Taking both your point and OP's, I think yes, the decimal to binary conversion is a representative of the data wrangling problems developers are asked to solve all the time, but the process should include giving the specification of decimal and binary numbers, so that prior knowledge of them is not required.
I still think the decimal to binary conversion is a problematic test.
Decimal to binary is a required problem taught to Computer Science university students, but not so pervasive outside of universities. University students will likely solve it simply by memory, while other people may have to figure it out by themselves. That makes the test:
unfairly skewed towards those who can afford a university degree
less able to differentiate within the university students
not closely related to the domain of the employer
A similar data conversion task taken from the domain of the employer would give the prospect employees an idea of what the work will look like, prefer those who are more familiar with the domain, and test the problem solving skills of university students as much as those of others.
I'm worried that you think exposure to the concept of binary and decimal for people working in computer science/development is limited to those with degrees. Anyone with basic math or comp-sci knowledge should know this. If someone applying for a comp-sci/dev position managed to completely skip a pretty basic and core concept in the field I'd wonder about the quality of their knowledge and worse still if they didn't know and couldn't ask for gaps in their knowledge to be filled in on site (which would take all of 10 seconds), I'd worry that they are going to be terrible at a job that is 90% about asking questions to things we don't know.
A similar data conversion task taken from the domain of the employer
You think that binary is beyond people without a degree but specialist domain knowledge isn't?
No, I do not think only those with degrees know of binary and decimal. I merely assert that those with degrees have a far higher chance to know of binary and decimal than those without. Please don't strengthen others' proposition unilaterally.
You think that binary is beyond people without a degree but specialist domain knowledge isn't?
No, I do not think binary is not beyond people without a degree. Rather, I think domain specific knowledge is beyond both people with and without a degree. As you say the job is
90% about asking questions to things we don't know
Then for those with a degree you are not testing this at all, since they know binary already.
I still think the decimal to binary conversion is a problematic test.
^ That is the wrong thing to focus on. Mentioning disparity between university CS and non-university CS is another level of irrelevant. I provide the test, candidate can ask "how do you do that" I say "you divide and mod the number by two, the division gives you the next step the mod gives you the first output" then they write the algorithm down in code.
Whether that problem is binary to decimal, decimal to binary, fizzbuzz or reversing a string, the thing to focus on is that it's an easy problem to explain, it's ripe for asking questions and it leads to a small amount of code that is representative of a candidates basic coding skills. What you get from that 30 second to a minutes worth of effort is:
Can this candidate analyse and understand a simple problem statement
Does this candidate have good CS basic grounding
Can the candidate admit that they have gaps in their knowledge
Can the candidate ask questions
Does the candidate have a basic grasp of loops, variables, operators, state
Can they converse in a clear way from a technical focus
We agree on many. We likely agree on more than you think.
But man, this whole thread started about the binary to decimal test. That's part of the topic. How can talking about the topic of the thread be "wrong"?
Thank you for the list of considerations. I like it. But let's cut it here.
Taking both your point and OP's, I think yes, the decimal to binary conversion is a representative of the data wrangling problems developers are asked to solve all the time, but the process should include giving the specification of decimal and binary numbers, so that prior knowledge of them is not required.
I still think the decimal to binary conversion is a problematic test.
Decimal to binary is a required problem taught to Computer Science university students, but not so pervasive outside of universities. University students will likely solve it simply by memory, while other people may have to figure it out by themselves. That makes the test:
A similar data conversion task taken from the domain of the employer would give the prospect employees an idea of what the work will look like, prefer those who are more familiar with the domain, and test the problem solving skills of university students as much as those of others.
I'm worried that you think exposure to the concept of binary and decimal for people working in computer science/development is limited to those with degrees. Anyone with basic math or comp-sci knowledge should know this. If someone applying for a comp-sci/dev position managed to completely skip a pretty basic and core concept in the field I'd wonder about the quality of their knowledge and worse still if they didn't know and couldn't ask for gaps in their knowledge to be filled in on site (which would take all of 10 seconds), I'd worry that they are going to be terrible at a job that is 90% about asking questions to things we don't know.
You think that binary is beyond people without a degree but specialist domain knowledge isn't?
No, I do not think only those with degrees know of binary and decimal. I merely assert that those with degrees have a far higher chance to know of binary and decimal than those without. Please don't strengthen others' proposition unilaterally.
No, I do not think binary is not beyond people without a degree. Rather, I think domain specific knowledge is beyond both people with and without a degree. As you say the job is
Then for those with a degree you are not testing this at all, since they know binary already.
The test is supposed to -in 30 seconds- show basic coding skills. NOT test knowledge of binary.
You are focusing on the wrong thing and creating a straw man.
My focus is
So you think that is the wrong thing to focus on?
^ That is the wrong thing to focus on. Mentioning disparity between university CS and non-university CS is another level of irrelevant. I provide the test, candidate can ask "how do you do that" I say "you divide and mod the number by two, the division gives you the next step the mod gives you the first output" then they write the algorithm down in code.
Whether that problem is binary to decimal, decimal to binary, fizzbuzz or reversing a string, the thing to focus on is that it's an easy problem to explain, it's ripe for asking questions and it leads to a small amount of code that is representative of a candidates basic coding skills. What you get from that 30 second to a minutes worth of effort is:
It also has the added discussion routes of
We agree on many. We likely agree on more than you think.
But man, this whole thread started about the binary to decimal test. That's part of the topic. How can talking about the topic of the thread be "wrong"?
Thank you for the list of considerations. I like it. But let's cut it here.
I think the test is stupid. I have BS CS but I vaguely remembers how to divide and get that binary stuff. You need knowledge on how to convert it.
There are also self thought developers and this test will prove nothing.