|
So many times have I started to explain a problem to someone and almost instantly figured it out as I was explaining the problem. I've known people who have completely scoffed off "rubber duck decoding" but it tends to work so well for me.
|
|
|
|
|
Never underestimate the power of the subconscious.
While you are walking the dog, enjoying a brewskie or lying in your bath the brain is still working away. When you come back to the problem it seems a lot simpler.
We're philosophical about power outages here. A.C. come, A.C. go.
|
|
|
|
|
This is so true.
It's easy to get stubborn and not want the problem to defeat you, so you stick at it for hours, only to wake up the next day and the solution literally just pops into your head.
Also, the other points people have made about talking the problem over OUT LOUD, with either a person or by yourself. That works wonders too. Something about being forced to express the problem clearly - such that someone else can understand where you're at - frequently unblocks the issue in your mind and the solution presents itself.
"And when I have understanding of computers, I shall be the Supreme Being!"
|
|
|
|
|
What you will ideally need is a tutor.
That is someone who is a skilled developer and wouldn't mind occasionally have a look at your problem...
I think there is a tutoring program on CodeProject (with volunteers..) give it a go sometimes, who knows?
(I never did, so I don't...  )
|
|
|
|
|
|
Ha yes, that's the word! 
|
|
|
|
|
You're doing the right things, so you should improve over time as the skills of design and debugging become more practiced.
It's often difficult to work out where to start on a complex problem, so sitting there for 2-4 hours isn't unusual: I've spent most of a week working out what to do before I've even fired up the editor (or Visual Studio). Sometimes that "thinking time" is the most important part of the project!
A couple of suggestions:
1) As has been mentioned, take breaks. I try to stop what I'm doing at least once an hour and do something totally different, as it "recharges my creative batteries" and stops me getting stale.
2) Never be afraid to say "this isn't working" and throw away an idea just because you have put effort into it. It's one of the hardest things to do, but sometimes we have to say to ourselves "this isn't going to work", or "that was a bad idea", even if we've coded most of it... 
3) Practice. Practice, practice, practice, practice, practice, practice, practice. And then practice some more! Development and debugging are both skills, and like all skills they only ever get better when you use them. Learn the basics of the VS debugger, and use them - ignore the complicated stuff for now - it's a "mind set" you need to develop, not an intimate knowledge of arcane debugging commands. Basic breakpointing, stepping into, stepping over, looking at variable contents is 99.99% of all I do with the debugger. It's also worth logging info (either to a log file, or the console) as that can give you a better "feel" for how code is running.
Keep at it - you seem to be doing pretty well so far! 
Bad command or file name. Bad, bad command! Sit! Stay! Staaaay...
|
|
|
|
|
You remind me that I learned the existance of "set next statement" after 3.5 years of developing. All in VS. With complex problems ranging from hardware management to layered integration of Sobel transforms with heuristic detection of patterns.
And that the first 3 months I didn't know how to use the debugger effectively and debugget with prints. Image processing. With prints.
As I say, I know nothing about computers. Beacuse each time I think of what I knew two years ago I end up saying "I didn't know anything back then". Since it happened regularly in the past 20 years I can predict that two years from now I'll be looking back and saying that now I know nothing about computers, so I accept it as of now.
GCS d--- s-/++ a- C++++ U+++ P- L- E-- W++ N++ o+ K- w+++ O? M-- V? PS+ PE- Y+ PGP t++ 5? X R++ tv-- b+ DI+++ D++ G e++>+++ h--- ++>+++ y+++* Weapons extension: ma- k++ F+2 X
If you think 'goto' is evil, try writing an Assembly program without JMP. -- TNCaver
When I was six, there were no ones and zeroes - only zeroes. And not all of them worked. -- Ravi Bhavnani
|
|
|
|
|
I would add only one thing to all the good advises... Choose pet project, that seems to be hard... That will force you to learn new things every day...
Skipper: We'll fix it.
Alex: Fix it? How you gonna fix this?
Skipper: Grit, spit and a whole lotta duct tape.
|
|
|
|
|
In my experience the type of "technical learning" necessary for programming skills in a given OS and "stack" varies from person to person; of course it varies by general cognitive abilities in dealing with logic and abstractions, symbol and semantics, but, there is, also, imho, a kind of "innate" cognitive style which is quite "individual."
Some folks, in my experience, are endowed with strength at learning from "the top down" ... from abstraction at a high-level to "nitty-gritty," from algorithm to code: give them a set of Backus-Naur diagrams and they can visualize how those abstractions "work" in a "real world" computer language as a dynamic set of parsing/object-construction, etc., "rules."
Other folks, like me, in contrast, are "bottom up" learners who learn best by making multiple passes over the high-level concepts and forma structure in the context of focusing on some specific technique, or problem to solve, most often while experimenting/prototyping. I believe this type of learner tends to need a lot of "hands-on" before they can form an accurate "mental model" of how the high-level abstractions and concepts "work" in the code.
I do believe that intense periods of all-out effort, total immersion, in learning are very valuable; at the same time, I believe it is an important skill to develop to know when (as in your example where the screen starts dancing) that you are over-loaded mentally, and the ratio of effort to learning turns negative.
A phenomenon I have noticed in myself, and others, is a kind of "carry-forward" from the first language/stack one learns in depth to your future learning new languages/stacks. For me, the first really deep-dive was with Lisp, and then, PostScript, and I find myself thinking, sometimes, that I wish .NET had the kind of "dictionary stack" to govern current semantic state, and other save/restore-state mechanisms, like the 'gsave/grestore graphic-state semantics, that PostScript has 
If you accept the premise of a tendency for learning in "bottom up" vs. "top down" style, and a connection with the general cognitive style of a person, then, from an educational point of view ... what follows ?
I'd say the strong "bottom upper" at times need to engage in intensive immersion in the higher-level abstractions, in the structure, semantics, algorithms, design-patterns. And, I'd say the strong "top downer" needs to "get their hands dirty," and work with lower-level idiosyncratic features, like dealing with the quirks of the UI Controls that .NET provides in WinForms, WPF, ASP.Net, etc.; like manipulation of bit-maps, using API calls to "fill in the gaps" where the front-end of the stack doesn't provide the functionality you need.
I see nothing wrong with reaching mastery of how to find what you need in the documentation as you need it ... without an accurate "mental model" and understanding of the micro-aspects of that topic's state/behavior paradigm. In my experience with teaching programming, the person who behaves/feels (most often compulsively) like they "understand nothing" unless they "understand everything" is in deep trouble, and needs being "brought to earth" ... gently
Perhaps there is an opposite to that hypothetical "syndrome" where a person may feel over-confident because they can make code that "works," without realizing that they have not accumulated the type of higher-level knowledge that allows them to generalize from their current achievements to other problem areas ?
When I hear someone say, like you did, that you really enjoy getting "deep" into programming: well, I think that is absolutely great !
Oh, I didn't mean to go on like this ...
cheers, Bill
«There is a spectrum, from "clearly desirable behaviour," to "possibly dodgy behavior that still makes some sense," to "clearly undesirable behavior." We try to make the latter into warnings or, better, errors. But stuff that is in the middle category you don’t want to restrict unless there is a clear way to work around it.» Eric Lippert, May 14, 2008
|
|
|
|
|
Excellent post!
I'm top-down AND bottom-up. I do best when I start with a high level view of the subject and dig into the theory. This gives me a visualization of what I'm dealing with, including rough parameters.
Then I get my hands dirty, digging into the details and building little pieces, and grouping them into bigger pieces. I have to have the hands-on work to truly learn anything (maybe everyone does to some extent).
The initial picture is always full of holes. There are too many details left out by verbal descriptions. The hands-on work fills in those gaps and corrects misperceptions.
When learning a new language I build an address book application. I've had the same data model since 1991 and know the requirements inside-n-out. This lets me focus on building an application, from data access layer to GUI.
|
|
|
|
|
Quote: P.S. Is there any standardized way to improve at debugging? I'm still trying to learn my way around the VS debugger.
not really, has has been said before - practice .. that being said, when I started out in 'c' etc, command-line stuff on a Unix platform, all we had was 'printf', none of this fancy IDE stuff .. so, I'd suggest, for my 0.02 worth
a) learn how to step through a program looking at variables etc, setting breakpoints
b) learn a logging framework - log4net or similar - avoid printf("At step 1, value of x is ....") throughout your code
c) look at techniques of 'incremental' building of programs/solving problems - test the basic algorithms/classes etc with unit tests etc that way you can be sure your 'base'/foundation is firm before you proceed
d) wether you use command-line or Atlassian SourceTree or GitKraken, as you develop bigger programs, use source control - sure, when you start out, you backup important files/directories as date/time stamped zip files, yes ? but 'rolling back' from these if you mess something up is a pain - so learn and like a Source Control System
above all - read voraciously, have a thirst for knowledge - its not necessarily about the language/framework, its about being able to break problems/issues down into solvable chunks and then modelling them
|
|
|
|
|
Experience, experience, experience.
It's over 4 years I program full time and before that I did all kind of microprojects - I still bang my head on some "simple" problems.
The best way to avoid long fruitless hours is to invest time in designing what you want to achieve and the structure of what you're building. That streamlines and divides the problems into smaller ones that can be tackled more easily. Also keep in mind that you may have to redesing one thing several times and recognize it after long coding hours - how to improve at that? Wasting time. Doing things less than optimal.
With time you'll know what worked and what not and you'll be able to work faster. It takes a year to make a year of experience, there are no shortcuts.
Just don't ever think you have nothing to improve or learn, keep your mind open and be ready to acknowledge when you did suboptimal things. It happens.
GCS d--- s-/++ a- C++++ U+++ P- L- E-- W++ N++ o+ K- w+++ O? M-- V? PS+ PE- Y+ PGP t++ 5? X R++ tv-- b+ DI+++ D++ G e++>+++ h--- ++>+++ y+++* Weapons extension: ma- k++ F+2 X
If you think 'goto' is evil, try writing an Assembly program without JMP. -- TNCaver
When I was six, there were no ones and zeroes - only zeroes. And not all of them worked. -- Ravi Bhavnani
|
|
|
|
|
I agree, good design can make a massive difference. Invest time in learning design patterns, that will save hours of debug/code time later.
|
|
|
|
|
I'll add a couple of other things I have found invaluable over the years.
Another set of eyes. Because you are so close to the code you can end up seeing what you expect to see, having someone else look at a problem has often identified an error that I just can't see no matter hown long I look at it.
Another developer you can talk the problem over with. Often as you try and explain the problem to your colleague you can solve the issue without the other person needing to say a word. This can really weird some people out but it does work.
Never underestimate the power of human stupidity
RAH
|
|
|
|
|
Mycroft Holmes wrote: Another developer you can talk the problem over with. Often as you try and explain the problem to your colleague you can solve the issue without the other person needing to say a word. This can really weird some people out but it does work.
Tested and true. Absolutely invaluable.
GCS d--- s-/++ a- C++++ U+++ P- L- E-- W++ N++ o+ K- w+++ O? M-- V? PS+ PE- Y+ PGP t++ 5? X R++ tv-- b+ DI+++ D++ G e++>+++ h--- ++>+++ y+++* Weapons extension: ma- k++ F+2 X
If you think 'goto' is evil, try writing an Assembly program without JMP. -- TNCaver
When I was six, there were no ones and zeroes - only zeroes. And not all of them worked. -- Ravi Bhavnani
|
|
|
|
|
Agreed. My cat is an excellent problem solver in this way. 
Bad command or file name. Bad, bad command! Sit! Stay! Staaaay...
|
|
|
|
|
I would not trust a cat needs you to solve a problem
Skipper: We'll fix it.
Alex: Fix it? How you gonna fix this?
Skipper: Grit, spit and a whole lotta duct tape.
|
|
|
|
|
You'd be surprised: they are pretty good, because they have a way of staring right at you with an expression of "So? Do I care?" all the damn time that makes you doubt what you are saying is important or relevant. And that leads you to doubt what you believe is right...which helps you go  and realize where you went wrong!
Bad command or file name. Bad, bad command! Sit! Stay! Staaaay...
|
|
|
|
|
I've used Griff's "ask the cat" strategy many times with great success.
A typical "cat programming" session might go something like "Okay, Mog, check this - for some reason this binding doesn't work even though x, y and z are all in place and I know that event gets fired and the dependency is there and that's always going to work unless ... unless ... of course! God! I'm an idiot, why didn't I think of that? Mog, your a genius! Have a kitty treat!"
It's not, needless to say, that the cat provides the answer but that you find yourself restating the problem in a slightly different way - it also helps that you're not physically looking at the code. When you look too hard at code, it's easy to go in circles.
I often find that I'll wake up in the morning with a solution to something that seemed utterly impenetrable the day before. Overnight, you subconsciously reframe the problem and then the solution becomes obvious in a way that just doesn't happen if you chain yourself to your desk.
|
|
|
|
|
I can give you the tip, to handle every return value, if it is an error. Often a logging message is enough.
Try yourself in throwing and handling your own exceptions.
Press F1 for help or google it.
Greetings from Germany
|
|
|
|
|
My number one tip when it comes to the situation where you spend a few hours figuring something out which in the end was fairly simple is to get the habit in to work in seeing it from another angle.
Hard to explain without having it sound cliche but this is one of those situations where often enough you just need to air the problem with someone else. Try to explain it out loud to yourself as if you were describing it to someone or if possible find someone to discuss it with.
It can actually be fairly useful to speak with someone that's not a programmer/coder because then you have to dissect the problem in order for them to understand what's going on and that will help you solve it.
Another thing I might suggest is to take on something a little bigger to challenge yourself and grow. Something where you just can't jump straight in to and requires more design and thought.
A few years ago I made a simple tamaguchi game in c++ for my wife where she easily just could make new pngs to add more clothing, faces etc.
At the time it seemed so trivial to me, I knew what I wanted to do and started doing it. But after a little while it grew so complex that it was a total mess and I realise that I had spent more time writing a UI library than working on a game and once I striped the game away I had something that was easy to reuse. Helped reinforce oop principles in me after realising my mistakes as well as how to better plan ahead.
I like programming games because they can range from super simple to complex behemoths. You can work closely with the gpu, on network or AI and UI. It's also easy to get ideas, just pick something and see if you can replicate it and improve to your liking.
|
|
|
|
|
If I understand what you're saying correctly, the issue is the "impedance mismatch" between knowing what you want to do and knowing how to do it. It can be very impeding!
For example, when doing Python programming, I know various basic things about the language and its syntax, but I don't know hardly anything about useful 3rd party utilities (or even the ones built into Python!) and I certainly don't know the Pythonic way of doing things. So, for example, in C#, if I want to iterate over an IEnumerable, I have a couple options that are well known to me:
for (var foo in bunchAFoos) {...}
or
bunchAFoos.ForEach((f)=>...);
But how do I even write a loop in Python? Is:
for foo in bunchAFoos:
The best way? What if I want an index? Well, I use an extension method in C#:
bunchAFoos.ForEachWithIndex((idx, foo) => ...);
I have google the answer for Python:
for index, foo in enumerate(self.buncAFoos):
And is that the best way?
It gets worse -- do I wrap my Python functions in a class (their called "methods" then)? How do create a static singleton of the class? How do I derive from a class? How do I call the base method? Discovering that I have to explicitly call the base class constructor was a WTF is going wrong debugging session.
And then there's the debugger (using Visual Studio Python plugin). Why can't I step over the code>if statement? On a fluke, I changed:
if self.state == AppState.performing:
to:
if (self.state == AppState.performing):
and weirdly, I can now step over the if statement in the debugger. Why?
How do enums work in Python? (Answer: use the "enum" module and write it like non-member assignments, like this:
class AppState(Enum):
offline = 1
online = 2
unconfigured = 3
waiting = 4
performing = 5
TheOnlyRealTodd wrote: Is this pretty typical for a new coder or coders in general?
It's typical of coders with a new language/framework/platform. Don't even get me started on the lunacy of trying to launch an app at startup in Debian (or any *nix OS.)
So, to re-iterate, the impedance mismatch between knowing something because I've read a book or watched a few videos, vs. knowing how to do a thing, well, that can be cough shocking.
Marc
|
|
|
|
|
Marc Clifton wrote: the impedance mismatch between knowing something because I've read a book or watched a few videos, vs. knowing how to do a thing
I call that "eloquent," and I speak from direct experience of a life-time of swimming upstream all-too-slowly against the flux of mis-matched impediments, kept afloat only by some talents in rationalization and denail
cheers, Bill
«There is a spectrum, from "clearly desirable behaviour," to "possibly dodgy behavior that still makes some sense," to "clearly undesirable behavior." We try to make the latter into warnings or, better, errors. But stuff that is in the middle category you don’t want to restrict unless there is a clear way to work around it.» Eric Lippert, May 14, 2008
modified 5-Jul-16 1:47am.
|
|
|
|
|
you hit denail on dehead with that one.
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
