« first day (704 days earlier)      last day (2711 days later) » 

12:28 AM
@RavenDreamer may I assume vecInput is storing your final result?
 
side question, I thought you were working with unity, why do you have xna tag in your question?
 
Because the question is not unity specific, and XNA has a higher follower count than C#
 
then you should not add any of those tags to your question.
you can't add any tag to your question, unless it describes what the question is about.
 
XNA uses C#. It does describe the question.
 
 
3 hours later…
3:22 AM
@RavenDreamer are the original angles always a cardinal direction?
if you want an anisotropic function that rescales a vector I'd do something like
scaledMagnitude = Sqrt((v.x / maxXSpeed)^2 + (v.y/maxYSpeed)^2);
if (scaledMagnitude > 2)
   v = v * 2 / scaledMagnitude;
 
3:42 AM
You know what I hate... tutorials with horrible variable names.
 
variable names such as "v"? :P
 
foo bar, v, variableprefixRealVariableName, etc.
 
4:18 AM
@tylerrrr07 then what do you use suggest for variable names?
 
@Gajet - I've got basic mouse input working now. One step closer to that RTS :P
 
:D
 
I can click stuffs and select stuffs using a drag + click
 
yesterday I've talked about same idea with a friend of mine,
 
And how about... DescriptiveVariableNameBasedOnWhatItsForWithoutAnyPrefix
 
4:18 AM
@tylerrrr07 that's too long
anyway it seems he once had that idea and implemented a tech demo with gamemaker
 
@Gajet I'm speaking in generic terms. But in all honesty... if that's what it requires for me to figure out what the variable does, then its just long enough
But to be specific, the example I'm looking at prefixes all of its class level variables with the letter m
 
though he got off the track after a little while and it ended up a game which you had to design some objects and their behaviour
 
Makes Copying / pasting / formatting to my specific needs a pain
 
I like prefixes.
 
Prefixes are OK when it comes to UI - otherwise meh... I'd rather not
 
4:21 AM
they usually declare variable scope. or variable type.
 
especially for something as trivial as scope
 
I use "m" prefix for fields too
scopes are not that trivial.
specially when you class contains methods with 50+ lines of code
 
Fields and local scoped variables start with a lower case letter for me. Properties start with an upper case letter.
And thats sufficient for me
 
this happens when you are developing core features
 
Yeahhhh... We're going to have to agree to disagree here
 
4:23 AM
for me all Fields start with "m", all methods and local variables and functions start with lower case letter, and all classes start with upper case.
 
Methods and Classes both start with Upper Case for me
 
I don't have set naming rules but I probably use what tyler uses
 
If it isn't obvious which scope a certain field / local variable belongs to (class or local) then the class / method design probably isn't quite right
 
that's pretty close to .NET standard style
 
@Jimmy - That's pretty much where I got it :)
 
4:26 AM
fancy naming systems make it harder for me to understand code
might as well alternate between rEVERSEcAMELcASE and igPay atinLay.
 
I'll never understand or even remotely like web development. :(
 
@WilliamMindWorXMariager - What's not to understand?
 
@WilliamMindWorXMariager it's not you, it' web developments problem :D
 
According to html5test.com, my phone supports canvas and 2d.
Running this in my browser: legeria.net/l2r
Works fine.
Running it on my phone.
White screen.
Now, I'm using firebug, to make sure there are no errors on the page. And there's not.
 
Blue BG + Green Box?
Is there more to it?
 
4:30 AM
looks fine in iPhone4
 
@tylerrrr07 Nope that's all.
 
Works on my Xperia Play
 
Only special part about the page is, it'll retain the same width/height no matter the resolution, as in, not scale.
I'm trying to test it on my Windows Phone 7.5.
Which, like I've mentioned, should support canvas with 2d context.
 
windows phone is using IE core, which I doubt has a version more than 8
 
Well, bad news, it doesn't work on IE 9
Which also supports Canvas according to that site
 
Maybe there is a specific implementation required for IE9
 
SCRIPT438: Object doesn't support property or method 'getContext'
oh wait
it's in quirksmode
 
@WilliamMindWorXMariager - Try including <!DOCTYPE html>
 
you have to doctype the demons out of that one
 
@Jimmy - I saw that
 
4:34 AM
Where should I place that? Top of the page?
 
Parsing rules 1
<!DOCTYPE html> triggers standards mode »
Yes - At the top
 
absolutely before anything else
 
Yay for wanna be game devs that are web developers IRL :P
 
Hehe
Next step is getting it to actually fill the page properly like it does in firefox.
So, should I test with IE9, and assume I get the same issues in it like I would on the phone?
I guess I could just download IE9 and test. :P
Anyways, I'll be back soon, you kids are awesome. <3's and huggles.
 
4:39 AM
Windows Phone is a mobile operating system developed by Microsoft, and is the successor to its Windows Mobile platform, although incompatible with it. Unlike its predecessor, it is primarily aimed at the consumer market rather than the enterprise market. It was launched in the second half of 2010, with a release in Asia following in early 2011. With Windows Phone, Microsoft created a new user interface, featuring its design language called Metro. Additionally, the software is integrated with third party services and Microsoft services, and sets minimum requirements for the hardware on w...
(Link goes to the web browser section)
Long story short - it uses IE Mobile with a rendering engine based on IE9 - so testing with IE9 first is probably a good idea but not a guarantee.
IE9 also has a developer toolbar built in that's similar to firebug (press F12) so you can use that for troubleshooting some.
 
Thanks @tylerrrr07
Right now I have a white bar in the bottom when in portrait mode, but not in landscape mode.
I guess I should see if I can detect portrait/landscape mode. :P
Since it doesn't fix the resolution right now.
 
 
1 hour later…
5:57 AM
Yay
Basic rendering loop is working: legeria.net/l2r
All written in SharpKit. <3
 
6:10 AM
@WilliamMindWorXMariager is it 6am your time?
 
Right now?
It's 8:10 AM
 
yes
 
 
6 hours later…
12:10 PM
@Gajet - I may have an idea for an RTS
Kind of...
It's more like I have an idea which involves writing an RTS toolkit that allows you to create custom Tech Trees using Custom sprites / animations and then I'll make the game that uses them
 
12:32 PM
Hey guys
 
@tylerrrr07 I'm listening
you mean I have to design the tech tree?
if that's the case, there is no need for any toolkit.
I can directly write all the tree into a file.
and you (or again me) can write a parser only to draw that tree.
 
1:02 PM
@Gajet - No I mean the players can design their own tech tree if they want
In my head I have a vision of notebook doodlings, stick figures with different themes (Pirates, Ninjas, Medieval, etc.) to start out with, and then a little wizard for creating their own and importing their own doodles as graphics
@dbomb101 - Hey
And then I thought it would be neat for online play where you could on the fly download the other players tech trees and use them kind of like a downloaded map
The tech tree building could have some sort of score system so that you can't just make a super OP tech tree (though without really thorough testing I imagine this could happen anyway)
This is just what my imagination came up with when I was trying to sleep last night
 
+1 for player customization :)
 
1:25 PM
hey guys
do you mix TCP with UDP for example in FPS game, when there's need for e.g. chat messages ?
according to my question on SO:
http://stackoverflow.com/questions/11361208/high-performance-scalable-udp-servers

the man says that "Using TCP for chat messages and loading level data at startup looks like a good idea, that's an usage pattern TCP is good at."
and again here
http://gafferongames.com/networking-for-game-programmers/udp-vs-tcp/

though Glenn says that "My recommendation then is not only that you use UDP, but that you only use UDP. Don’t mix TCP and UDP, instead learn how to implement the specific pieces of TCP that you wish to use inside your own custom UDP based protocol."

Later in the comments he replies about mixing the protocols that "now, of course if you just had the occasional event sent via TCP, and the bulk of your time sensitive data in UDP, no problems at all"
or should I just ask separate question on this site ?
also, high five gustavo, I'm too 16
 
@Anonymous It all depends on what, exactly, you expect your networking system to accomplish. In my opinion, implementing TCP features on top of UDP just because is probably a painful and unnecessary exercise.
 
yea I also think so
just wanted to make sure
anyway it's good news
 
1:40 PM
That being said, if (down the road) you determine TCP to be sub-optimal for your needs, you can look at building a layer on top of UDP.
Bottom line: it's a premature optimization issue.
 
it always is
lol
 
Not always. Sometimes it's an architecture/design issue.
Such is the case with the problem I've run into.
 
@TreDubZedd That's my fear. The architecture issues.
I just fell into one too.
 
I've got two threads: A and B. Thread A produces data at a given interval (and I'll call each interval a frame). Thread B produces data at multiples of that interval (say, 2x, 3x, 5x and 7x). How do I determine when a frame has all its appropriate data (from A, B or both)?
I can't just use A to trigger the "end-of-frame" event, since its data might come in before the data from B.
I can't just use B to trigger the event, since there may not be any B data associated with a frame.
 
And making thread B wait for thread A may be pointless.
very interesting
 
 
2 hours later…
3:23 PM
@Anonymous, The reason you shouldn't mix, is that TCP can block the UDP traffic. Which is why, using it for chat and level loading should be relatively safe, but using it to stream anything would be a bad idea.
 
That's assuming you're using the same port for TCP and UDP, right?
I can see that being a very bad idea.
 
3:48 PM
@TreDubZedd hm? which one is very bad idea ?
 
@Anonymous using UDP and TCP on the same port is a Bad Idea™.
 
@TreDubZedd do you absolutely need to synchronize A and B frame by frame?
 
4:13 PM
@TreDubZedd Does the same port really have anything to do with it? Ports are only logical
 
4:35 PM
@JohnMcDonald I just don't know, anymore :)
@Jimmy Yes. I absolutely need all the data from A and B for a given frame to be collected together--even if that means "nothing" from B.
 
4:46 PM
but A and B are on different threads
so are they synchronized to a "tick" somewhere?
 
Yes...
both sets of data come from a process which is spawned by my application. One set comes via StdOut, the other via TCP.
Both have a timestamp
 
so both dump data to a threadsafe queue, you write out all available in the queue right before the 'tick'
oh, you're matching by timestamp. So possibly the TCP ones could be delayed and you have to wait for it
 
right
So the question, then, is "When do I wait for B--or when do I process A immediately?"
 
unless you can look into the future, you can at best have a timeout window
it will add latency to the whole system though
delay A by up to 50ms or so, the train leaves if B doesn't show up
 
I can look into the future, in the sense that I know that B will produce some data every 3 frames, and some every 5 frames, etc.
 
4:53 PM
so I don't see wyh your question is "when do I wait for B or do I process A immediately"
 
That is, multiple queues are funneling into B
 
so B produces on frames 3, 6, etc but also possibly on 1,2,4,5 but you don't know
 
well, it's more that I expect a set of data on frame 3, and a different set of data on frame 5...and both sets on frame 15...
and nothing (from TCP) on frames 1, 2, 4, 6, etc.
 
this isn't secretly a FizzBuzz interview question is it?
 
haha it would be, except for the A thread
...and the possibility of more than just Fizz and Buzz
there can be n intervals, all funneled into B.
Actually, the mod solution is the only thing I've been able to really come up with.
I'd just hate to have to iterate through all my intervals on every frame, if I don't have to.
 
5:03 PM
how many intervals are you speaking of?
mod is pretty cheap
 
n > 100?
 
Realistically, anywhere from 0 to thousands
 
if you have thousands of n, would you have data almost every frame?
 
Assuming they're not all higher primes, yes.
(which is really a pretty safe assumption)
Hm. Here I was getting on Anonymous's case for premature optimization...and I was doing the same thing...
Thanks for being a sounding board so that I could clear my head, @Jimmy
 
5:07 PM
Sounds more like an architectural thing. What's this for?
 
I'm running a satellite simulation within another application, for analysis
 
Ah, ok
 
np. I had to do something similar (but just combining TCP streams) but I had to use a timeout window since I didn't have any foreknowledge of periodic behavior
 
I was imagining a strategy game where each unit (n) has its own reporting cycle
 
Yeah...not gamedev-related, but you're a bunch of pretty bright guys, so I figured I'd ask. It turns out, as a side-note, that gamedev is actually pretty similar to satellite control systems.
 
5:10 PM
in a game, you don't need a thread-per-unit. You also have the option of double-buffering the game state.
 
Oh I know, that's what I was thinking
 
5:25 PM
I feel more safe knowing TreDubZedd has to comply with ITAR. I wouldn't want Iran to get ahold of our FizzBuzz satellites.
 
6:22 PM
I feel like I should propose FizzBuzz as the name for our next satellite project.
 
6:34 PM
Couldn't hurt
 
6:56 PM
I'd pay for a FizzBuzz satellite
 
Maybe I'll propose the idea to SpaceX. They can call the launch vehicle, "Mentos&DietCoke"
 
oddly enough, this server doesn't seem to have a reasonable limit on input
I just asked for the first 20 million numbers. it's taking a while
 
oops Error 500
 
It seems to get confused if your input is backward
 
7:03 PM
well the first 200k came back
 
 
1 hour later…
8:29 PM
@Gajet - I don't think I want to do an RTS.... lol
Boy do I lack motivation
 
Valve increasing the efforts for indies to get on Steam: store.steampowered.com/news/8367
5
 
8:47 PM
@tylerrrr07 I can say that RTS games aren't easy. You either have to have networking for hundreds of units, or some great AI, not to mention collision avoiding/detection against moving targets and group-based pathfinding
@WilliamMindWorXMariager Neat
 
9:21 PM
So, curious though.
Lets say we have a system, similar to redstone in Minecraft.
 
Are you implementing redstone?!
 
I'm implementing something similar yes.
I was considering, since it runs on the server, maybe it's worth parsing the circuit into a C# function and then compile it.
Eh, it goes on the thoughts document. :P
 
:/ does it work in 2D?
 
Premature optimization if I ever saw one.
 
I don't think it would be the same
 
9:24 PM
Redstone?
 
I was thinking of using wirebusses to avoid the whole missing depth part.
And multiple colors, that can cross paths.
 
So you could have red wire and green wire, and they could travel in the same tile and cross paths and such.
But, wires will likely just serve for powering.
 
well, that's all they did in Minecraft too
 
9:25 PM
Signals would be wireless with a limited range of course.
And repeaters for more distance.
This way, you can cut the power supply for a circuit or kill the connections.
More PvP aspects.
 
yeah, kinda reminds me of this demo video of a really neat indie spy game. Let me find it
Here it is: Gunpoint: youtube.com/watch?v=T6a0WR2-tLg
 
You are allowed to get all hot and flustered.

A layout framework for [XNA](http://create.msdn.com/) on Windows & Windows Phone 7 from [Red Badger Consulting](http://red-badger.com/blog/tag/xpf/). Inspiration is drawn from the [WPF](http://msdn.microsoft.com/en-us/library/ms754130.aspx) and [Silverlight](http://www.microsoft.com/silverlight/) XAML stacks, [Reactive Extensions](http://msdn.microsoft.com/en-us/data/gg577609.aspx) is used to keep things nice and fast.
aaand the MD failed of course :)
in about 2 weeks we will be pulling the trigger and FOSsing the project
 
Sexy. :P
Also, markup fails when you have linebreaks.
 
9:47 PM
ah
 
Cool
You see the wiring system that Gunpoint used @Will?
There's a good example @1:54
 
Yeah, it's something just like that I'm planning.
I got the idea from WireMod a mod for Garrys Mod 10.
 
Oh yes
 
So basically, various components that can be stuck together. Some needs power. Like, a laser turret. Can be placed in a tile, with a direction. Needs power to work and signal to shoot.
The tile automatically acts as the housing.
So picking a better material makes it more resistant.
More advanced laser turrets could perhaps have a scripted direction.
 
10:01 PM
So the turret would have multiple inputs, like fire, rotate left/right and outputs, like current direction and online/offline.
 
man, I want to buy Gunpoint. Why can't they sell me an alpha like Overgrowth does?
 
@tylerrrr07 only one question, how do you prevent tool abuse?
I mean what if someone creates a tech tree with only one node unlocking everything?
now that I read rest of your idea, it seems tech trees are constructed before the game start.
that's a nice idea too, but quite different with what I had in mind.
@tylerrrr07 do you need motivation again? finish my game :D
 
10:17 PM
the wiring stuff reminds me of the early demos for Love
 
 
1 hour later…
11:41 PM
@WilliamMindWorXMariager sounds really cool.
 
11:54 PM
I think voxel/tile building games are best when they have a purpose beyond building.
 

« first day (704 days earlier)      last day (2711 days later) »