> In programming, it’s quite difficult to determine the point at which you actually know something versus when you don’t. A person may be strong in systems development, yet when interacting with someone of a comparable skill level who works in web development, they can feel incompetent.Are there any objective metrics for assessing a developer’s level? For example, can someone who has built LFS be considered technically mature? Or someone who has reached a senior position? These are just examples. I’m interested in a subjective perspective on what criteria actually make sense for evaluating an engineer’s level.
>>107824032Literally only retard managers who don't know how to code actually believe that it's "difficult to determine who is a good coder".The more middle management and upper management is away from know how the harder it is for them to know who is bullshitting and who knows the substance. I worked in shitty companies that were built by pompous ex-CERN ex-Professors with doctorates to the brim but knew jack shit about coding, and I worked in companies where CTO was a self taught programmer without even a bachelors but knew coding inside and out (from lowest level to abstract ideas). The best companies were where a programmer was a manager, without a doubt. They created teams that were specific to the field and they knew how to test the candidates.Programming is a skill that requires relentless drive and ambition beyond the monetary gain, willingness and desire to learn beyond "I know enough" and constant desire to be humble before others in terms of sharing knowledge (as they might know more than you). That is literally it. Not language specific knowledge. Not even CS knowledge overall. Sure, these things will overlap, someone who is driven will know shit beyond what the usual curriculum on CS major is, but if I don't know what a CS major curriculum is how will I even judge his knowledge?See? That's why it's said "it's difficult to determine". No, it's hard to actually control people who have knowledge and power to go away and build their own companies from ground up, that's the real issue. Anyone you can employ isn't the sort of person who will build a SaaS on their own to create additional side gig income. Hell, with bar of entry so low in vibe-coding world the only real thing that limits someone is the drive and willpower.
>>107824032These days it's actually quite easy.>do you use AI?If yes, they're incompetent.
a trial period is the only way you are going ot evaluate anyone's ability meaningfully.otherwise you are just optimizing for the biggest liars and grinders. even those with low IQ could memorize a bunch of leetcode solutions, its not that hard.
>>107824032Number of engines built.
>>107824032Number of websites built.
>>107824120I get what you’re saying. In practice it really does come down to having someone technical involved in hiring, otherwise it’s easy to miss who actually understands the work and who just talks well. Teams led by people who’ve coded seriously tend to structure interviews and expectations much better. Motivation and long-term drive matter more than formal labels, that’s hard to disagree with.
>>107824362fact
>>107828123thanks for not adding anything to the conversation
>>107824032>Are there any objective metrics for assessing a developer’s level?Very hard to evaluate I'd say. What I'd most expect out of an experienced developer is a general design & problem solving capability. This is obviously very hard to evaluate, in an interview I mostly focus on questions and toy problems which show how the guy thinks and how he approaches problems. If it looks like he knows his shit in the interview then I guess he's worth a try and the only real way to tell is to see him on the job. Have a trial period maybe or something like that.>For example, can someone who has built LFS be considered technically mature?No, this is almost certainly never a skillset you will make any use of on a real job.>Or someone who has reached a senior position?It's more than nothing but I've seen alleged seniors or "architects" who came off as very incompetent in the interview.
>>107824032Yes. Simply create a truth table for every single thing a programmer might know (based on questions one very single domain). You'll probably have to painstakenly fill in over 9000 forms, but it'll be the most detailed representation of the skills.
>>107826457If it's above 0, become concerned.
>>107824032IQ test + Years of experience
>>107828202ygu
>>107824032Write 2 basic tests which determine whether or not the person can solve basic problems by using variables and functions.Write 1 complex test which determines whether or not the person can solve a multi-step problem.They don't need to pass the test, but if they display an ability to break down a problem into smaller steps, then write functions, then you can bet that they will be able to write a large scale program given enough time if they *want* to.A real method is just a checklist of program methodologies and evidence that they used such a methodology. Variables, using the variable correctly, int usage, operator usage +-/* modulus, function, using the function correctly, designing functions appropriately, recursion (very rarely useful), lambdas, external libraries, file usage, datatype class definitions, a math test can be very very good for long term aptitude. There is only 1 thing that matters and that is the will to do it. 10 years of will can prove to be surmounting in experience and skill so beyond that of other developers that catching up in speed may seem impossible.
