[RESOLVED] JUST FOR GRINS (ZCE Practice Test?)

fuzzy1

Hey All,
Just for grins, I've taken one of the practice tests for the ZCE exam.
For my money, it seems the test is rooted in a bunch of silliness?
Take, FOR EXAMPLE the following question...

Question:
What is the best approach for converting this string:
[INDENT]$string = "a=10&b[]=20&c=30&d=40+50";[/INDENT]
Into this array?
[INDENT]
array(4) {
["a"]=>
string(2) "10"
["b"]=>
array(1) {
[0]=>
string(2) "20"
}
["c"]=>
string(2) "30"
["d"]=>
string(5) "40 50"
}[/INDENT]

POSSIBLE ANSWERS PROVIDED: (choose 1)
[INDENT]

Write a parser completely by hand, it's the only way to make sure it's 100% accurate

Use the parse_str() function to translate it to an array()

Pass the variable to another PHP script via an HTTP GET request and return the array as a serialized variable

Just call unserialize() to translate it to an array()

Write a string parser using strtok() and unserialize() to convert it to an array

[/INDENT]

However, it seems to me (and this is where I'm hoping someone might show me the error in my thinking...) but it seems to me that the BEST answer would be "F"

[INDENT]
F. NONE OF THE ABOVE (this is a trick question comprised of nonsensical GIBBERISH!!!)[/INDENT]

That, or option "A" but to my eyes, "this array?" isn't an array at all. But then... maybe that just goes to show how far I am from realistically considering ZCE???:p

laserlight

That, or option "A" but to my eyes, "this array?" isn't an array at all. But then... maybe that just goes to show how far I am from realistically considering ZCE???

It is also a test of whether you recognise that the "GIBBERISH" is a possible output of [man]var_dump/man. Consider:

<?php
$string = "a=10&b[]=20&c=30&d=40+50";
parse_str($string, $array);
var_dump($array);
?>

Consequently, B is probably the best answer.

fuzzy1

WELL! Is my face red?:o

just goes to show how far I am from realistically considering ZCE!!! :eek:

Thanks lazerlight!

Kudose

A question very similar is on the exam, and yes, the answer was the equivalent of B.

fuzzy1

Because I didn't get enough egg on my face yesterday...

I offer the following as an example of what I still think are
largely a lot of silly questions on the ZCE practice exams.

If you wanted a variable containing the letters A through Z,
that allowed you to access each letter independently,
which of the following approaches could you use?
(Choose 3)
a)$str = "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
b)range('A','Z')
c)explode("","A...Z")
d)You would use the ALPHA_ARRAY constant
e)None of the above

A & B WORK, C doesn't, I can't find anything on D "ALPHA_ARRAY", AND E, negates the whole thing right???

<?PHP
$alpha =range('A', 'Z'); 
$str = "ABCDEFGHIJKLMNOPQRSTUVWXYZ"; 
echo $str{0};
echo $alpha{1};
echo $str{2};
?>

returns 'ABC'
Now, either there is some obscurity called "ALPHA_ARRAY",
or I'm completely misunderstanding the question.
or maybe -- just maybe -- the correct response is A, B, & E :queasy:???
Anyone???

Kudose

The answer is A.

If you read carefully, which is what they are going for, it says

If you wanted a variable

So B and C are out automatically ... D is a trick and Yes, E nullifies the question. So A is the correct answer.

i.e.

echo $str[2]; // prints C

fuzzy1

Hmmm...

laserlight

Kudose;10888147 wrote:
The answer is A.

If you read carefully, which is what they are going for, it says

So B and C are out automatically ... D is a trick and Yes, E nullifies the question. So A is the correct answer.

i.e.

echo $str[2]; // prints C

The second part of that clause ("containing the letters A through Z") is just as important, since range('A', 'Z') returns an array whose elements contain the letters A through Z.

EDIT:
Oh wait, it says "choose 3" rather than "choose the best answer", in which case B should also be a correct answer (since such an array would transitively contain the letters A through Z). Unfortunately, I am also not able to determine the third correct option.

fuzzy1

YES, Thank you lazerlight!!!

If you read carefully, which is what they are going for, it says
CHOOSE 3

and it also says

which of the following approaches could you use?

and clearly, "range('A', 'Z');" is an approach which you could use e.g.

$alpha =range('A', 'Z');
 // wherein $alpha IS a VARIABLE employing "approach" b) "range('A','Z')" No?
echo $alpha{1}; //prints "B"

fuzzy1

With fleeting pangs of vindication...
I offer the following for your consideration as another example of a POTENTIALLY SILLY ZCE question .

What combination of boolean values for $a, $b, $c, and $d will result
in the variable $number being equal to 3?

<?php

$a = null;
$b = null;
$c = null;
$d = null;

if($a && !$b) {
  if(!!$c && !$d) {
    if($d && ($a || $c)) {
      if(!$d && $b) {
        $number = 1;
      } else {
        $number = 2;
      }
    } else {
      $number = 3;
    }
  } else {
    $number = 4;
  }
} else {
  $number = 5;
}
?>

While there IS a correct answer...
In the context of the question the code block provided is represented as being
generally viable, however at least 40% of the code is wholly untenable.

In all permutations of "true or false" for $a, $b, $c, and $d,
only 4 of 16 will satisfy the first condition, and only one of those will satisfy the second condition
(e.g. the correct answer to an otherwise flawed question/problem).

There are NO possible permutations that will satisfy either the third or fourth conditions.

Now if that isn't silly enough, be honest and ask yourself if you think that you,
or anyone else short of some peculiarly gifted savant (think Rain Man here)
could be realistically expected to "see" the permutations in their heads
(or scribble them on a scrap of paper even -- assuming that is allowed)
and come up with a correct answer in under -- say -- 5 minutes time?

Missing the obvious var_dump component of the answer to the question presented in my initial post (duh),
by the time I got to said question I was already kind of ticked-off at the "silliness" of many of the test
questions that had preceded it and perhaps rushed to dismiss it as silly as well.

Nonetheless ... assuming the practice tests are representative of the actual ZCE Test,
I stand by my original assertion that the ZCE test is largely "rooted in a bunch of silliness."
Anyone???

Kudose

haha, I need to read carefully.

I disagree that the test is based on silliness though.

laserlight

fuzzy1 wrote:
Nonetheless ... assuming the practice tests are representative of the actual ZCE Test,
I stand by my original assertion that the ZCE test is largely "rooted in a bunch of silliness."
Anyone???

I agree that this third sample is an example of a bad test question. Unfortunately, it is far easier to come up with silly puzzles than to come up with questions that genuinely test PHP knowledge in the light of actual programming ability.

Kudose wrote:
I disagree that the test is based on silliness though.

However, if I remember correctly, you have a vested interest in debunking any claim that the ZCE tests are "based on silliness" :p

Kudose

laserlight;10888234 wrote:
However, if I remember correctly, you have a vested interest in debunking any claim that the ZCE tests are "based on silliness" :p

😃

I will admit that some of it is silly, but most of the the questions still hold validity. Also, the ZCE practice exams from whoever ... I can't remember but you can buy them from Zend ... are similar to the actual exam, but are still very different. From what I saw on the exam, the "silliness" was vastly eliminated.

As a side note, I can't imagine they would develop and deploy a real test that people didn't believe in.

Weedpacket

Kudose wrote:
I will admit that some of it is silly, but most of the the questions still hold validity. Also, the ZCE practice exams from whoever ... I can't remember but you can buy them from Zend ... are similar to the actual exam, but are still very different. From what I saw on the exam, the "silliness" was vastly eliminated.

I would agree with this, based on my recollection. I remember going through the practices and realising that in some cases even the questions were wrong, probably because of version changes; the practice exams don't necessarily track as closely as the real exam. But if you can get through the practice without grief then you're well-prepared for the real thing (which is the whole point of the practice, really).

But still:

fuzzy1 wrote:
There are NO possible permutations that will satisfy either the third or fourth conditions.

So? Are you telling me you've never written a test with the comment

// This should never happen?

Now if that isn't silly enough, be honest and ask yourself if you think that you,
or anyone else short of some peculiarly gifted savant (think Rain Man here)
could be realistically expected to "see" the permutations in their heads

I didn't have any trouble (signed: Raymond Babbitt).

Not that it's necessary to answer that question.

It's a skill that comes with experience (translation: I know what I'm doing). $number should equal 3? Okay. The first if() statement that guards that statement requires that $a be true and $b be false (that's two variables set right there, and I'm only on the first test). The second if() statement requires that $c be true and $d be false. Ooh, that's all four variables set. Of course, there's one more test ($d && ($a || $c)) and that has to fail. Which it will if $d is false. And it is. Who needs to arse about with every possible assignment, and every single if statement just to answer that question?

Obviously, an exam question is not going to have eight or nine lines of arbitrary code when just assigning a value to $number does everything the question requires; and the tests are not going to rely on arbitrary function calls that will do no-one knows what, unless they're native, in which case the question becomes one of understanding what those functions do. But that's not what the question is checking. The question is checking that you are capable of logical thought.

laserlight

Weedpacket wrote:
But that's not what the question is checking. The question is checking that you are capable of logical thought.

My first impression when reading that question is that it is testing whether the reader knows about the conversion of null to false when in a boolean context. Sure, once you know that, you can solve the question using logical thought, but you can also solve the question by mechanically applying the rules of boolean algebra, especially since the variable names are not descriptive. That seems rather pointless in a ZCE exam.

On the other hand, if the ZCE exam only tests such language features directly, it would not actually test programming ability. Perhaps one can say that the ability to reduce a problem to boolean algebra is indeed a sign of programming ability, in which case I would concede that the question is not silly - but a glance at the programming ability of certain academics shows that academic knowledge of programming related subjects does not necessarily correspond to programming ability in a real world context, yet that is precisely what the ZCE exam intends to test.

Weedpacket

The problem with testing "in a real world context" is that there is all this real-world context that would have to be loaded into the question. The MSCE exam questions come with four or five paragraphs of waffle, and 95% of the work consists of decrypting the question (with an answer that tends to be of the Lego-block "use Foo.NET and Microsoft Bar" variety).

laserlight wrote:
but you can also solve the question by mechanically applying the rules of boolean algebra

Many many things can be solved by mechanically applying rules - but that's what computers are for. The question is: are you smarter than that? The approach described by fuzzy1 would work, but apparently not without idiot-savant levels of concentration or plenty of time and scrap paper. The alternative is to have a bit of understanding of what you're doing. Remember that these are practice questions - they won't exactly go on your permanent transcript. They're geared to allow potential candidates to decide for themselves whether or not they understand enough to take the real exam.

especially since the variable names are not descriptive

Of course the variable names are not descriptive. The code would work exactly the same no matter what the variables were called - descriptive, nondescriptive, or just plain misleading. Actually, maybe that would make it a better question. Have a variable called "$false" that would actually need to be true, or $null which may or may not be null, or "$doesNotHaveNoEmptyValue". See how many people who think that variable names actually many anything to the computer get tripped up by it. (Let's load in some context. You've been contracted in to add some new features to some ghastly kludge that originally in Perl before being ported to PHP back last century and you're trying to determine what conditions lead to a particular error message. You've found the routine that generates the message and it's buried in a tree of if-statements testing a little maze of twisty boolean expressions, all different.... Rewrite it? May be a good idea if (a) you can convince the owners who have obviously resisted the idea up until now that it's worth their money, and (b) you already understand all its business rules sufficiently to replicate them, including (c) all the owners' favourite bugs.)

As for a "sign" of programming ability, I'd posit that logical reasoning is a requirement of programming ability. A glance at the programming ability of certain "programmers" (certain MSCEs included) shows that programming ability doesn't correspond much to programming in the real world either. But I don't think the ZCE exam does much to test one's office politicking or job interview skills either. But you already know that most real world software is crap.

laserlight

Weedpacket wrote:
See how many people who think that variable names actually many anything to the computer get tripped up by it.

Programs should be written for people to read, and only incidentally for computers to execute (pardon my paraphrase, but it might be an accurate quote).

Weedpacket wrote:
You've been contracted in to add some new features to some ghastly kludge that originally in Perl before being ported to PHP back last century and you're trying to determine what conditions lead to a particular error message. You've found the routine that generates the message and it's buried in a tree of if-statements testing a little maze of twisty boolean expressions, all different.... Rewrite it? May be a good idea if (a) you can convince the owners who have obviously resisted the idea up until now that it's worth their money, and (b) you already understand all its business rules sufficiently to replicate them, including (c) all the owners' favourite bugs.)

Yes, rewrite it. Not necessarily the entire thing, but that portion of code, after creating suitable unit tests to ensure that your rewriting does not break anything else. Effectively, that would be similiar to mechanically applying rules for simplification. The difference is that with mere inspection, the difficulty of deciphering the code remains, whereas a suitable refactoring means that the code will be easier to understand and debug even if another problem crops up. In my opinion, this is a more useful "real world" skill than just being able to determine the origin of that particular error message by inspection.

Weedpacket wrote:
As for a "sign" of programming ability, I'd posit that logical reasoning is a requirement of programming ability.

I agree, and it should be a given, at least where these kinds of tests are concerned. I daresay that a simple programming problem presented in an interview would better serve to determine if a candidate has the ability to reason and solve problems than an MCQ quiz that is not supposed to focus on testing that in the first place, and for which the reasoning process cannot be observed.

Weedpacket

laserlight wrote:
My first impression when reading that question is that it is testing whether the reader knows about the conversion of null to false when in a boolean context.

....Nah. Too complicated for that; too many ways to stuff it up without getting that bit wrong.

Programs should be written for people to read, and only incidentally for computers to execute (pardon my paraphrase, but it might be an accurate quote).

Not incidentally: they have to be executed for them to be any use. Doesn't matter how readable it is if it's wrong. My point is that the computer doesn't understand anything about what's going on; determining what a computer will do with a program can't rely on anything that isn't available to the computer - like the meaning of variable names.

whereas a suitable refactoring means that the code will be easier to understand and debug even if another problem crops up. In my opinion, this is a more useful "real world" skill than just being able to determine the origin of that particular error message by inspection.

'Course, first you have to find the offending code before you can clean it up.

Like asking that the code in that question be rewritten (I mean, refactored) as something more like

if(!$a || $b) {
	$number = 5;
} elseif(!$c || $d) {
	$number = 4;
} else {
	$number = 3;
}

which, you'll note, is precisely as effective as the original at testing the testee's knowledge of the effect of casting a null to boolean. But it would make the question asked about as difficult as "in which sport is tennis player Roger Federer famous?".

I agree, and it should be a given, at least where these kinds of tests are concerned.[emphasis mine]

I agree too; but more generally: logical reasoning should be a given for programming. Anyone who has aspirations of being a programmer has to be capable of it (and want to do it). If you can't reason logically your programming is going to be crap even if it does (through liberal amounts of voodoo and lots of "plz hlp" posts in forums) work most of the time. If you don't want to reason logically then programming really isn't for you.

But you'll have to admit, these practice exams have shown that at least one person needs to learn a fair bit more about PHP programming before claiming competency. And in that sense they've been successful.

laserlight

Not incidentally: they have to be executed for them to be any use. Doesn't matter how readable it is if it's wrong. My point is that the computer doesn't understand anything about what's going on; determining what a computer will do with a program can't rely on anything that isn't available to the computer - like the meaning of variable names.

Obviously, at least other than for a researcher to obtain tenure :p But that misses Abelson and Sussman's point: the readability of source code is very important. I would say that if the source code is misleading, then it is unreadable. If it is wrong, then how readable it is matters, since it affects how well you can spot the problem and fix it. (Incidentally, according to Paul Graham's Hackers and Painters essay, I almost got the quote correct word for word, except that "computers" should be "machines". Gah.)

However, my comment about undescriptive variable names is just a passing remark about why I am inclined to use boolean algebra. The names are at a meta-level, resembling variable names often used in the realm of mathematics instead of programming, or in programming to dissociate from context. This seems contrary to the actual ZCE exams, which supposedly "use a set of testing questions that reflect real-world scenarios". In that sense, perhaps your idea of misleading variable names would be a Good Thing, though the problem of it being awkward to refactor code in such a test environment remains.

'Course, first you have to find the offending code before you can clean it up.

Yes, and such would make for good ZCE exam questions (and I presume that such questions already exist). Provide readable code that contains one or more bugs, and then have the candidate identify them (and maybe match it with a suitable fix). No refactoring needed, so it becomes a challenge on whether you can spot the bugs and fix them, and this would indeed be realistic.

But you'll have to admit, these practice exams have shown that at least one person needs to learn a fair bit more about PHP programming before claiming competency. And in that sense they've been successful.

Yeah. I know I have done some "Ten Year Series" questions back in school that did not really help for the real exams, except that they made me study more, heheh.

fuzzy1

Originally Posted by Weedpacket
Anyone who has aspirations of being a programmer has to be capable of it (and want to do it).

Aspirations? ME.
Desire (and want to do it)? -- a bit redundant, but... also ME. I enjoy coding immensely!
"has to be capable of it"? well... that WAS kind of the point in taking the practice exam. (note title of post?)
That is, to gauge my capability, identifying -- with a modicum of specificity -- the holes therein (Just for Grins😃.)

Originally Posted by Weedpacket
As for a "sign" of programming ability, I'd posit that logical reasoning is a requirement of programming ability.

In context of all the aforementioned it did seem logical/reasonable then, to seek validation from knowledgeable
sources as to whether or not it was in fact ILLOGICAL to conclude that many of the questions in the practice tests
-- presumably designed as a means of gauging the programming ability of others, by folks of unimpeachable logic themselves --
were in fact, often silly if not just plain wrong.
On the often silly where not just wrong count, there DOES appear to be some consensus.

Originally Posted by Weedpacket
you'll have to admit, these practice exams have shown that at least one person needs to learn a fair bit more about PHP programming before claiming competency. And in that sense they've been successful.

That would be ME also Weedpacket, a point on which --in my initial two posts here-- I have already expressed earnest willingness to concede.
Why don't we leave it there then, before it gets nasty and somebody gets hurt? :p