A method of saving a mission

[foofee]

Suppose you don't want to complete an entire mission.

Here's a way to save your progress and later reload the mission and continue where you left off.

 

During a mission ESC to quit and select "Save Track".

 

When you want to resume the mission where you left off, run the saved track using the "Replay" selection from the main menu.

After starting the track running press Ctrl-Z several times (maybe 10) to increase the playback speed. Watch the playback.

 

Just before the ending point set the playback speed back to normal

by pressing Shift-Z and ESC to quit.

 

Select the "Take Control" option and resume the mission.

 

The only thing that keeps this from being a true save/load is having to watch the playback up to the save point. If you speed up the time enough that shouldn't take too long.

[Fifi]

Yeah, was very surprised the other day when at the end of track i could take back control!

Nice feature!

[Milli]

Thank you. Never thought of that. Will come in useful.

 

Regards,

Milli

[Jona33]

This won't work in a campaign though as I discovered to my cost.

[Devrim]

But you may not get an exact replay track that you did before.

Because sometimes "replay" doesn't re-read track properly (or track isn't stored properly - I don't know).

 

So, this way is not trustable to me.

[nickexists]

I really wish they would just add a simple save function.

[Kaktus29]

a save function is a must, and its more than doable.. don't know why is this such a big issue to enable.. its even easier than the "save track" thing since the computer would not need to save the whole track but only that one particular moment in time, .. and then you open the save point and continue from there on.. i don't get it, why don't we have this by now? ..

 

being able to save a point would make dynamic campaigns much more possible with better continuity along the war as it happens..

[winz]

don't know why is this such a big issue to enable.. its even easier than the "save track" thing since the computer would not need to save the whole track but only that one particular moment in time,

Saving the one particular moment is where the problems start in sw as complex as DCS. Track needs to save just timestamped keystrokes and replay them from a defined starting position, that's prety straightforward.

But to save a particular moment is big effort for a sw that wasn't build with such functionality in mind. You have to store tons of variables (physics model state, ai routines state, systems state, models state...etc), which you must first identify - that is a lot of work that has to be done for every module.

[foofee]

Why can't you have a save function which saves the track as it does now,

but when you play back the track it zips through all the key presses really

fast and then resumes the normal speed mission at the end?

[cichlidfan]

Why can't you have a save function which saves the track as it does now,

but when you play back the track it zips through all the key presses really

fast and then resumes the normal speed mission at the end?

 

We can't fast forward now without introducing errors, so how do you expect that to work?

[Kaktus29]

@cichlidfan ... okei, let me put it this way.. when in ME we put waypoints, planes, tanks, etc.. we build a mission.. this is in ME..then we press FLY and this materialized in the Sim.. why can't we have this process in reverse? .. like, the moment you stop the mission this waypoints that were already in ME and planes that were already in ME at the beginning of the mission now get UPDATED in the ME when you come back to de-brief.. its basically debriefing in ME .. translating the debrief we see after the mission where you can see who fired which missile and killed which guy etc.. basically translate this data into pics on ME.. so a crashed plane is GONE-not needed on the map, and the one alive is now on the airfield it landed and was recorded where it landed...

 

i still refuse to accept why this is so hard.. a freaking save option for cryin out loud.. i don't care if it wouldn't be 100% correct..like a tank column of 20 tanks was driving on road on a waypoint of 30 miles.. it got hit with ground assault planes and 4 tanks were destroyed.. mission ended at lets say 12:24 ..meaning mission lasted 24 minutes and in this time column moved so much miles ahead to its goal waypoint..

 

we already have the GRID in the F10 map.. we can implement this to "TRANSLATE" the picture to ME.. you have a tank on this coordinates here and here, you have helis on this coordinates.. this are just numbers, no huge task for modern CPU's.. no graphics needed.. nothing.. but AI capable of registering this things before mission finally shuts down and takes you back to ME or whatever..

[cichlidfan]

@cichlidfan ... okei, let me put it this way.. when in ME we put waypoints, planes, tanks, etc.. we build a mission.. this is in ME..then we press FLY and this materialized in the Sim.. why can't we have this process in reverse?

 

My comment was with respect to using fast forward on a track to accomplish the goal.

 

Regarding your proposal...

 

You would also need all information about the aircraft current motion in space, and many other factors beyond the current position of the units.

[Fishbreath]

You would also need all information about the aircraft current motion in space, and many other factors beyond the current position of the units.

 

Not really. Speed, altitude, boresight vector, and velocity vector would be enough, wouldn't it? The flight model can start with that, just like it does in an air start.

[Kaktus29]

why would we need "current motion in space"? .. i think ending the mission would be possible only when planes NOT in combat.. so if they are just flying and not being engaged or engaging something you could press "end mission" and "saving" of ground units to ME and all other flying crafts would automatically be stationed at their home bases..

 

so when you start a new mission it would be from that point BUT with planes stationed at their bases.. i think this would be a low-cost option to make "save" possible..

 

no matter how we look at it, SAVE is more than doable.its a shame it hasn't been implemented even in the most modest ways..

 

especially with CA as it expands this makes things easier.. since you could do stuff if you are not flying and waiting for some 5-10 minutes for an engagement of some planes to finish..

 

missions can be designed in such a way that they "close" by 2 hours..reactionary forces like GAI, or CAP are irrelevant as when there is no opponent they can easily be put back to their home-bases for next mission (plus minus the ammo they used and planes they destroyed as info stays with them-again easy to do since this is what debrief already does now)..

[Kaktus29]

kaktus29 points with hand at fishbreath.. -what he said! ))

[winz]

@cichlidfan ... okei, let me put it this way.. when in ME we put waypoints, planes, tanks, etc.. we build a mission.. this is in ME..then we press FLY and this materialized in the Sim.. why can't we have this process in reverse? .. like, the moment you stop the mission this waypoints that were already in ME and planes that were already in ME at the beginning of the mission now get UPDATED in the ME when you come back to de-brief.. its basically debriefing in ME .. translating the debrief we see after the mission where you can see who fired which missile and killed which guy etc.. basically translate this data into pics on ME.. so a crashed plane is GONE-not needed on the map, and the one alive is now on the airfield it landed and was recorded where it landed...

 

i still refuse to accept why this is so hard.. a freaking save option for cryin out loud.. i don't care if it wouldn't be 100% correct..like a tank column of 20 tanks was driving on road on a waypoint of 30 miles.. it got hit with ground assault planes and 4 tanks were destroyed.. mission ended at lets say 12:24 ..meaning mission lasted 24 minutes and in this time column moved so much miles ahead to its goal waypoint..

 

we already have the GRID in the F10 map.. we can implement this to "TRANSLATE" the picture to ME.. you have a tank on this coordinates here and here, you have helis on this coordinates.. this are just numbers, no huge task for modern CPU's.. no graphics needed.. nothing.. but AI capable of registering this things before mission finally shuts down and takes you back to ME or whatever..

Because a ton off properties are not exposed in the mission editor (or stored in mission files) and are initialized from default values when you hit the fly button. Those properties need to be saved as well, because the default values no longer apply in mid-mission situation. Cockpit switches poisition would be the most basic example.

Another simple example - Fuel for A-10C. You specify your fuel-load in mission editor. When you hit the play button the module code takes this value and distributes the fuel evenly between your two tanks. Now imagine you get damaged tank, and lose all your fuel in one of yours tank. What happens if you would like to save now? The fuel load as whole is no longer sufficient, because using it would (once again) distribute fuel evenly once you load your savegame. Resulting in half of your remaining fuel beeing put into the ruptured tank..not good. For savegame functionality you need to store fuel in each tank.

There are tons of such variables, because DCS modules are the most complex flight simulator modules available for civilian market. What is 'worse' is that DCS is modular, so the modules themselves would have to tell DCS World what they need to save.

 

Yes, it woud be a great functionality, I agree. But it is a lot of work and ED doens't have a big team and there things higher on the TO-DO list than this.

Edited November 13, 2013 by winz

[Kaktus29]

@winz.. as i said, what we need is a functioning AI of sorts.. an AI to monitor the units and "write them down" as you hit "SAVE" button.. i mean I CAN DO THIS right now.. PAUSE the game, write the coordinates of units, groups, their waypoints, altitude, speed, banking, yaw, pitch.. it will take me some 45 minutes to put down some 30 - 40 units but i can do it.. then in ME i can put those numbers i got from F10 map to a NEW mission and voila.. we have "SAVED" a mission..

 

i can't see why this is so hard.. i mean a computer can "write" this numbers down much faster than i can.. it shouldn't take more than 1 second.. the problem as i see it is not-functioning AI logic.. we would need ME-F10 AI logic tha would understand what "SAVE" the picture means.. and translate it to "new mission" that is updated with the situation..

 

about mid-air A-10C etc..and all the new things that are changed and complications as you said it, why? why is it so hard.. as you said, in the beggining of the mission there is a DEFAULT setting for the plane.. so plane that is NOT shot up, that both engines work, fuel tanks not ruptured etc.. all that "save" button would do is do a reduction equation from the default setting.. like default was this and this - (as in MINUS) new realities.. (lower fuel, less ordnance, damage to left wing ) ..

 

i know that if one really wants SAVE mission to work when major AIR WAR is happening is very hard, and i'm not saying this to be done since it could prove lot of work-not impossible-but lots of work..but even so, people are forgeting that the biggest problem to creating Dynamic campaign IS the SAVE option... in that view i think its the most important thing in DCS ever.. more than new modules, new engine, new whatever..

 

A Save option gives you ability to actually START a war with so and so units and prosecute a WAR.. over limitless missions or until war is won/lost (considering the damage you have done or territory you have won/lost)..

 

after SAVE is enabled we can start working on AI logic in understanding Air POWER/Air dominant vs. contentious air space/SEAD/Tactical ground attack with Air cover-Escort/CASARTY/Armor/Infantry/- ..

 

Right now its all dry and empty cuz its basically one mission at a time-without any real connection between one and the other especially since there is no SAVE enabled that would give you a sense that when you destroyed a bridge to prevent armor to cross it you see this result in next mission that forces enemy armor to go around to different bridge or wait for engineering brigade ...

 

SAVE option is more important that anything right now..

[upupandaway]

I don´t really see the problem in saving a ton of variables, transforming the mission state into a save file that the ME could load again later. There are enough other game that create/handle huge savegame files (TES for example). If some of the variables aren´t stored right now, would it be so complicated to add them to mission files?

Apart from that I don´t really think it´s THE most important thing in DCS right now, certainly not more important than a new engine or modules as tose will add different things which are at least equally important to me personally.

[Vedexent]

Ah, playback errors. I remember the first time I rewatched a mission and went "wait a minute! I didn't blow up there!" :P

 

-----

 

It seems that the TacView program "hook" exports and saves game data pretty much like people seem to expect a "saved game" function to work.

 

I don't know if it stores enough variables, or whether all the required variables are exposed to grab.

 

I also don't know if it would be possible to re-create units with enough fidelity by using scripting functions and libraries like Mist.

 

Still, it looks like a third party plugin is implementing a large part of what people seem to want in a save game feature: if the system design won't do something you need, hack around it :P

[winz]

@winz.. as i said, what we need is a functioning AI of sorts.. an AI to monitor the units and "write them down" as you hit "SAVE" button.. i mean I CAN DO THIS right now.. PAUSE the game, write the coordinates of units, groups, their waypoints, altitude, speed, banking, yaw, pitch.. it will take me some 45 minutes to put down some 30 - 40 units but i can do it.. then in ME i can put those numbers i got from F10 map to a NEW mission and voila.. we have "SAVED" a mission..

Because it would not work. Those data are not enough to recreate the mission. Imagine you saved with your method with an AI in the middle of landing approach. After load it would in better case abort the landing, in worse case crash into the ground.

You are also ommiting one huge aspect of mission - triggers and lua scripts. In this method you don't perseve their states. Not to mention that all units will be recreated new, without any damage they received.

 

about mid-air A-10C etc..and all the new things that are changed and complications as you said it, why? why is it so hard.. as you said, in the beggining of the mission there is a DEFAULT setting for the plane.. so plane that is NOT shot up, that both engines work, fuel tanks not ruptured etc..

I don't need to damage the plane to acheive this. I can fly on one engine for 20minutes, or I can leave the APU running for the same thing to happen. And that's for A-10C. In P-51D, where you manually select tank used, there is no way to recreate the fuel values from a single number. You need to store fuel load for each tank separately.

 

all that "save" button would do is do a reduction equation from the default setting.. like default was this and this - (as in MINUS) new realities.. (lower fuel, less ordnance, damage to left wing ) ..

Yes, in order to climb on mount everest, you only have to climb the damn mountain.... The problem is that the program doesn't know which variables needs to be stored to recreate the object state. A bunch of programers need to spent time indetify those variables. And there are tons of them, the fuel is just the most basic, obvious expample. If you look how many parameters you change in CDU and other systems in the A-10C you realize how many variables there are. Then there is damage persistance, wear and tear on equipment (i.e. engines)...etc. It's not hard, in sense that it's a complicated process requiring some insane research paper, it's just very time consuming.

 

SAVE option is more important that anything right now..

That's your opinion. My opinion is that EDGE + Dedicated Server is more important than save functionality.

 

It seems that the TacView program "hook" exports and saves game data pretty much like people seem to expect a "saved game" function to work.

TacView only have to save basic object properties, like location, velocity and major game events. I.e. it doesn't know how much fuel was left in tanks of an airplane, or what kind of damage was done to an unit. It doesn't need to, it's not required for what its aim is.

Edited November 13, 2013 by winz

[Mosley]

How many times have you replayed back a track to end with gun runs into the ground with no target it sight or fly off into the distance for no reason.

 

No offense to ED but the replay system is not up to par. If this system was modified to save all the different variables as some have mentioned you would more than likely get a different variant of your save than what is actually going on in the sim.

 

ED would also have to implement some sort of way to re insert all the destroyed objects. Think bridges, buildings, aircraft, vehicles, etc. For me these become landmarks or references to the current fight. Should they be simply left out? In theory this type of stuff should be easy to put back in sim. At cord x,y place destroyed t-80 tank, but in the end the sim was not designed with the "Save" feature in mind and would require a lot of work to get this working right. Not to mention the time added to get the replay system working properly. Also as other's have mentioned the LUA scripting would have to save all those variables plus older ones if they effect the newer variables.

 

Personally I would rather them fix the bugs that we currently have such as CBU's causing slideshows or the 97's ganging on up 1-2 targets among other things. Continue working on EDGE and eventually pick back up development on the dedicated server would be time better spent imo.

 

Also keep in mind, quite a few 3rd party projects are being released and ED has to help these guys as well. Their plate is simply so full this "Save" feature would be at the bottom of the list.

 

I guess we can hope they wanted to add something like this and are working to add it to a EDGE version because if not I doubt it will ever happen.

 

Its a small team and they are doing what they can but it all takes time.

[ShuRugal]

Because a ton off properties are not exposed in the mission editor (or stored in mission files) and are initialized from default values when you hit the fly button. Those properties need to be saved as well, because the default values no longer apply in mid-mission situation. Cockpit switches poisition would be the most basic example.

Another simple example - Fuel for A-10C. You specify your fuel-load in mission editor. When you hit the play button the module code takes this value and distributes the fuel evenly between your two tanks. Now imagine you get damaged tank, and lose all your fuel in one of yours tank. What happens if you would like to save now? The fuel load as whole is no longer sufficient, because using it would (once again) distribute fuel evenly once you load your savegame. Resulting in half of your remaining fuel beeing put into the ruptured tank..not good. For savegame functionality you need to store fuel in each tank.

There are tons of such variables, because DCS modules are the most complex flight simulator modules available for civilian market. What is 'worse' is that DCS is modular, so the modules themselves would have to tell DCS World what they need to save.

 

Yes, it woud be a great functionality, I agree. But it is a lot of work and ED doens't have a big team and there things higher on the TO-DO list than this.

 

aaand the simulator already handles all of these variables, and processes changes to them based on very complex interactions, all in real-time.

 

A very crude way of saving the game state would be to simply take a snapshot of the current state of the simulator (basically, the entire memory chunk allocated to the sim) and save it to the disk for later use. This would result in absolutely massive save game files, you would not want to have more than a handful at any one time, but it would be entirely feasible.

 

 

Yeah, there's a crapton of information being processed at any one moment in the sim. But guess what, the sim is already handling all that information in real time. All you have to do for a save file is to make a copy of that information and save it somewhere. The hard part isn't in that, it's in trimming out the un-necessary info and setting routines for the simulator to fill it back in on file load.

[winz]

No, you cannot just dump an app memory and restore it at a later date, that doesn't work. I.e. the the dump wouldn't contain kernel memory = handles to kernel resources, like opened files would be invalid after the restoration.

 

Yes it is very clear what you have to do to save a game, it's the implementation part that is time consuming.

 

And to add a personal note to the 'quick & dirty' solutions people try to figure out here. I can tell you, from my experience as SW developer, that quick & dirty solutions will in the end cost you more development time than proper solutions. Because while you hack it quickly together, and it kinda works in your happy end scenario you visioned, it doesn't work in any other, so you spend a ton of time plumbing all those holes the quick and dirty solution has...more ofthe than not with another bunch of quick & dirty solutions.

Edited November 15, 2013 by winz

[Fifi]

A very strange thing happened the other day, when i watched back my P51 track:

 

In my mission, i landed back home nicely, taxiing to park, and cuting off my engine.

Watching the record track, i saw my plane not landing on airfield tarmac as i did, but completely off in the grass...and crashed into a bunker!! :surprise:

 

How come??

[Merlin-27]

A very strange thing happened the other day, when i watched back my P51 track:

 

In my mission, i landed back home nicely, taxiing to park, and cuting off my engine.

Watching the record track, i saw my plane not landing on airfield tarmac as i did, but completely off in the grass...and crashed into a bunker!! :surprise:

 

How come??

 

Did you use fast forward while watching?

It's a known issue that a track may become "corrupted" if you speed up or slow down the playback. Hopefully we will get a more refined replay solution down the road.