Which concept took you the longest to grasp?

This could be anything from some specific technologies/tools, data structures, design patterns, programming paradigms to advanced concepts like Machine learning, Neural networks, AI, etc.

Which concept took you the longest to grasp?

Comments

[Matt Ellen-Tsivintzeli]

Back in 1999 I was trying to learn c++, but I got stuck on template classes. A few years later I learnt about generics in c# and suddenly I understood what template classes were :D

I'm still stuck on monads, so I guess that will be the biggest gap when I finally get back to it.

[Madza]

Would you say it would be a great first language?
I'm not considering the learning curve, but the way the lang is designed to learn programming basics. Most say C is super solid. Although, I've heard C++ is good, too.

[Dave Cridland]

Python.

The advantage of Python is that it's quite a "broad church" language, supporting many different approaches to solving a problem - but it's also very consistent.

There are some things it won't teach you, most notably generics/templates (which are different, but similar, capabilities), but generally most concepts in other imperative languages exist in Python (and there are a few functional techniques about as well).

The lack of special-casing and "gotchas" makes it very simple to work with and explore, though, and things like Jupyter (a sort of web-based super-REPL) help things further.

Moving on from there, C++ is a great language to understand because it has both very low level elements to it and high level constructs built on top - allowing you do dig very deeply into how things like coroutines or containers actually work if you need to.

[Madza]

I have no experience in using it, but as far as I've seen some code, I've liked the syntax of it 😉 Very intuitive and easy to read for those starting out 👍

[Dave Cridland]

There used to be quite a number of languages designed for teaching - BASIC was one (designed by, amongst others, a nun), and Pascal was another. Universities and schools seem to have universally ditched them all in favour of Python.

[keogami]

Sauce for the nun thing?

[Dave Cridland]

Sister Mary Kenneth Keller

Took vows in 1940, and while doing a PhD - the first woman to get one in Comp Sci - ended up in Dartmouth College, implementing the first version of BASIC.

[Christopher Kruse]

Wow, I didn't expect to find a Sr. Keller reference on DEV!

I went to Clarke, where she founded the Computer Science department (and there's a building named after here there now).

[keogami]

Lol a nerdy nun, thanks for thr fact!

[Matt Ellen-Tsivintzeli]

My first language was BASIC, which I think was a good place to start. It's more important to have fast feedback when you're first starting out (IMO) to get you hooked. So I would recommend an interpreted language over a compiled one, just because you can get the immediacy in a read-evaluate-print loop (REPL) where you type something and it causes something to happen.

Also, the "batteries included" nature of most popular interpreted languages means that you're going to find it easier to do things off the bat, not trying to figure out how and what to include.

I can't say I'm a C++ dev, because it's been well over a decade since I last wrote any, so maybe the STL and whatever exists now has a batteries included approach, so perhaps that's not a concern.

But for the basic concepts I'd start with a language that outputs comprehensible errors and doesn't make you wait to find out when you make a mistake.

C and C++ are good second languages.

[Madza]

Fair enough 😉
Some solid points there 👍

[peerreynders]

but the way the lang is designed to learn programming basics.

"Programming Languages" are designed to represent programs that can get "stuff done" - not to teach programming basics - unless of course the language is expressly designed to "teach programming basics".

How to Design Programs, Second Edition

HtDP teaching languages

Pyret

PAPL - Programming and Programming Languages

The Structure and Interpretation of the Computer Science Curriculum (2004):

In general, we believe that the HTDP project has validated the usefulness of functional programming and functional programming languages in the first programming course. We have found that teaching Scheme for Scheme’s sake (or Haskell for Haskell’s sake) won’t work. Combining SICP with a GUI-based development environment for Scheme won’t work better than plain SICP. The two keys to our success were to tame Scheme into teaching languages that beginners can handle and to distill well-known functional principles of programming into generally applicable design recipes. Then we could show our colleagues that a combination of functional programming as a preparation for a course on object-oriented programming is an effective and indeed superior alternative to a year on just C++, Java, or a combination.

We are hoping that other functional communities can replicate our success in different contexts. We suggest, however, that using plain Erlang, Haskell, or ML and that teaching programming in these languages implicitly will not do. We all need to understand the role of functional programming in our curricula and the needs of our students. (p.13)

quote

“It’s mind boggling that your HtDP students are better C++ problem solvers than people who went through the C++ course already”.

[Mohamed Dahir]

I can't say I fully understand what a monad is but after banging my head days against parser combinators, I am a bit more confident about it.

[David Snyder]

Pointers

[Ben Halpern]

This made me curious

Do they teach "pointers" in bootcamp? Should they?

Ben Halpern ・ Dec 3 ・ 1 min read

#discuss #computerscience #codenewbie

[Peter Witham]

This is high on my list, foolishly I dived into C++ with no prior programming experience way back when and immediately starting crying lol

[David Snyder]

same, C++ was my first language and I was having fun for a few months until I got to pointers and decided I would switch to JavaScript lol

[Peter Witham]

It always seems like every language has that moment that comes down to "you loved me until now"

[Matt Curcio]

Bayes theorem! I use it occasionally but I have to teach myself over and over. It's like learning a new language, 'use it or lose it.'

[keogami]

What do you use it for?

[Matt Curcio]

The last time I used it I was testing the false positive paradox. I was used to work in the biotech industry. So, What's the chance you get a disease versus the fact that you tested positive for that disease, tested once or even twice. Alternatively, one might come across it for A/B testing in stats. What are the chances you will see false positives in your comparison testing.
HTH

[keogami]

I dont have much of a mathematics background so anything related probablity hits me hard. Could you suggest me study material? Oh and what kinda data you were storing for your purpose?

[Matt Curcio]

Starting out in Statistics, I suggest something like Kahn Academy. It looks well thought out.

But I also like Kahn Academy because you can jump around as you want.
HTH

[keogami]

Oh i love that guy! He runs a new channel by the namr "Three blue one brown", i learned differential from on there 😍😍

[OmegaCerberus]

For me it was Dynamic Programming,
Had a hard time understanding when to use it in problems and still now i'm kinda afraid of it mainly the ones coupled with trees and graphs

One of the reasons is that i was unable to find appropriate resources on how to use and understand DP and also cuz i'm kinda dumb

[Jeffrey Desir]

Lol only a self-deprecating genius could have a hard time with dynamic programming, and still feel kinda dumb 😁
If intelligence could be quantfied, then maybe the ability to transform data outside of its structure would be a better start than the joke that is IQ.

[Carlos Córdova Sáenz]

For me, even until now, after 3 years from the first time I read and studied about DP, I still forget almost everything about it and have to study again to keep it fresh in my mind.

[Jeffrey Desir]

Soon (enough 🙄) my affinity for LISPs & Clojure will be matched with my discursive posts. The most rewarding & self-damning attempt I'd ever made in my CS career is learning Clojure 🌟macros🌟.
random example:
often when we repeat declarative patterns (do this with this) we capture their behavior in functions or methods

(ns example.macros)

;; simple script to compare variable against 0. 
(defn heat 0.001)

(if (not (= 0.0 heat))
     (println "Cannot cohere!"))

To capture the if atoms behavior, let's try a function first.

(defn if-cohereable [value dispatch]
  (if (not (= 0.0 value))
     (println "Cannot cohere!"
      (dispatch)))

;; nice! but a few problems in practice...
(defn run-program [ :keys [ actions, H* ;;Heat Operator ]] 
   (if-cohereable H* (actions)))

(run-program :H* 0)

;;psuedo-output
=> #<clojure.function>

looks helpful, but we can't actually use this as an 'if' statement, just a comparator that returns a function.
With the power of Macro though

(defmacro if-cohereable [ :keys [H* actions-body ]]
   (if (not (= 0 H*))
      ;;throw error if comparison fails
     (throw "Non-zero-energy prevents entanglment!")
     ;;otherwise run the body of code passed into macro call
     ~@(actions-body)))

;; now in action...

(defn run-program [ energy ]
  (if-cohereable :H* energy ;;named arities with if-statement
    (do
      (init! energy)
       (simulation/start!)))

😅 a rather innocuous example but point is I couldn't grasp the use-cases for this meta-programming feature until I labored through documents & examples & uses before I understood the sheer power in talking about code the same as data.

The closest thing I know to LISPs macro power are most other languages Proxy Generator or Iterator capabilities and although they mirror much of LISPs flexibility, they are SYNTACTICALLY inefficient to express data transformations (all code is) fully.
✌🏿 It was a pain and a pleasure both learning and sharing ~

[ProMikeCoder2020]

Some of them were(please don't judge me as I am only a 13 guy learning to program in his spare time):

• The concept of classes
• Git and GitHub( Tech with Tim video on this helped me a lot to understand this stuff)
• Responsive CSS(Kevin Powell responsive course really helped me on this)

What I still have problems with:

• Yarn workspaces
• Yarn and npm dependencies problems and such
• CSS weirdness
• Choosing the correct semantic html tag

[Jean-Michel 🕵🏻‍♂️ Fayard]

Simplicity

[Madza]

Well put 👍
Also 'simplicity' !== 'easy', both directly and indirectly 😉

[Jean-Michel 🕵🏻‍♂️ Fayard]

Every developer should watch and digest this video from Rich Hickey :P

infoq.com/presentations/Simple-Mad...

[Madza]

Thanks for sharing, this is golden 🔥🔥

[keogami]

Lol

[Gautham Vijayan]

When I was in 11th grade I did not understand C++ concepts which was taught in my school. Eventually I grinded hard enough to score full marks in all of the C++ tests they had given about the fundamentals of C++ like classes, inheritance and conditionals.

[LemmaEOF]

Whenever I come across something difficult, it's usually a sort of there-is-no-spoon sort of thing. The first I can remember was wrapping my head around Java generics only existing at compile-time, and then grasping what first-class means in the context of programming languages. Now I suppose it's monads.

[Jeffrey Desir]

you're like an inch away from the cliff of Clojure 🙏🏿

[Forrest Cahoon]

I don't know if this took the longest for me to grasp, but it took a certain example. Way back in the 1990s when I was first learning object-oriented programming, it was hard for me to get a grasp of what an object really was. Every explanation of OO programming made objects out to be these things with almost mystical properties.

It was only when I encountered Perl's LWP module, which uses an object as an "agent" to make web requests, that I realized that an object could, in fact, be anything.

Any grouping of behaviors that arbitrarily decide are being done by "the same thing", hey, you're perfectly free to declare that's an object. It's all up to you.

The problems with all the previous explanations I encountered is that they all conflated what an object should be (in terms of sensible code organization) with what an object could be (in terms of language structure, which is to say, no limits).

I still think it would be great, when introducing OO, to show examples of silly decisions of how to break a problem up into objects, just to make this arbitrariness perfectly clear, and open the door to creative thinking about what objects could be.