Memorising Formulas

off.the.walls

Want to see if theres any debate around this in a day where if I don't remember a formula I can go and look it up and bang I can them implement that in my code.

Does think theres any needs for knowing formulas off the top of your head anymore?

RedXIV

Depends what you mean by formula I suppose?

If you mean something mathematically based to implement some sort of real world physics in a game, I can't see anyone objecting to needing to look that up.

If you mean an algorithm for something common enough i.e. searching or sorting algorithms, I'd be worried that interviews would make you seem a bit more unprepared if you've to try and figure out something that you see usually ingrained via experience

Very much length of a piece of string argument though

srsly78

My IQ drops by 75% when internet goes down. Can only remember the stuff I actively use, even code I wrote a few months ago is forgotten. And yeah, there are a LOT of equations in my work.

jmcc

It is not about knowing formulas. It is about knowing the right algorithm and the right formula.

If you are doing any serious programming work, always have a paper notebook and pen to hand and write down any ideas, formulas or algorithms that you may find useful or use in a program. And put comments in your code so that you will know what you were thinking at the time you wrote it.

Regards...jmcc

5rtytry56

Personally speaking, memorizing a formula is a bit like memorizing a really long password ( > 10 characters ). You could trying a memorizing them like that.

colm_c

Formulas are from physics and maths, in programming it's algorithms and design patterns that are the important parts IMO.

I don't think you'll be asked to write the Formula for Velocity in C in any developer job interview!

If you can nail design patterns, Singleton, factory, decorators etc. you'll cover yourself for any interview, and be a better developer.

What the are for, when to use them and a high level of how to use them.

Manach

Memorise songs, poems and ballads - Yes.
Memorise formulae that can be safely read into from a file - No.

Like any skill, the more you practice the better it becomes, so with memory. But I find that while it is important to read and gain understand, rote memorisation of technical formula is not really worth it. A good book on the subject, Moonwalking with Einstein: The Art and Science of Remembering Everything by Joshua Foer does go through memory tips to handle routine items.

14ned

off.the.walls wrote: »

Want to see if theres any debate around this in a day where if I don't remember a formula I can go and look it up and bang I can them implement that in my code.

Does think theres any needs for knowing formulas off the top of your head anymore?

I'd personally far prefer if a new hire looked something up before implementing it instead of using their memory.

I'd even personally far prefer if a new hire did prior art research to find a preexisting implementation of an algorithm and used/cloned that instead of writing a new, buggy implementation.

The older I get the more I stitch together other people's software. My productivity keeps rising too and the average hourly rate I can charge/get away with/convince clients keeps rising with it, so as I'm so fond of saying to clients, my main value add is to write as few new lines of code as possible

Niall

BrokenArrows

14ned wrote: »

I'd personally far prefer if a new hire looked something up before implementing it instead of using their memory.

I'd even personally far prefer if a new hire did prior art research to find a preexisting implementation of an algorithm and used/cloned that instead of writing a new, buggy implementation.

The older I get the more I stitch together other people's software. My productivity keeps rising too and the average hourly rate I can charge/get away with/convince clients keeps rising with it, so as I'm so fond of saying to clients, my main value add is to write as few new lines of code as possible

Niall

Exactly. A lot of people think that reinventing the wheel is a good thing. There are such vast resources out there at the moment stackoverflow, github, codeproject etc. Almost anything you are about to write has been done before, and in a lot of cases its been done better than you could do yourself.

At the very least search for an existing implementation to compare it against the way you are thinking of doing it and gain some new ideas.

One point on doing this however is that some people will take the existing code some a 3rd party source, use it, but they have no idea whatsoever how it works. Its absolutely necessary to understand the code you are reusing.

Yourself isit

I agree with the sentiment here but there's a growing trend to ask algorithm questions I'm interviews.

dazberry

Yourself isit wrote: »

I agree with the sentiment here but there's a growing trend to ask algorithm questions I'm interviews.

A company that shall remain nameless ha(s/d) a couple of those questions on their test back in the nineties, AFAIR a sort algorithm and a prime number algorithm. They figured that would sort the chaff from the wheat, I remember being a bit incredulous actually, as I said to them - I'd look them up in the pink Sedgewick book. And it turned out to be a Clipper job too.

Interviewed there again in the early naughties (needs must) and they had the same test - seemed a bit less relevant at that stage of class libraries / frameworks which generally contained some of that functionality - so why would you be reinventing the wheel?

In the mid-naughties the tables turned a bit and I interviewed a number of developers from that company - their faith in their test turned out to be a bit misplaced perhaps.

Subsequently I've covered that sort it all academically as part of a functional programming module - (back to the pink book), but that was a good few years ago and I wouldn't be interview fit without some revision to cover it again as a coding test.

Yourself isit

Apparently this is what FB, Google and others are doing now and it is spreading - at least in silicon valley.

14ned

Yourself isit wrote: »

I agree with the sentiment here but there's a growing trend to ask algorithm questions I'm interviews.

Most of the tech multinationals have four or five hours of interviews on algorithms.

If you really want to work for one of those, buy one of the many books of worked algorithms and memorise them by rote, practising on a whiteboard for speed. Don't think you're cheating, everybody does it. On the day you'll look effortless if you're lucky in the questions put to you.

I think that sort of interview discriminates between how badly the interviewee wants the job and how good they are at faking it only. And, to be honest, that sort of hire is probably what most tech multinationals want to hire, so it's all good.

Niall

Talisman

BrokenArrows wrote: »

One point on doing this however is that some people will take the existing code some a 3rd party source, use it, but they have no idea whatsoever how it works. Its absolutely necessary to understand the code you are reusing.

It's especially true of college students.

beauf

Some here

https://www.quora.com/What-algorithm-questions-were-you-asked-at-an-Amazon-Microsoft-Google-interview