subject being tested.