Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Skyrunner]

Nice o.O

I sometimes wish I was better at algorithms and dynamic programming, since all the programmingncontests are basically informatics olympiads that are all about dynamic programming. To this date I have exactly 1 award (bronze, inschool) from such a competition.

[Putnam]

Always a fun kind of thread to read through.

[MorleyDev]

Wait, what?

Code: [Select]

Field f = Integer.class.getDeclaredField("value");
f.setAccessible(true);
f.set(5, 6);

Integer i = 5;
System.out.println(i);

Java logic! We don't support operator overloading because it has the potential to be misused by a programmer, but all those other things that can be misused? A-ok!

[Skyrunner]

What does that code snippet do? What is it supposed to do/why is it being abused?

[da_nang]

What does that code snippet do? What is it supposed to do/why is it being abused?

Through reflection, it makes the Integer wrapper class represent the primitive int 5 with primitive int 6.

[MorleyDev]

Through reflection, it makes the Integer wrapper class represent the primitive int 5 with primitive int 6.

Essentially. This feature has some more serious use-cases. PowerMock, for example, uses this kind of reflection functionality to allow you to Mock or Stub implementations where you don't have access to an interface (i.e final classes).

Now, I do have some issues with using PowerMock (Most situations, I've found it cleaner to develop an lightweight, abstractable, manual and integration tested facade around the 3rd party libraries) but I can see some usage, especially in migrating towards a better system. A very useful temporary band-aid for working with legacy code.

I just find it somewhat arbitrary at times what features in Java were excluded because people can misuse them (like operator overloading), and what features they went out of their way to include inside the language despite being misusable (complex reflection logic).

[alexandertnt]

Probably because operator overloading is "nice to have" but not really a big problem if its not there, wheras reflection is a powerful and useful tool that can do all sorts of magic. Both are abusable, but one is more justifiable.

I am guessing anyway.

[Telgin]

To be honest, I haven't run into many situations where operator overloading was useful.  Vector and matrix math so far is the only example of anything I've written outside of a classroom where it was used.  Operator [] is one I can see being reasonably useful otherwise if you're writing smart containers of some kind, and of course I use that with vectors but not in any class I've written.  The dereferencing operator is useful for smart pointers, but in Java that's kind of pointless.

On the other hand, it is useful for math libraries, and excluding it because it can be misused sounds awfully heavy handed.  It's not like raw pointer levels of dangerous.

[tahujdt]

I am going to embark on programming a text adventure with nothing but a TI-89 Titanium calculator. I will post the results later.

[MorleyDev]

On the other hand, it is useful for math libraries, and excluding it because it can be misused sounds awfully heavy handed.  It's not like raw pointer levels of dangerous.

More or less my thoughts. Not sure how I feel about the opposite extreme yet. Languages like Scala have it so you can use almost any symbol in a function name. So <++> could be a valid function name. They also support infix operations, so you can ommit the . and paranthesis when invoking a function. So x + y is equivalent to x.+(y).

It's mostly used so that they can apply the "everything is a function and every function is an object" concept, making addition and subtraction just predefined functions on certain types. Powerful, certainly, but people can take it a bit too far. Though I do kinda like how + is addition, whilst ++ is concatenation.

[Skyrunner]

I have a problem.

Spoiler: image (click to show/hide)

Take this DF map of oceans and rivers. It shows the width of the rivers in quantized levels (around five of them, plus Ocean). I want to make a tree that represents each river, with the parent node being the part where the river joins the ocean and branching off where the river splits. Each node represents a pixel. The problem is, given a piece of the map like



I have no way to decide the parent node. It wouldn't matter normally except that if I make the tree in the wrong direction I can't handle branching of the river :<. How should I go about solving this? I suppose it would be easier to solve if the entire map is given as input (since ocean is guaranteed to exist along the edges), but also take longer to solve.

[Sheb]

So a friend of mine is required to use AIMMS for project in college, but they never taught her how to use it. Anyone here knows good ressources, or would be willing to answer a few questions.

[Levi]

I have a problem.

Spoiler: image (click to show/hide)

Take this DF map of oceans and rivers. It shows the width of the rivers in quantized levels (around five of them, plus Ocean). I want to make a tree that represents each river, with the parent node being the part where the river joins the ocean and branching off where the river splits. Each node represents a pixel. The problem is, given a piece of the map like



I have no way to decide the parent node. It wouldn't matter normally except that if I make the tree in the wrong direction I can't handle branching of the river :<. How should I go about solving this? I suppose it would be easier to solve if the entire map is given as input (since ocean is guaranteed to exist along the edges), but also take longer to solve.

I don't think that can be solved if you are doing it in chunks like that.  Maybe you could make a bunch of unrooted nodes with those chunks, and then do a second pass that connects them up to a final tree? 

Honestly I think it might be faster to trace along the edge of the ocean to find the root nodes first, and then just trace the branches recursively.  You could still just load the chunks as you need them while tracing if memory is a concern.

[Rose]

I would go through the whole map first and convert the bitmap data into a more vector-ish format that has direction, elevation, and width, then work from that.

[GlyphGryph]

Welp, I have a confession to make. I am TOO STUPID to figure out how to use IDEs. Its just... its not going to happen. I just can't handle the complexity. At this point I am pretty sure that anything beyond vim and g++  and the command line tools is just going to be forever beyond the grasp of my feeble mind. It is just.... too complex for me. :/

Spoiler: Danger, whine ahead (click to show/hide)

And honestly, everywhere else is moving past me to At work, I am having to learn and keep up with ember and build an ember app that can both be embedded in a rails environment  and  within a mobile app, and I don't think I am going to be able to do it.

I don't know why I even entered this field. I should have known, in college, when I was the only person who managed to have trouble compiling their eclipse app, that this wasn't going to  be a field I could thrive in, and yet... Ugh. I just can't do it any more. I can't keep up, and I can't move forward, and I can't afford to stop or I will be swallowed up by my college loan debts.


I officially hate programming now. Its just a constant exposure to people who effortlessly manage the things I just can't grasp, no matter how long and hard I try..