PDA

View Full Version : Begging 1C to include tracks from final testing with patches.



WWMaxGunz
10-05-2005, 02:11 PM
There is so much discussion on player tests of the planes.
And not always does the flying meet with Oleg or 1C standards.

In all fairness to 1C and also to players I would like to see;
tracks of proper flights showing planes making best speed, roll, climb, dive and turn.

I am assuming that these must exist as part of beta testing all through the process.
So is it so much extra work to zip up the final tests and make them available?

Especially considering this gives two benefits. First it shows the good way to fly
the different planes at least those things in the tracks. Second it shows capabilities
when done right and no need for so many posts and questions unanswered, fights over if
the player did this or that right and general not good feelings that end some at 1C.

When players have something to measure their own abilities to, they may improve.
So please as before do the tracks but please on performance as well as battle.
I and many others really did appreciate the tracks that came with ... was it AEP?

Neal

Mysha76
10-05-2005, 02:25 PM
good idea

crazyivan1970
10-05-2005, 03:49 PM
Maybe we should take alternative approach... there are many people in this community that are dedicated to "a plane", meaning know its best abilities, etc. Maybe even better then developers http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif.

TAGERT.
10-05-2005, 04:50 PM
Originally posted by crazyivan1970:
Maybe we should take alternative approach... there are many people in this community that are dedicated to "a plane", meaning know its best abilities, etc. Maybe even better then developers http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif. Got Track? http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif

BBB_Hyperion
10-05-2005, 06:12 PM
As i dislike the human uncertain factor i will describe my old idea to have ai routines rdy for player controled planes like hold alt climb at degrees ,keep speed at , roll degrees , keep torque neutral etc or even automate main testing procedures for the sim in ai scenarios .

TAGERT.
10-05-2005, 06:20 PM
Originally posted by BBB_Hyperion:
As i dislike the human uncertain factor i will describe my old idea to have ai routines rdy for player controled planes like hold alt climb at degrees ,keep speed at , roll degrees , keep torque neutral etc or even automate main testing procedures for the sim in ai scenarios . You could do something like that now with the mission builder.. but, prob is you can not read out the AI plane data via DeviceLink

WWMaxGunz
10-05-2005, 06:22 PM
Hyperion, do you remember the guy that was working on a straight and level program using
devicelink for purposes of level bombing through the sight? He was unable for some
reason to make it work as good as he wanted at the time, I don't know why. But offline
we have tools for maybe doing those things.

---------------------------------------------------

Ivan... many times tracks sent to 1C get comments of not being flown right.
I just figured that how do they say FM is ready without flying it?

But if we can have members who are recognized as being so good, always get the charger
and prop and method right then hey I will be happier as they are likely to answer the
questions and help others out.

http://forums.ubi.com/images/smilies/11.gif

TAGERT.
10-05-2005, 06:28 PM
Originally posted by WWMaxGunz:
Ivan... many times tracks sent to 1C get comments of not being flown right.
I just figured that how do they say FM is ready without flying it? My guess is they dont. By they, I mean Oleg and crew. They dont need to fly it to see how the math works, they can just run the equations and see what they spit out for a given input. Then have a util that checks the results aginst the values they expected. Like Rollrate, Climbrate, et.

The beta testers are good for general purpose testing, user interface, map building, or general feedback on how it feels. The FM is such a small part of a combat flight simulator. Thing like AI and DM and Graphics issues is where beta testers can give useful feedback.. In that all that stuff is just a wag anyways! http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif

But, if a beta tester finds something in the FM that is out of wac, it most likly is due to that is what Oleg expected, thus it was not caught by the crew's util that checks for it, because it is what they expected it to be. Which is why we have seen changes in the FM over time.. People see something that is not right, find the real world data to suport it, send it to Oleg, he changes is util that checks or that and makes the change in his FM code.. Assuming that Oleg is convinced the real world data is valid. This simple concept is something I have been trying to explain to Stigler for years now.. But he wont hear it. More fun to blaim the FM and Oleg, In that he thinks it makes Targetware sim look better because they have had less FM changes.

crazyivan1970
10-05-2005, 06:48 PM
Originally posted by TAGERT.:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by crazyivan1970:
Maybe we should take alternative approach... there are many people in this community that are dedicated to "a plane", meaning know its best abilities, etc. Maybe even better then developers http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif. Got Track? http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif </div></BLOCKQUOTE>

No, but got something heavier then that http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif

TAGERT.
10-05-2005, 06:58 PM
Originally posted by crazyivan1970:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by TAGERT.:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by crazyivan1970:
Maybe we should take alternative approach... there are many people in this community that are dedicated to "a plane", meaning know its best abilities, etc. Maybe even better then developers http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif. Got Track? http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif </div></BLOCKQUOTE>

No, but got something heavier then that http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif </div></BLOCKQUOTE>Got Ego? http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif

WWMaxGunz
10-06-2005, 05:24 AM
Originally posted by TAGERT.:
My guess is they dont. By they, I mean Oleg and crew. They dont need to fly it to see how the math works, they can just run the equations and see what they spit out for a given input. Then have a util that checks the results aginst the values they expected. Like Rollrate, Climbrate, et.

I sure hope not! That's about first stage only. Not checking how the code actually runs
within the dynamic of the machine and all else operating is luney! RT code is never just
look at the math no matter how good your results have been so far. You tweak, you check.
Anything less is sloppy practice sure to error once things get the least complex, which
IL2 is a world past from the start.

And for beta testers... they are second stage testers. First are the alpha testers.

BBB_Hyperion
10-06-2005, 07:12 AM
Originally posted by WWMaxGunz:
Hyperion, do you remember the guy that was working on a straight and level program using
devicelink for purposes of level bombing through the sight? He was unable for some
reason to make it work as good as he wanted at the time, I don't know why. But offline
we have tools for maybe doing those things.

---------------------------------------------------


I checked the possibilites a while ago and it is indeed possible but as you know devicelink process 1 set command in time so you have n commands for each flight correction after pulling data out. That means you call 1 data package and you send n corrections for flight attitude and settings. This means after every set command the attitude changes and result must be corrected this leaves a working but very inaccurate testing method. Only method i found possible with multi input that reacts fully on all controls is to implement a virtual joystick driver which has all axes and sliders for all things imagineable so all inputs can be send at once . But didnt try it out yet no time .(


Originally posted by TAGERT.
You could do something like that now with the mission builder.. but, prob is you can not read out the AI plane data via DeviceLink


Current AI is not able to use CEM or trim etc right that is why i would deny the option of tests just with ai climbs etc results would be off.
Even when you set player controled craft and set auto pilot on you could archive a ai controled flight with devicelink data but this data is not compareable to normal human player .

In fact i send in ai tracks showing fuel usage of some planes using this ai technique which turned out that fuel usage or fuel capacity was just wrong on some .

quiet_man
10-06-2005, 08:22 AM
would at least add some things to discuss at this forum http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif

quiet_man

TAGERT.
10-06-2005, 08:40 AM
Originally posted by WWMaxGunz:
I sure hope not! That's about first stage only. Not checking how the code actually runs within the dynamic of the machine and all else operating is luney! I think you missunderstood what I said, I didnt mean they just look at the text of the math on a screen or printout, they actully run the math with a rage of inputs and graph it on the screen. That is how we bench test portions of code in aerospace, checking for end point wierdness, roll over, etc. We even test the code for values it would never even see to see if it locks up or goes off into la la land. Thus, they dont have to sit down with a joystick and fly all 100+ planes each time they make a change, they just run these types of tests with uppper and lower limts for each plane.


Originally posted by WWMaxGunz:
RT code is never just look at the math no matter how good your results have been so far. You tweak, you check. Anything less is sloppy practice sure to error once things get the least complex, which IL2 is a world past from the start. Disagree 100%, by the way, what does "RT" stands for, real time?


Originally posted by WWMaxGunz:
And for beta testers... they are second stage testers. First are the alpha testers. Beta testers are mainly good for doing one thing.. Testing your code to see how well it holds up for stupid inputs! http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif 90% of Windows code is to account for stupid people.. or not so stupid people that do stupid things!

For example, guys who fly thier planes upside down 10ft off the ground and acedently hit the ground with their prop BUT THEY DIDNT BLOW UP.. But if your fling right side up you do die when your prop hits the ground. WHO KNEW! WHO THOUGHT TO CHECK FOR THAT ONE! Nobody, thus beta testers are great for stuff like that! http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif

Stigler_9_JG52
10-06-2005, 10:43 AM
You forgot a few factors in your story, Tagert.

What about the part where Oleg dismisses your real world data as "US [or other country not Russian) propaganda", won't name what his numbers or his source is, and just digs in his heels and refuses to change something. This goes not only for FM, but for things like view out of cockpits *cough*FW190*cough. Or, if he just relents to endless complaints here and changes "something" but doesn't tell you what or how much...you just notice that, for example, the .50s are "different" effectiveness-wise, but it's nothing measurable. How do you "get track" on that, hmm?

Thing is, with Targetware, you can easily see, get, and test the data. The files are open. There's a handy dandy debug screen you can open and see, on the fly, exactly what your speed, alt, G-force, temperature, whatever, is so you can more confidently say "this is off" than you can looking at a track (especially one that doesn't always reproduce what happened when it was recorded).

Then, getting a change is pretty simple, too. You report your findings, and either get someone who knows how to change it for you...or you can change the FM file yourself, and submit it to whoever's in charge of the server. If it passes muster, he'll replace the old file for yours and problem solved.

And even if the server host thinks you're full of it, you have the option of running your own server, with your version of the file, and others can come in and check it out and compare it. Perhaps then enough people will build a groundswell of support, based on real, observable numbers, and with all the data from both sides of the argument on the table, and perhaps then the main server host can be rightfully cajoled to change the "official" FM (or alternatively, to successfully defend the official numbers). Not based on simple whining or a popularity contest, or anyone's sense of entitlement to win in Plane X, but based on numbers everyone can see and generate, and data freely available to everyone.

Oh, and these processes are usually started by someone simply noticing something that seems obvious, like a plane not being able to hit its sea level speed, or having an overly generous climb rate or turn rate... and then, based on that, they begin to take a harder look at it. They don't just **** it off based on whether anyone has recorded a track of it.

TAGERT.
10-06-2005, 10:13 PM
Originally posted by Stigler_9_JG52:
You forgot a few factors in your story, Tagert. Which is not to be confused with missing facts.


Originally posted by Stigler_9_JG52:
What about the part where Oleg dismisses your real world data as "US [or other country not Russian) propaganda", What about it? Oh, wait, right, forgot, you know better.. right? We should take your word over Oleg's? Just what is it you do for a living and what is it that Oleg does and did for a living? Given a choice, I would take Oleg's word over yours any day of the week.


Originally posted by Stigler_9_JG52:
won't name what his numbers or his source is, and just digs in his heels and refuses to change something. I hear alot of whiners use that line.. but I have yet to see Oleg not answer important questions. Problem is some dweebs think he has to answer every question put to him, and if he does not, Well they get thier feelings hurt and then they come here to this forum and say things like you just did. Just to get back at him for not answering thier question. So, just which one of your questions did not get answered? Maybe I can answer it for you and you wont feel so bad and then cut Oleg some slack?


Originally posted by Stigler_9_JG52:
This goes not only for FM, but for things like view out of cockpits *cough*FW190*cough Ah, thanks for proving my point! Perfect example, Oleg, on several ocations explained his reasoing for not changing the Fw190 view. Just because you didnt understand it or agree with it does not give you the right to lie and say that he did not provide an explaintion.


Originally posted by Stigler_9_JG52:
Or, if he just relents to endless complaints here and changes "something" but doesn't tell you what or how much... you just notice that, for example, the .50s are "different" effectiveness-wise, but it's nothing measurable. How do you "get track" on that, hmm? Last I check, Oleg is not obligated to tell us what he had for breakfast, let alone what he changed in his code today.


Originally posted by Stigler_9_JG52:
Thing is, with Targetware, you can easily see, get, and test the data. The files are open. There's a handy dandy debug screen you can open and see, on the fly, exactly what your speed, alt, G-force, temperature, whatever, is so you can more confidently say "this is off" than you can looking at a track (especially one that doesn't always reproduce what happened when it was recorded). You clearly dont understand track files yet.


Originally posted by Stigler_9_JG52:
Then, getting a change is pretty simple, too. You report your findings, and either get someone who knows how to change it for you...or you can change the FM file yourself, and submit it to whoever's in charge of the server. If it passes muster, he'll replace the old file for yours and problem solved. Yet, you still come here and fly IL2 most of the time.. why is that? What with them catering to your every whim, I would think you would be over there sucking them dry day in and day out.. Yet your here? Let me guess, this week, or this day, they decided to change the FM based on someone elses whim, which is in conflit with your whim, thus, until the swicht it back to your whim, you feel the need to hang out here?


Originally posted by Stigler_9_JG52:
And even if the server host thinks you're full of it, you have the option of running your own server, with your version of the file, and others can come in and check it out and compare it. Perhaps then enough people will build a groundswell of support, based on real, observable numbers, and with all the data from both sides of the argument on the table, and perhaps then the main server host can be rightfully cajoled to change the "official" FM (or alternatively, to successfully defend the official numbers). Not based on simple whining or a popularity contest, or anyone's sense of entitlement to win in Plane X, but based on numbers everyone can see and generate, and data freely available to everyone. I dont think you can use the word groundswell to describe any aspect of Targetware.. Unless your definition of groundswell means six people flying at any one time. But, it is all relitive, I guess 6 people relitive to the typical 4 people playing would be a groundswell.. relitivly speaking.


Originally posted by Stigler_9_JG52:
Oh, and these processes are usually started by someone simply noticing something that seems obvious, like a plane not being able to hit its sea level speed, or having an overly generous climb rate or turn rate... and then, based on that, they begin to take a harder look at it. They don't just **** it off based on whether anyone has recorded a track of it. Yet.. your here and not there? Must be something lacking in Targetware.

Stigler_9_JG52
10-07-2005, 10:36 AM
I visit both places all the time, so there's no preference stated.

And whether there are many people or few at Targetware, that's irrelevant to a discussion on which system allows for independent testing and correction and which doesn't.

Now, true, Oleg is not obligated to tell us anything, but you have to admit if all the data's on the table, easy to glean from the sim itself and clearly spelled out, and the sources made clear for comparison, there's a lot less room to split hairs on the findings. After all, as you say, if you're really going to come up with a definitive answer, your numbers have to be TIGHT. You can't get them as tight as you yourself challenge people to get them with just a track.

See, here's where you clearly distinguish yourself as a sycophantic fanb*i: you assume (in true Benny Hill style) that Oleg's an infallible god and we're all stupid idiots who can't think critically for ourselves and observe things and bring up apt criticisms. Then, he gets a pass on having to "put up or shut up" as far as showing his cards, but we all need a master's thesis and dissertation to prove (to you) that, say, energy retention is off... and all the time we can see aircraft doing loops right off the tarmac with barely wheels up (such as certain 109s and Hurricane IIs did a few versions ago). And obvious observations such as "well, since the FMs of planes swing wildly between versions, they can't ALL have been right, even when you factor in changes to the game system" just roll off your back... simply because you know you can't disprove that with "got a track?"

WWMaxGunz
10-07-2005, 11:30 AM
Stig, 1C is making a product not a reality. Same with Targetware except that the
latter is at some point going to the pay for play model, ie open ended budget.

Neither is perfect. Both have things the other does not but I won't say the balance.

So maybe don't knock 1C for avoiding things that they decide are poor, bad or even
impossible measures of return for effort when you don't know how their code is.
It is bad form and worse the more you know or understand about industry workings.

I'm sure I could find a lot of shortcomings with Targetware. If I was paying to play
or wanted any future with it which does include that, I be finding a lot more as time
and my investment in the sim went up.

WWMaxGunz
10-07-2005, 11:47 AM
Yeah Tagert, RT means Real Time. In the case of our sims it does include using the
joystick as well as all the hardware and variations and the elements of timing in
the hardware, a very tricky thing as things get more and more freely run. Ask any
EE... oh wait, ask yourself about hardware and timing issues.
There's that guy who started out the whole "design by evolution" thing using, I think
it's PLA's that found that simulations of the same just don't cut it. You know, I am
sure. Fascinating but also limited in what kinds of results can be realistically
aimed for. Neatest part to me is that even with a working circuit they are usually
unable to say exactly why it works! It's just smaller and sometimes faster, whoohoo!

beta testers are good for testing the unexpected which includes stupid but also not
stupid events and series of events. They are also good for seeing what happens on
a range of hardware possibly as wide as the group, generally to some fraction. I've
been a beta before and there was more to it than screwing up and reporting on it.

Also contrary to popular opinion, beta testers are not responsible for final states
of games. Sometimes I've wondered if the betas did or did not report exploits or
other undesireable traits but in one case I know about (my NDA was over in 2003) the
final release was very different than any beta test and much inferior to the last.

Stigler_9_JG52
10-10-2005, 02:15 PM
Originally posted by WWMaxGunz:
Stig, 1C is making a product not a reality. Same with Targetware except that the
latter is at some point going to the pay for play model, ie open ended budget.

Neither is perfect. Both have things the other does not but I won't say the balance.

So maybe don't knock 1C for avoiding things that they decide are poor, bad or even
impossible measures of return for effort when you don't know how their code is.
It is bad form and worse the more you know or understand about industry workings.

I'm sure I could find a lot of shortcomings with Targetware. If I was paying to play
or wanted any future with it which does include that, I be finding a lot more as time
and my investment in the sim went up.

You miss my point, Gunz. Tagart is asserting that you have to have a track to even think anything is wrong with the modeling, which is patently false. Many of us feel that some parts of it are so obviously OFF as to not need a .ntrk to notice they're wrong; it is true that the track might help you quantify "by how much" it's off, but it's certainly no infallible scientific tool itself.

And I'm not going to sit here and say I can determine what Oleg should put resources into modeling correctly; but that doesn't mean he can't come out and say what the limitations are, rather than tell people who criticize that "they're wrong, be sure". Or make changes that are unquantifiable, and leave us to determine what has changed, and if it's more accurate than before. Or making changes based on who's whining loudest, rather than data (specifically, data he can name, and stand behind). Or ignore questions/observations altogether, which seems to be the most recent modus operandi.

And, as for TW, yes, you probably could find things wrong with some of its FMs. It happens all the time there. The difference is that the discussion (not to mention the DATA itself) is much more honest and open, and because the data cards are out there on the table, there's a lot more potential to change it for the better, free from any particular bias or assumption.

WWMaxGunz
10-10-2005, 05:25 PM
I really think that Tagerts' point is not about what you are allowed to think.
That's a false counter if this is right:

When you make a claim, a track is the only current way to back it up.