NewCommer's First Impressions

[Per]

Note that the problem does not lie in path-finding, because optimal paths are found quickly using threaded A*. The problem lies in path walking, ie following the path and dealing with obstructions. For those who want to follow the discussions, there are at least two tickets in the ticket system dealing with this, with several different patches, already.

[Zarel]

Note that the problem does not lie in path-finding, because optimal paths are found quickly using threaded A*. The problem lies in path walking, ie following the path and dealing with obstructions. For those who want to follow the discussions, there are at least two tickets in the ticket system dealing with this, with several different patches, already.

I find this distinction to be somewhat absurd, considering nearly every source I know of considers path walking to be an aspect of pathfinding. I mean, especially if you're not walking the exact path calculated by A*, then the path you do walk is technically the path "found" by the game engine.

I think it's easier just to refer to the whole thing as pathfinding, and refer to A* as "path calculation" or just "A*" or something like that.

I don't know, but it's certainly something that should be investigated

It already is being investigated:
http://developer.wz2100.net/ticket/1649

[stiv]

I find this distinction to be somewhat absurd, considering nearly every source I know of considers path walking to be an aspect of pathfinding. I mean, especially if you're not walking the exact path calculated by A*, then the path you do walk is technically the path "found" by the game engine.

I think it's easier just to refer to the whole thing as pathfinding, and refer to A* as "path calculation" or just "A*" or something like that.

If the distinction is difficult for you, think of pathwalking as 'navigation'. Pathfinding (the A* part of the process) simply finds a path between nodes of a graph. Navigation involves going from Point A to Point B in a changing dynamic environment that might involve small moving obstacles or local barriers too small to appear on the graph.

[dak180]

There is a library that we could use to do pathwalking: OpenSteer.

The real issue is whether the work necessary to utilize the library would be worth improvement that it would bring.

[Zarel]

If the distinction is difficult for you, think of pathwalking as 'navigation'. Pathfinding (the A* part of the process) simply finds a path between nodes of a graph. Navigation involves going from Point A to Point B in a changing dynamic environment that might involve small moving obstacles or local barriers too small to appear on the graph.

I know what the distinction is; that's not the issue.

The issue is that everyone else refers to path walking as if it were an aspect of pathfinding, and every once in a while Per corrects them by saying that it isn't.

Now, regardless of who is technically true, the simple fact is that it's far easier simply to go along with the trend, and include path walking as an aspect of pathfinding, and create a new term such as "path calculation" to refer to the A* part of pathfinding.

Until now, I've really just tried to avoid using the term "pathfinding" at all, precisely because of this problem, but I've come to realize that it's really quite futile. We do need a term for "the code that causes units to move from point A to point B." I've been using the abbreviation "pathing" in the hopes that it's vague enough for Per not to yell at me, but if everyone else uses the term "pathfinding" for that, why not?

[stiv]

I've been using the abbreviation "pathing" in the hopes that it's vague enough for Per not to yell at me, but if everyone else uses the term "pathfinding" for that, why not?

Why not be clear so we can know what we are discussing. After all, programming (and magick) is all about knowing the True Names of things.

It is important here because path finding (A*) is algorithmically different from path walking ( stumbling from A to B). A* says "cross the street to get to the bank". A to B says "don't run into that mailbox. Hey, watch out for that truck!"

[Zarel]

It is important here because path finding (A*) is algorithmically different from path walking ( stumbling from A to B). A* says "cross the street to get to the bank". A to B says "don't run into that mailbox. Hey, watch out for that truck!"

But why not call that something else, like my proposed "path calculation"?

As in:
Path calculation says "cross the street to get to the bank". Path walking says "don't run into that mailbox. Hey, watch out for that truck!" Pathfinding says both.

Heck, I'm pretty sure my definition is the correct definition:

http://en.wikipedia.org/wiki/Pathfinding
"Pathfinding in this context concerns the way in which a moving entity finds a path around an obstacle" - doesn't mention whether or not the obstacle is moving, and appears to refer to the path the entity actually walks, not the one it calculates.

I mean, really, why should the path that was calculated actually matter to pathfinding, if it's not the path you end up walking?

[bornemix]

... I'm pretty sure my definition is the correct definition ...

I'd say it's the definition...

[Per]

I don't really care which definition is used, and referring to dictionaries is really a waste of time. The point is that what I have chosen to call path finding and path walking refer to two very different parts of the code, and when I started making the distinction was when I saw people were looking at the wrong part of the code to solve the issues with movement.

(But really, when you read the wiki page on path-finding, you see that it is all about finding a graph between two points. It has very little to do with the actual movement between the two points. OpenSteer, on the other hand, is all about movement, and they call it 'steering' and 'path following'. So I think my definitions are both correct and accurate.)

[123vtemp-]

@ Zarel

Is there a way to increase the screen scroll with out making your mouse scroll faster atm?
Screen scrolling seems t be stuck on a low speed.

[123vtemp-]

btw I find current pathing and group formation very agreeable

[bornemix]

What's the difference between screen scroll and mouse scroll? Are you referring to zooming?

[Zarel]

Is there a way to increase the screen scroll with out making your mouse scroll faster atm?
Screen scrolling seems t be stuck on a low speed.

Sorry, it's not adjustable. You can try the NUM- and NUM+ shortcuts, or wait for the new scrolling system we're working on.

[cybersphinx]

Is there a way to increase the screen scroll with out making your mouse scroll faster atm?

There's a scrolling speed setting somewhere in the options, is that what you want?