Jump To Lightspeed Archive

Thread: X-Wing Space Superiority Fighter....NNNNOOOOTTTT!!!!!

DarthDiesel
Wed Nov 10, 2004 1:23 am
#40






ArianSix wrote:





DarthDiesel,


Actually, it *would* take up additional CPU cycles.


Think of the program as:


if(xwing) then

// blah

else if(ywing) then

// blah

endif


If they added adjustubale firing then we'd have something like:



if(xwing) then

if(autoconverge) then

// blah

else if(manual) then

// blah

end if

else if(ywing) then

// blah

endif



Just checking to see if the xwing is auto converge or manual requires 1 extra branch, then if manual, it would have to retrieve the angle and make the calculations.


The current system may not even be making big calculations to determine where the beams are because trig functions are expensive on CPU cycles. It's quite possible they have precalculated trig tables for whatever angles they currently use in order to improve performance.


Both ways, adding in adjustable convergence will be more CPU intensive than the current system.


Now, I'm not saying this will break the server as servers tend to be processing beasts, but I am saying that it is more CPU intensive which makes it technically morelikely to break the server under the new system than the current system.


So in summary, be nice and don't flame other forum posters because they could be right and it makes you look all that more silly for doing it.


-- A6


Message Edited by ArianSix on 11-09-2004 06:40 AM





Very cute. Did I say it would not require more processor power? NOOO.


All I said was it wouldnt really make much difference. AND I AM RIGHT.


I've written code before. While I don't know the details of processor costs for calculating a trig function compared to a look-up-table or whatnot, you have to be on crack to think this simple operation will in any noticable way decrease gaming performance, or "break the server."


I didnt fly off the handle for fun (frankly, I don't think my post was at all uncalled-for or outrageous - you guyshave some pretty thin skins) - I did it because the suggestion was frankly absurd.


Like I said, GIVE ME A FREAGIN BREAK!
DarthDiesel
Wed Nov 10, 2004 1:26 am
#41






Citairen_Mugarrii wrote:

I wonder where people get this whole "all the other games had adjustable convergence" crap from anyways.


The only one with that was Xwing Alliance, and that was part of what killed that game.


I am not however saying there was no convergence in the other games, it was just fixed (as it would be, the cannons are mounted, not on mini turrets). I can't be sure without opening up the files, but I believe the convergence point was around 750m with max range around 1500m (note, this would be where the shots met, the range where they would be inside the size of a Tie cockpit would be much sooner, also of note is that the range and speed values in JTL area little funny compared to the XWing series. Our spped measurements seem to be one tenth the x-wing equivalent, or rather, flying at 1000 speed in JTL would probably feel like 100 in X-wing. The distance I haven't found a decent exchange rate for yet, but the range seems rather short on the weapons.)







Oh for God's sake. Laser convergence killed the game?!?!? BWAHAHAHAHAHAAAAAAA!!!!! DUDE?!? What the HECK are you smoking??!!


In a lot of the lterature, and HECK, EVEN IN THE MOVIES, there are suggestions that that's how the X Wings laser system works. And it didnt ruin gameplay at all. The convergence in XWA was automatic, and as I recall, could be switched to infinite - THERE WAS NO DOWNSIDE!!


My God, you guys are pathetic.

EvanHalcyon
Wed Nov 10, 2004 1:48 am
#42

Just as an FYI, WWII Online does have this option. However, the convergence can only be set before climbing into an aircraft. If done in-flight, the new settings won't be implemented until you return to base and get a new plane. The command is .conv (enter distance.)


Personally this is something I would like to see in the game with the same principle. You can set the convergence before launching into space, or during space, but won't take affect until your next launch.


Flamers, you have the floor.



Finn Halleck

Gorath
Ducimus
Wed Nov 10, 2004 4:11 pm
#43

>>
"In this game, if one laser hits all four do, it may not look that way but that is the way the code is written."
>>

So let me get this stright. If i group two projectile weapons in the same group (which shows up as 4 blasters), and i fire both weapons at the same time, if one of the four blaster rounds hits, it counts as if all 4 blaster (both projectile weapons) hit? Even though it looks like only one blaster hit and the other 3 wizzed by the target?



Major Rapax Victis - Commando
Time wasted playing MMO's:
Meridian59, Ultima Online, Everquest, Asheron's Call 1 & 2, Dark Age of Camelot, Anarchy Online, Earth & Beyond, Eve Online, Planetside, Star Wars Galaxies

"I'm really bored."
...has mastered the Pilot profession
Dreamland
Wed Nov 10, 2004 4:16 pm
#44

As the person in question,I have no knowledge of programming or game design past making my own levels for FPS games. I am not educated in the world of programming. However it seemed like simple logic to me that to have to combine target prediction that is already not working, (the lead indicator is out to lunch on account of lag) With an infinate angle of convergance on a ships blaster, would seem to be a bit of an excess to me.
QBallz
Wed Nov 10, 2004 4:41 pm
#45

Actually this would require virtually no processing power if you want to get into it. I mean every shot is already tracked regardless of covergence, so really the only change would be the angle fired relative to the ship, so basically treat each weapon like a fixed turret.


The only extra information would be keeping those custom angles for each player, and I really don't think that would be much of an issue. I mean they already track every shot that is fired as far as who owns it.
Dreamland
Wed Nov 10, 2004 5:48 pm
#46



But do they track each single bolt? A tie interceptor fires 4 bolts with 1 gun. I have heard people say that if even one of the bolts hit it does 100%damage. Are they tracking every single bolt or is the pattern nothing more than a cosmetic representation for continuity? The 4 bolt spread from the interceptor just flies outward to its disapearance point in a perfectrectangle with the bolts pointing inward. Does each single bolt on a tie interceptor have a 25% portion of the single weapons damage attached to it? or is there just a 4 pointed hit or miss roll in place?


Now would it not be more complex with every bolt moving in an arbitrary direction set by the player based on the range of the target? I realy wasn't speculating so much about processing power asI was about latency issues for that matter.

Message Edited by Dreamland on 07-08-2003 04:54 PM

DarthDiesel
Wed Nov 10, 2004 6:23 pm
#47















ArianSix wrote:






DarthDiesel wrote:

Sorry man - I don't know what world you live in.


Apparently I'm forced tolive in the REAL one.


Sure, multiply it by the thousands of users on a server (galaxy) and it can add up. You seem to forget that, compared to ALL THE OTHER CALCULATIONS GOING ON, this one would be practically insignificant!! The processors in question can handle A LOT LOT LOT LOT more than such a simple calculation, even thousands of such calcs.


You keep insisting that the calculation is simple. Trig calculations are *not* simple. If you ever programmed 3D graphics you'd know this.


I've been doing experiments at home on the calculations involved in non-hardware accelerated (i.e. no 3D graphics card) trigonometric graphics manipulation (i.e. skewing 3D points into 2D space, 3D point rotation, texture mapping onto a quadrilateral)... the performance hit from computing these functions using trig is large and noticable. In this case we're talking a single extremely high end machine with only one user on it.


Given an arbitrary angle, a function cannot be optimized because the trig needs to be recalculated each time the angle changes. Give a fixed rotation (i.e. 45 degrees), performance drastically increases because you can whip out a calculator before hand, figure out the answer and plug the number directly into the program... avoiding the trig calculations entirely.


Again, take these calculations, multiply them by thousands and you do create a huge amount of additional load.


Will it crash the machine? Maybe not immediately... however it will make a significant impact on processor costs / player which means the machine will take much less punishment before hitting max capacity.


Clueless my ass - look in the mirror. Like I said, compared to all that crap you claim will crash a server, that simple calculation is like the risk of breaking a car's suspension by choosing to carry a 12lb book with you instead of a more efficient 10lb, or even 2lb book, where the car's total mass is like 3000lbs.


And next are you going to claim that a computer will never need more than 640k of RAM?


Putting a 12 lb book in the car will not significantly affect it's performance. Putting 2,000 of these books into the car will.


You're looking at the server and you're looking at yourself... but you're forgetting that there are thousand of other people on the server doing the same calculations at the same time.


As an oversimplification:


max People served = server capacity \ load for each person


if current people served x load for each person is greater than server capacity you have problems.


Obviously, as load increases, max people served decreases, and the ease of surpassing max capacity increases.


Again, I never claimed no performance cost. Why not use look up tables to do the auto convergence? I don't care what method you use - you admit right there an easy way to include it. Heck, why not just let the user's CPU handle the computation, and send the convergence angle value or whatever to the server?


Again, lack of understanding of programming. Look up tables are best when the angles are set (I wouldn't be surprised if that's what's happenning now). Adjustable angles means either the requirement of many more look up tables or a dynamic trig calculations. I never did admit there is an easy way to do this, I said that look up tabes are fast and good with fixed angle calculations.


Why not let the user's CPU handle the computation? If you haven't realized this already, most of the processing happens server-side on a MMORPG to keep people from cheating. Also, it is imperative that the calculations be done serverside because this information is needed not just by you, but by the server and every other player in the area. If you're computer is suffering from poor performance we don't want this to affect the other players around you... that's why this is happenning serverside.


With lag (network overhead), question performance of the game machine, and a necessity of everything around you to have access to that information, it's just not realistic to have to wait up on your machine.


Think of PvPing on the ground. The people with 3 Mb/s connections havea distinctive advantage over the guy with a 28.8 modem because they're getting the information faster. You don't want these guys to sit around and drop to 2 frames / second just because another player in the area has a lousy machine/connection.


If you want to complain that such a simple game feature would risk a server crash, then lets start adding up the THOUSANDS OF OTHER simple little code features that are included in the game, and judge them by the same standard. Bottom line is, those thousands ARE the standard I'm using - this simple addition would be puny compared to all of them.


Stop insisting on simple addition. What would you consider complex? Multiplication? Division? In theory you can pretty much consider all of these variations of addition (subtraction is also addition in a sense). Yes, every feature costs cycles, and no change, by itself, will crash the server per se... it's about the big picture... everyone using those changes at the same time.


Also again, if max load is considered a constant any increase in load leads you that much closer to the demise of the machine, process, or whatever.


Yes, system resources are important. AND THEY MUST BE BALANCED OUT WITH GAMEPLAY VALUE!!!


You remind me of a programmer here at work who claimed to be the is-all know-all of programming... he turned out to know absolutely nothing, got a lot of projects in trouble, and is now history.


If you haven't realized it already, I'm not against the use of the extra trig functions. What I'm fighting for in the person you stepped on - whos points are actually well grounded - with your self-acclaimed computer proficienct which is nothing but wrong and ill-grounded.


And since you clearly have problems comprehending reality - I wasnt the first to flame.I didnt attack any posters at all, just made a statement. Then I'm told I'm being an ass:


You were being an ass and have been good at making a point of being one. You have no concept of the reality of computer programming, architectures, concepts, and graphical manipulation... so unless by "reality" you mean "never never land", I'm sorry but you may be confused.


One thing I don't stand for is other people getting flamed for absolutely no reason. Whether or not if that's how you see it, you were flaming that person.


If you're going to tell someone they're wrong, first, be polite about it and, second, make sure you right before you do it.


As my late grandfather used to say:


"It's better to keep your mouth shut and be thought a fool than to open your mouth and prove it".





-- A6










I'm sure you have more 3D programming experiencing than I do.


I'm a well-studied mechanicalengineer and have no problem understanding the issues you raise. I appreciate your transparent desire to show me up. I don't know nothing of this technical issue, and so I didn't make my remark lightly.You still havent really debunked my point, nor do I think you can, because my convergence suggestion is perfectly doable I'm sure. In fact, you reinforced my point -thereare no solid grounds for claiming my suggestion wasnt worthwhile based on "processor" loads.Sure, it would increase processor load, but to an extent not worthwhile? You didnt demonstrate the distinction, given all the other little featuresa server processor must handle that I and many others might consider not worthwhile.


Your example of why the user's computer couldnt participate is apparently inconsistent. So, you seem to state that PvP ground combat, as currently in the game, rewards fast connections. If it does, youve defeated your own contention. Separately, there are plenty of things the server relies on the player's computer to send I'm sure - I highly doubt one more small piece of data will make much difference, and don't doubt contingencies can be adopted for the concerns you raise. There are several other technical clarifications I might care to discuss. However, your preoccupation with pedantic pretense now leaves you too inclinedto defend it. It sounds like your grandfatherwas a wise man - you might try actually following his advice.


Cheers.

Message Edited by DarthDiesel on 11-10-2004 07:13 PM

DarthDiesel
Wed Nov 10, 2004 6:26 pm
#48






Dreamland wrote:

As the person in question,I have no knowledge of programming or game design past making my own levels for FPS games. I am not educated in the world of programming. However it seemed like simple logic to me that to have to combine target prediction that is already not working, (the lead indicator is out to lunch on account of lag) With an infinate angle of convergance on a ships blaster, would seem to be a bit of an excess to me.







I never suggested it need be infinite angles. Surely, a finite number of angles from a look-up table would suffice, especially since our weapons can't hit targets approaching infinity, which I imagine might run into significant figure concerns.
DarthDiesel
Wed Nov 10, 2004 6:28 pm
#49






QBallz wrote:

Actually this would require virtually no processing power if you want to get into it. I mean every shot is already tracked regardless of covergence, so really the only change would be the angle fired relative to the ship, so basically treat each weapon like a fixed turret.


The only extra information would be keeping those custom angles for each player, and I really don't think that would be much of an issue. I mean they already track every shot that is fired as far as who owns it.






Right, and I still don't see why the user's computer couldnt do the calc, and send a small data packet to the server, or whatever, with the necessary identifier value from a predetermined set of angle values. Seems freagin simple to me, but gee, wouldnt want to overload the server (rolleyes).
DarthDiesel
Wed Nov 10, 2004 6:31 pm
#50






Dreamland wrote:



But do they track each single bolt? A tie interceptor fires 4 bolts with 1 gun. I have heard people say that if even one of the bolts hit it does 100%damage. Are they tracking every single bolt or is the pattern nothing more than a cosmetic representation for continuity? The 4 bolt spread from the interceptor just flies outward to its disapearance point in a perfectrectangle with the bolts pointing inward. Does each single bolt on a tie interceptor have a 25% portion of the single weapons damage attached to it? or is there just a 4 pointed hit or miss roll in place?


Now would it not be more complex with every bolt moving in an arbitrary direction set by the player based on the range of the target? I realy wasn't speculating so much about processing power asI was about latency issues for that matter.


Message Edited by Dreamland on 07-08-2003 04:54 PM







It wouldnt be some "arbitrary" direction, where suddenly there are 4 completely discrete objects for the processor to worry about instead of one. A simple correction factor can be applied to a source object (theconvergence pointvector)to determine4 distincteventoutcomes. Nominally, all 4 laser blasts would hit the exact same point on the target.


Now, as a matter of fact, people are saying the current appearances are just that - appearances, and the appearance of 4 bolts is just for looks, while actually they impact as if they wereone blaster bolt. If that is NOT true, then it would appear the game must already concern itself with 4 individual blaster vectors. So, what a convergence system might do is simply multiply the vectors by the appropriate adjustment/correction value so theyconverge on one point in space. Been a little while since vector calculus, etc., so I'm just throwing out thoughts which may not be 100% on the ball.

Message Edited by DarthDiesel on 11-10-2004 05:36 PM

Dreamland
Wed Nov 10, 2004 6:33 pm
#51


I'd also like to point out i never said it would overload the server, i said it seemed like it would be overload. A descriptive term meaning that it would take much more effort on the developers part to code for it and make it work realy well taking latency into account. Especialy considereing there are targeting problems now relating to latency with the system that is in place.

Message Edited by Dreamland on 11-10-2004 05:35 PM

gera
Wed Nov 10, 2004 6:58 pm
#52

Point is, I cannot handle X-Wing myself in Jump to Lightspeed. I like my Y-Wing better with 'better' mass to put more higher weapons. Currently NPC AI is just and just drawing a donut in space. Turn and turn and turn. With limited capability of Rebel ships turn rate against NPC Tie's and you can never compete with NPC TIE army alone in a crippled equippment because you cannot put equipment of your level vs. your level TIE army.




Armor removed from Jedi so they can be kited 'as designed' - Blixtev

Page 4 of 5