The future of SF?

[ZoltanTheHun 101]

I do have to wonder what TK pays himself. For instance, if as mentioned here he does 90%+ of the coding himself, and he says "to put in feature X will cost $100,000," and coding is all labor, what is he saying about length of time to make the feature? Does he mean it's going to take 2 years, or that for 9 months of work it will cost $100k? Um, you charge yourself too much, dude.

 

 

I am a coder in an Eastern European country(HUN), and I several times played with the idea of going solo and start my own company, not necessarily a gaming company, because that is probably  the most volatile option out there. Without the security of some kind of corporate backing you need to think about future savings as well, because you never know when you get into the situation that you cannot work for month(s), without any paid holiday or insurance. I calculated about  this several times, but I think If I want to get the same kind of options with the paid holidays that I live on now, I must make at least twice as much as I earn at the moment. Just think about that companies need to keep up all kind of different not money making facilities  like HR  next to the pay of the real moneymakers. Even if TK is just solo, and works with contractors maybe, a lawyer and a bookkeeper is a must. So 100k might look good on paper, but if you start to spend that money on all

 

Maybe I am wrong, but I do not think that TK is greedy, I rather think that he is just trying to survive, and he found some more profitable opportunities in the mobile business. And if that is truly profitable, that doesn't mean that SF3 sees the light one day. I say this, because he is a simmer, hack, I cannot imagine someone working on a sim for more than a decade, if he is not a simmer himself. So I would expect him, that if he is financially satisfied he can spend more time on what he enjoy: making flying games.

[PFunk 198]

To my mind, and this is only my opinion, the ne plus ultra of the SF2 series was the Desert Storm mod.  It represented the best of what the native engine could do, using the resources that engine was most familiar with.  Cold War 70s and 80s-era warhorses and period-accurate SAMs and ECM environment.  It makes the best use of what the stock engine and programming of the SF2 series can render.  Anything more modern than that requires some serious fiddling with the capabilities and offers a much more limited rate of return for effort.

 

Anything of a newer era of air combat is going to require a significant rewriting of the avionics code in order to really be able to render the world in a way that's as entertaining as the SF2 engine is known for being and as believable as we would like for it to be.

[beachav8r 360]

Agree with both Zoltan and PFunk's last posts. While it is probably a fruitless endeavor to try to get inside TK's mind, I think it is probably a fair assumption that he wants to provide the best user experience and still put a roof over his head simultaneously. I have yet to meet a developer that is a sim enthusiasts that doesn't want to put every bell and whistle into a sim to put their stamp of awesomeness on it. Those that have been in the business for a decade + (like TK) temper that enthusiasm with the reality of raw monetary flow. Who wouldn't? Razorworks was awesome - where are they? Jane's was awesome - where are they (please don't answer that..it hurts to know). Microprose was awesome - where are they? So we have to live with the fact that longevity and concessions go hand in hand.

 

That said...LOL..the bottle of wine is talking now..forgive me. I can give you a couple insights. Developers DO care what we banter about here. You might think that we represent the vocal minority (and statistically we do), but devs are no different than us. We want feedback, affirmation, and measured (read: polite) critical commentary. Where else are you going to get feedback than online forums? The people that play 10 hours and shelve a title aren't going to be expansive in their lists of pros and cons. But you also have to throw out the outliers - the guys that have flown 20 minutes and the guys that have spent their entire lives flying the same sim - neither represents your target audience which is (usually) the solid middle. Does TK read this stuff? Probably. Some birdie probably taps him on the shoulder and points him in this direction whether it is praise or infuriating criticism. Hopefully, the ratio of good to bad is equivalent to what his product represents. The reality of the situation (unfortunately) is that often the squeaky wheels are the shrillest and if a developer isn't smart enough to take that into consideration, they might get a warped sense of the reception their product is getting. It's always been my aim to represent proportionally what is good and bad in a sim.

 

Getting back around to it (sorry, I'm wandering - casualty of cheap wine I suppose) - Zoltan - I agree...TK isn't greedy. I can't imagine anyone even putting forth that argument. Greedy how?? The series has advanced (both in graphics and features) and represents a solid place between the ridiculous (Hawx) and the obscene (DCS - albeit obscenely fun in my opinion..lol..). SF2 hits the target nearly perfectly in regards to what it shoots for. Personally, I see no need to revamp anything graphically with SF2..it looks great, plays great, and is a pretty solid platform to mod on. I truly hope TK comes back to roost and that when (if) he does we can throw the full weight of our support behind him. This isn't meant as a shot toward anyone other than myself...I simply didn't have the time available to recognize the great fit that the SF series has in our community. I recognize the error of my ways now.

 

PFunk - agreed, the Desert Storm mod is really, really good. The beauty of the situation is that the era between Korea and Iraqi Freedom (is that what it's called?) is probably the most requested period out there. WW2..ugh..so, so not interested (personally). Korea - again..cool..but not my cup-o'-tea.. Vietnam. Bam. I'm there. Falklands? Yeah. Right on. Middle East? Timeless..you can run a plausible conflict there from the 50s to the 2080s (edit: ETERNITY, or at least until the oil runs out) probably.

 

So mobile sims - eh..I've reviewed some, enjoyed some, but they are flash-in-the-pans. I think developers will find a point of diminishing returns on them. When they do, hopefully we will receive the developers back with open arms rather than a cold shoulder. At the risk of using a "corporate" term..it is a symbiotic relationship.

 

So TK - SF3?

 

BeachAV8R

 

(all sics caused by this horrific $9 a liter wine..really? $9???)

Edited April 16, 2015 by beachav8r

[Do335 382]

That said...LOL..the bottle of wine is talking now..forgive me. I can give you a couple insights. Developers DO care what we banter about here. You might think that we represent the vocal minority (and statistically we do), but devs are no different than us. We want feedback, affirmation, and measured (read: polite) critical commentary. Where else are you going to get feedback than online forums? The people that play 10 hours and shelve a title aren't going to be expansive in their lists of pros and cons. But you also have to throw out the outliers - the guys that have flown 20 minutes and the guys that have spent their entire lives flying the same sim - neither represents your target audience which is (usually) the solid middle. Does TK read this stuff? Probably. Some birdie probably taps him on the shoulder and points him in this direction whether it is praise or infuriating criticism. Hopefully, the ratio of good to bad is equivalent to what his product represents. The reality of the situation (unfortunately) is that often the squeaky wheels are the shrillest and if a developer isn't smart enough to take that into consideration, they might get a warped sense of the reception their product is getting. It's always been my aim to represent proportionally what is good and bad in a sim.

Heh yes Beach we're all human with emotions. It's been mentioned that commercial ware devs are "paid to shut up" on the forums. That and, of course, how they're supposed to manage their time, forum blathering or coding?

As a loyal reader of your AARs since like '05, I think they are certainly very spot on to represent the thoughts of mainstream players.

 

And I think (it's just me here, but I sure reckon others'll agree) that SF is good to still have ya around despite TK's departure from PC.

Edited April 16, 2015 by Do335

[Stratos 3,188]

I agree with PFunk too, the ODS mod is just amazing and at the moment is at the top of SF development. But adding a working A-G radar can enhance the experience a lot, specially for planes like the Hornet or the Strike Eagle, does not have to be more complicated than the current A-A radar, but that and CCRP will improve the game a lot and will let the player play until nowadays planes.

 

Can you imagine flying a argie Super Etendard in bad weather over South Atlantic and searchig for a radar eco for your Exocet without knowing If you will hit a british carrier or not...

 

Or searching for SCUDS in the open desert at night using the radar to hit'em hard with a GBU...

 

The same goes for NVG, are those that really hard to add? Not need to be the lighted circle of real life NVG but a simple filter that let you see outside and that can be activated and deactivated trough player in game input.

 

I agree about A-A refuelling as being useless cause maps are not to scale, but not about some guys saying is too hard and does not match SF. Anyone remember the Falcon 4 refuelling options? there is one that let you do all the approach and once you're close to the boom it takes control of the plane until is fully refuelled. Air to Air refuelling are part of every air campaign in the last 50 years, not having them is a lost to the sim.

 

So many options...

Edited April 16, 2015 by Stratos

[+russouk2004 6,958]

Future of SF2..?

if we get the nod from tk to be able to rip it apart   then the future for it is limited only by, 1,  lack of imagination and 2, skilled modders eh...lol

Edited April 16, 2015 by russouk2004

[Stratos 3,188]

Future of SF2..?

if we get the nod from tk to be able to rip it apart   then the future for it is limited only by, 1,  lack of imagination and 2, skilled modders eh...lol

 

 

[+russouk2004 6,958]

Indeed,if thats not permission then I dont know what is....

 

quote  " Please be warned that any mods that change the binary files ( exe,dll,cat etc)  blah blah"

 

hes saying if we mod binary files?...mod them is alter them open them....so they need be original for future packs....a clear message that we can open the files to mod them from his own mouth...

cmon mue...cat tool for locked cats please...

at CA if we decide to break open the game files...and we have tools here to download to do it....we shud make agreement before downloading to use it for sole personal use...and proof of ownerhip of the games be posted via screenshots...im gonna get my bud to crack open the files we need.....TK clearly allows us to do it...so long as we dont swap\share the game core files.

Edited April 16, 2015 by russouk2004

[+streakeagle 871]

The key is, you aren't supposed to distribute his files, presumably even hex-edited versions.

So he is condoning doing anything you want with your installation, but you don't have the right to post/distribute anything/everything you create from his work.

[Wrench 9,846]

isn't that called "Catch 22"???

[FastCargo 412]

Oh come on guys...I posted about this already.

 

If such a mod is created, make a patch that can be distributed. The patch looks for a legitimate, unaltered game file, then alters it. The patch does the rewriting, and the game files (even altered ones) are not distributed.

 

That way the mod can be used, but still requires an unaltered, stock game to exist.

 

FC

[+russouk2004 6,958]

 

So he is condoning doing anything you want with your installation, but you don't have the right to post/distribute anything/everything you create from his work.

What im saying is....If I opened locked cats....modded an a\c that you may not have....and I upload skins etc   then im not doing wrong  as the mod is no good to you if you havent the a\c ive modded...so thats not breaking his guidelines for modding..

uploading skins is in essence distributing TK`s "original files" as he labels them  ...so that mean we must stop making skins?

[+streakeagle 871]

In general, he locked down the files he didn't want distributed, i.e. the lods... but then left the new DLC unlocked which leaves them open to easy distribution/piracy if you are trying to make money selling skins.

TK's original goal of being as open and moddable as possible conflicted with his later goals of preventing piracy and tech support issues caused by widely distributed user mods.

Only TK really knows what he means since he clearly contradicts himself :P

But based on the LOD issues, I am pretty sure he doesn't want the exe, dlls, and/or lods being distributed for free.

But apparently he wouldn't be against distributing patches that alter legitimately owned copies, per his email response approving local edits and Fast Cargo's logic that doesn't involve distributing the edited files.

 

If someone has the time and energy to try to disassemble TK's code to fix minor bugs or maybe even add new features, more power to you. I have some experience doing down that path when I was younger, single, and more of a machine code/C++ programmer. I have no interest in doing that now and seriously doubt anyone following that path will get very far, especially trying to add features.

 

I remember when the Battle of Britain source code was released and chatting with Osram on SimHQ about progress on successfully compiling, running, and then improving the game. I didn't even try to join the team as I was more focused on Jane's USAF as well as the future promise of SFP1 (both having F-4E Phantoms), but watched the BoB modders become the BDG and work hard to produce what would become BoB2, which is still a great sim and game even compared to DCS level realism. My point being: they had the source code and it took them many years to achieve a stable build with any useful changes/fixes/improvements.

 

From what I have read online posted by others that worked on EA Jane's projects, having TK's source code might not help too much since apparently his organization and documentation skills are less than optimum: i.e. while the code works great, it is very difficult to read and follow if you are not TK.

Edited April 17, 2015 by streakeagle

[Stratos 3,188]

Oh come on guys...I posted about this already.

 

If such a mod is created, make a patch that can be distributed. The patch looks for a legitimate, unaltered game file, then alters it. The patch does the rewriting, and the game files (even altered ones) are not distributed.

 

That way the mod can be used, but still requires an unaltered, stock game to exist.

 

FC

 

FC is right, we can do that with TK permission. I agree too in signing an agreement to mod the files. Should we start?

[+gerwin 451]

At the end of my DiD article I tried to send some business TK's way with a bit of encouragement to readers to make it known to ThirdWire that they like their products. *shrug* Can't hurt I suppose..(now watch, TK will shut down the PC purchasing page tomorrow..lol..)

 

http://www.mudspike.com/sf2-war-for-israel-did/

 

BeachAV8R

 

Nice Article! Will read it properly later.

 

 

As to TK's permission about EXEs,DLLs,CATs. Allthough it is funny he replies at all to such a question. AFAIK his reply is basically the law, like it has been for decades.

Posting Hacked No-CD Executable = illegal

Posting No-CD Patch (without part of the Target Software itself) = Not illegal, but often frowned upon.

[ZoltanTheHun 101]

In general, he locked down the files he didn't want distributed, i.e. the lods... but then left the new DLC unlocked which leaves them open to easy distribution/piracy if you are trying to make money selling skins.

TK's original goal of being as open and moddable as possible conflicted with his later goals of preventing piracy and tech support issues caused by widely distributed user mods.

Only TK really knows what he means since he clearly contradicts himself :P

 

 

I think what is stated here is quite obvious and it is a fully legal approach: to mod/hack a game you have all the rights from the very beginning, you can do whatever with the files on your hard drive you want, but you cannot distribute any original files/file parts.

A patch is always a legal way doing this, if it was done through reverse engineering. So redistributing lods or parts of lods is a big no-no, but creating new lods is fine. 

 

I think there was never a legal problem to do any modifications to the game, it quite easy to comply with law in this regard. Reverse engineering afaik was always supported in most western countries. The problem is that it is way easier to create a program which injects the required bytes into an existing dll, then to reverse engineer the dll itself. So the problems are two fold: who would do it, and what is possible within reasonable resources(which is mostly time)?

 

Anyway, there are two welcome things in this: TK responded and he is not denying that future patches are possible.

 

Again, I truly think he must be in love with flight if he worked on this series so far, and everybody needs a break and think about their financials sometime... and honestly, I am sure, if he would be done with Strike Fighters for life, he would release it for the community as open source...(though probably part of the source code is reused on mobile)

[JediMaster 451]

I don't know how much time TK spends on mobile vs what he spent on PC or whatever, but when he said "MP would cost $150,000" or whatever, why didn't he make it a side project? SF2 came out SIX years ago. He couldn't make it work tinkering on it in his spare time over a period of greater than six years (figuring he would've started on it before SF2's first release) or he didn't want to?

 

I think that's the real thing--what he wants to do. He may or may not have reasons he can communicate to us for why he does things, and I think he slaps some arbitrary giant money value on them sometimes just to avoid having to say "I really don't want to do MP because...I just don't want to."

[ZoltanTheHun 101]

May be he didn't want to... The problem is that the multiplayer is the hardest part of them all. I read about implementing multiplayer in strategy games and probably the same implies here.

 

There are 2 basic architectures when it comes to multiplayer:

Client-server systems where clients send their input to the server, and the server sends back the gamestate to the player. So in a shooting game, you send your input to the server, and the server sends back the location of everyone in a given step. Now in a shooting game you have a very small gamestate, think about the position of the players and where they look, you can basically simulate the game on your computer based on this information. There might be small differences, if someone is slower, than it appears lagging, but to you the game is always fluid. If there is a lag, then someone just jumps in position, it is no big deal. 

 

The problem with this approach, that the gamestate grows very fast with each game element, which results in larger and more packets to send. As a result, it takes more time to send the data through the network, and if you want to send it the fastest way, you will send it through UDP, which is a fast protocol, but it is less reliable than TCP. This might result in packet loss, the more data you send the more often it can happen. In shooting games it is no real problem, because the enemy jumps ahead or backward, and maybe you lose a frag, but if it does not happen often, you can live with it.

 

So, we have a simulation here, there are tanks marching on the ground, AI and player planes flying around. Rockets chasing bandits, whom throw flares and chaff all around while engaging in precision bombing. It is quite easy to understand, that we will have a huge gamestate and if we want to have a networking game, than we need to limit our packet size.

 

How can be done? Well, the answer is easy, the implementation is not: let's simulate the games synchronously, and share the inputs the player give. This way our packet size depends on the player only. We can send a data to a server, which is an overhead, or just share between each other directly.

 

This last method is called peer-to-peer lockstep. It sounds cool, even the name of it, but in reality this is a big pain in the ass. The problems are twofold: first, the simulation needs to be deterministic. Not fully deterministic, but the more the better. If 2 games simulate the same input and they are not deterministic, than you will see two different games to play, and the players won't necessary notice it, actually they can both win, and be surprised on the end results.

 

But hey, it is a computer, it cannot generate even random numbers, so sure it is deterministic. Unfortunately they are not. The problem is afaik is with floating point numbers, which we love so much when we implement physics and 3d transformations. All those nice vectors are made of very nice floating point numbers, and they have an ugly tendency of not being accurate. This is caused by the way the computer stores them. They have  a finite mantissa and a limited precision and when you work with them you always introduce bigger and smaller rounding errors, which in sever cases can zero a number out. To make the simulation deterministic, we can use integer math throughout our program and handle this manually, but that would be incredibly slow, so other techniques, very careful programming and lot of testing needed which is truly a huge amount of work, depending the current state of the engine, it could be almost the size of a full rewrite, especially in AI and physics.

 

There is one more drawback to this model: input cannot be lost  to keep perfect sync, so the speed of the simulation is slowed down by the individual computers and the network. For example in an RTS the tick rate(how often physics, gamestate, AI updates per second) can be 20, even if the FPS is updated more often. This ensures that every computer can simulate a current step and receive the next input for the next step.

 

 

These 2 things: simulating deterministically, yet fast enough makes it a very big challenge, especially if you want to support mods too.

.

However if the first thing is done, we have the playback function basically. Which is half of the multiplayer. It is fun to look at long playbacks in Il2... mine is usually completely wrong after a while. Playback in games is done by recording the player input and feeding that to the computer to replay the game. If the engine is not deterministic enough, it will fail after a while.

 

I hope this helps understand it.

 

Here is 2 links on the topic which is more accurate than my explanation:

http://www.gamasutra.com/view/feature/3094/1500_archers_on_a_288_network_.php

http://gafferongames.com/networking-for-game-programmers/what-every-programmer-needs-to-know-about-game-networking/

Edited April 20, 2015 by ZoltanTheHun

[+streakeagle 871]

If you don't already have your own code library to handle a complex programming task, you have to study and/or learn by trial and error to build it or use someone else's. Multiplayer was not TK's specialty, so he went the easy (and at the time, common) route: Microsoft's DirectPlay. Without such a well-documented and easily accessible (free!) library, you can bet SFP1 would have been single player only. However, DirectPlay had its disadvantages. It didn't really handle the latency of the internet so well, data transfer rate wasn't particularly high, not designed to handle network address translation (users hiding behind routers) or DHCP (temporary/random assignment of IP addresses).

 

For reference, Operation Flashpoint started out with DirectPlay, but was quickly retrofitted with an alternate "sockets" method of connection that was both faster and more reliable. TK's solution to the bottleneck caused by DirectPlay and the SFP1 game engine was to limit the objects and view settings in multiplayer. As implemented, SFP1 multiplayer was best in the everyone vs everyone dogfight mode which kept traffic to a minimum and all players happily flying since they could respawn after being shot down. The co-op/team vs team mode exposed the limitations of SFP1 multiplayer. The only ground objects/AAA/SAMs were around the target area with no clouds or aircraft carriers. The massive multiplayer online flight sim, Aces High, has specialized proprietary code to try to minimize warps and lag despite hosting hundreds of players with various internet speed/latency issues. I always felt that both TK and HiTech (founder/lead programmer of Aces High) could have benefited each other by melding TK's content with HiTech's network coding and servers.

 

Microsoft omitted DirectPlay from DirectX10. Multiplayer tech support had always been a thorn in TK's side. It should not have been a surprise that he abandoned multiplayer entirely when SF2 transitioned into DX10. At the same time, TK could see how well multiplayer was working out for other flight sims. The people flying online with even the most popular games like IL-2 and Aces High didn't even begin to total enough income to waste even one more penny trying to develop or purchase new multiplayer code.

Edited April 21, 2015 by streakeagle

[Icarus999 70]

Indeed,if thats not permission then I dont know what is....

 

quote  " Please be warned that any mods that change the binary files ( exe,dll,cat etc)  blah blah"

 

hes saying if we mod binary files?...mod them is alter them open them....so they need be original for future packs....a clear message that we can open the files to mod them from his own mouth...

cmon mue...cat tool for locked cats please...

at CA if we decide to break open the game files...and we have tools here to download to do it....we shud make agreement before downloading to use it for sole personal use...and proof of ownerhip of the games be posted via screenshots...im gonna get my bud to crack open the files we need.....TK clearly allows us to do it...so long as we dont swap\share the game core files.

I fully agree on the  cats .... and  the ability to take a real look at the inner workings of the locked  LOD based terrain system would lead to a new era in terrain development.