The Big Programming Thread - Page 797

Prev 1 795 796 797 798 799 1032 Next

Thread Rules
1. This is not a "do my homework for me" thread. If you have specific questions, ask, but don't post an assignment or homework problem and expect an exact solution.
2. No recruiting for your cockamamie projects (you won't replace facebook with 3 dudes you found on the internet and $20)
3. If you can't articulate why a language is bad, don't start slinging shit about it. Just remember that nothing is worse than making CSS IE6 compatible.
4. Use [code] tags to format code blocks.

Blitzkrieg0

United States13132 Posts

November 11 2016 01:56 GMT

#15921

Doing an insertion with recursion actually seems really annoying, but ^_^

NovemberstOrm

Canada16217 Posts

November 11 2016 07:41 GMT

#15922

Anyone familiar with swift/xcode?
xcode 7.1
swift 2 i believe

I have one Textfield(playerAnswerTextView) on ViewController1 and a UI label(userAnswerTextBox) on ViewController2, I want the textfield to display its text to the ui label, segues are setup, i have a button setup to go to the next view controller. I've tried to get it work but can't seem to for some reason :/

Help is appreciated.

Nesserev

Belgium2760 Posts

November 11 2016 09:54 GMT

#15923

--- Nuked ---

Prillan

Sweden350 Posts

November 11 2016 13:48 GMT

#15924

On November 10 2016 12:52 Manit0u wrote:
[image loading]

This is great!

Djagulingu

Germany3605 Posts

November 12 2016 21:40 GMT

#15925

I just switched from Unity desktop to Gnome and I'm LOVIN it so far. Workspaces are such a nice touch. Lovin it. If I could move my bottom file bar to left, it would be perfect. Can't I just do that with some extension or shit?

Deleted User 3420

24492 Posts

November 13 2016 16:05 GMT

#15926

I need to write a weighted graph in java. Any opinion on the best way to implement this?

I am thinking that my graph class will have a map members:


Map<node, Map<node, intege>>

where the integer represents the weight of the edge

So My map would store each node, as well as a map of nodes it connects to and the weight of the edge between them.

Is this a good solution?
Are there alternatives any of you would prefer?

Also in lecture, the professor kept mentioning "nodes" and "vertexes". Though for the life of me I didn't understand the difference and he didn't (couldn't?) really explain the difference.

Is it that in a typical representation, the vertex is simply a representation of the spot on the graph, and the node would be an object that contains the vertex and any other associated data?

But in that case if there was no other associated data it would make sense to just represent it with nodes instead of vertexes, right?

Okay, thanks!

Acrofales

Spain18302 Posts

November 13 2016 17:12 GMT

#15927

There are many ways of implementing weighted graphs. It really depends on what you want to do with them. Generally a decent method is:

Your graph is simply a list of nodes

Every node has a list of edges in the form:

Map<Node, Integer> weightededges

However, this doesn't work for some common graph problems (such as max-flow-min-cut) where you want a list of edges at the top level. It also means that if your graph is not directed, you are storing each edge twice, which is slightly redundant (but usually not a problem).

There is no difference between nodes and vertices. Node is used in networks, and vertices in mathematics. Both mean the same thing.

If you need your edges at the top level for the functioning of your algorithm, generally a good solution is to have an edge data structure with fields:

Node start
Node end
int weight

And have a List<Edge> at the top level. However, in this case you will have to do extra work to ensure your graph is sane, and you don't have multiple edges between the same 2 nodes.

I'm not a big fan of your solution of Map<Node, Map<Node, Ingeger>>: it takes care of keeping the graph sane, but it is not as easy to do operations on (such as order all edges by weight): it is a compromise between the node-based solution and the edge-based solution, without the advantages of either.

Deleted User 3420

24492 Posts

November 13 2016 17:22 GMT

#15928

Thanks acrofales

that all makes sense

So, if my graph got really really huge, using lists of edges would be bad, right? Wouldn't sets be better?

edit: your last edit touches on my project requirements a lot. im gonna have to think about that.

Let's say I was going to have to implement djiksta's algorithm
What structure would you recommend then?

RoomOfMush

1296 Posts

November 13 2016 18:25 GMT

#15929

On November 14 2016 02:22 travis wrote:
Thanks acrofales

that all makes sense

So, if my graph got really really huge, using lists of edges would be bad, right? Wouldn't sets be better?

Why would sets be somehow better than lists?
An ArrayList is pretty much the fastest data structure for append operations and iterations. Only when it comes to random access insertions and deletions is the ArrayList somewhat slow (but still pretty fast).

On November 14 2016 02:22 travis wrote:Let's say I was going to have to implement djiksta's algorithm
What structure would you recommend then?

I would recommend having a Node class and an Edge class. An Edge connects two Node's. Each Node has a List of Edge's to other Node's which are connected. You can also have a Graph class which has a List of all Node's.

The List of Edge's is useful because it can be sorted by weight which makes it easy to pick the "cheapest" path first.

Deleted User 3420

24492 Posts

November 13 2016 18:28 GMT

#15930

Okay so I basically went with what I think you were saying, I am going to use an arraylist of edges.
yeah basically exactly what you are talking about.

there's so many ways to implement this lol

Acrofales

Spain18302 Posts

November 13 2016 18:28 GMT

#15931

The first. Just think about how that algorithm works.

E: had this reply open without posting.

Having both nodes and edges is a "complete" representation of a graph, but you'll find it's really redundant most of the time. For starters, at no point in Dijkstra will you be interested in knowing about all edges, so why bother representing them explicitly. Having a list of edges in each node is effectively the same as having a list of nodes in each node. It's just encapsulated in an imho over-engineered OOP manner.

The reason you want a Set (in my example, and your initial idea, the keys in the map) is so you don't have to check uniqueness of each edge.

Deleted User 3420

24492 Posts

November 13 2016 18:42 GMT

#15932

Okay cool guys it's all coming together now

I'll probably be back in a few days with more questions lol

Mr. Wiggles

Canada5894 Posts

November 13 2016 18:49 GMT

#15933

On November 14 2016 01:05 travis wrote:
I need to write a weighted graph in java. Any opinion on the best way to implement this?

I am thinking that my graph class will have a map members:


Map<node, Map<node, intege>>

I don't believe there's any technical difference between nodes and vertices. They're basically synonyms when talking about graphs. For example, I usually refer to vertices, but others use the word nodes.

In addition to what's been suggested, another possible graph representation is the adjacency-matrix:

https://en.m.wikipedia.org/wiki/Adjacency_matrix

This is generally useful for doing random lookups of if two vertices are connected.

On November 14 2016 03:28 travis wrote:
Okay so I basically went with what I think you were saying, I am going to use an arraylist of edges.
yeah basically exactly what you are talking about.

there's so many ways to implement this lol

Instead of repeated sorting to find the minimum edge distance, you can also look at using a priority-queue implementation.

https://en.m.wikipedia.org/wiki/Priority_queue
https://docs.oracle.com/javase/7/docs/api/java/util/PriorityQueue.html

This will generally be faster than re-sorting the list at every step of Dijkstra's algorithm. It's also a very useful data structure to know in general, when you want to keep a sorted queue of something where you're popping and pushing members.

Deleted User 3420

24492 Posts

November 13 2016 19:01 GMT

#15934

In java, is there a time you need to use "this" to specify what member you are referring to *other* than when you have duplicate field names? (like, are passed parameter "key" and have a class member "key" so in your constructor you do this.key = key ... which is obviously bad anyways)

phar

United States1080 Posts

November 13 2016 19:16 GMT

#15935

Yes. When implementing an equals() method, the canonical example uses syntax like if (this == other) { return true; }

RoomOfMush

1296 Posts

November 13 2016 19:27 GMT

#15936

On November 14 2016 04:01 travis wrote:
In java, is there a time you need to use "this" to specify what member you are referring to *other* than when you have duplicate field names? (like, are passed parameter "key" and have a class member "key" so in your constructor you do this.key = key ... which is obviously bad anyways)

There is a few obvious situations like returning "this" or passing "this" as an argument. Other than that it might happen when dealing with non-static inner classes that the inner class and the outer class have a method with the same name (or perhaps a member with the same name) and you need to refer to both of them, for example:

public class Test {
	
	public Test() {
		// ... do stuff
		ActionListener al = new ActionListener() {
			public void actionPerformed(ActionEvent e) {
				Test.this.actionPerformed(e);
			}
		};
		// ... do more stuff
	}
	
	private void actionPerformed(ActionEvent e) {
		// something something
	}
	
}

If you would leave the "Test.this." out you would get an infinite recursion.

Deleted User 3420

24492 Posts

November 13 2016 19:33 GMT

#15937

okay another weird question

I have an arraylist filled with objects
each object has a map

I find one of the objects
I do a shallow copy

I then update the map of the shallow copy

I then do a indexOf(shallow copy) on the arraylist

will it find the index of the original object?

It will, right?

edit: I don't even need to put the shallow copy in the arraylist do I? by updating the map I also update the map of the original object? Is that correct?

edit: yes. yes that is correct.

AKnopf

Germany259 Posts

November 13 2016 20:13 GMT

#15938

If you change the map of the copy, then yes, the map of the original is altered, too. This is because they both have a reference to the very same map. Primitive types are passed by value while complex types are passed by reference.

About indexOf(), it depends on how you implemented the equals()-Method in your custom class. The ArrayList just goes through its elements until it finds the first element that returns true for equals().
If you have not defined your own version of equals() and you are not inherting from a class that implements equals(), then you use the default implementation, which is for complexe types a reference check. This means that only the exact same object is equal. The copy is not equal.

Further reading:

Different variants of equality in java:
- http://www.leepoint.net/data/expressions/22compareobjects.html
- http://stackoverflow.com/questions/4178997/how-default-equals-and-hashcode-will-work-for-my-classes#4179011
indexOf() uses equals() http://stackoverflow.com/questions/22236009/why-does-list-indexof-use-equals-method-of-passed-in-object#22236559

Most importantly: Always also implement hashCode() when you implement equals() in a way, that two object that are equal also produce the same hash code. Otherwise you will land in programmers hell when you want to use your objects in a HashMap.

Cheers!

Update: Fixed sentence about reference vs value. Thanks, Manit0u!