1911 replies to this topic

[Karl Berg]

Heffay, on 02 April 2014 - 12:37 PM, said:

You make it sound like "Sure, no problem, we'll slap in the integration code in a couple of hours, do some QA and *poof* TrackIR support. New bug: Hitting the control key to enable free-look in 1PV causes the mech to Stackpole.

That's exactly why I'm going to hand off to gameplay when I'm done. Then I can blame them for the bugs..

[Krinkov]

I noticed in your presentation overwriting of manual edits was a problem. Is that problem at all related to the reappearance and swift patching of some minor bugs that pop up? For example, the scrambled hud after overheat bug popped up again after a patch in February but was gone the next patch. Does your system only generate code for the back end or do we see some of it's results in game?

[Monkeystador]

Hello, i have been kindly pointed out this post and it's reference to TrackIr. I am interested in the implementation of this. Mainly though because i like to have this pave the way for any headtracker. My main interest is support for the coming Oculus Rift.

I wrote another post outlining the most simple way to get it integrated:

* x,y orientation for player head in the controller assigment list, make this absolut. controller x/y some value like 0 is straight ahead.
* make head control active all the time
* the hard: decouple the crosshair from the player head view vector. Unlike the freelook currently implemented.

could you comment on that?

[Karl Berg]

Krinkov, on 03 April 2014 - 02:03 AM, said:

I noticed in your presentation overwriting of manual edits was a problem. Is that problem at all related to the reappearance and swift patching of some minor bugs that pop up? For example, the scrambled hud after overheat bug popped up again after a patch in February but was gone the next patch. Does your system only generate code for the back end or do we see some of it's results in game?

Ahh back on topic!

Short answer is no.

Long answer is the segment on clobbering manual edits is about best practices with automatic programming, and the lessons learned from Microsofts visual code editor thingy. The problem is this; if a programmer comes along and edits a few lines or functions in a 100,000 line autogenerated C++ file, you've just increased the size of your maintainable code base by those 100,000 lines, which is devastating to productivity. Our solution is a tough-love approach. If a programmer hand-edits an autogenerated file, the next time he builds those changes will be lost.

This applies almost exclusively to server code, although with UI 2, the UI guys have been taking tentative steps towards leveraging more of our code autogeneration platform, for both C++ and actionscript, to improve a few sore points regarding communication between the two languages. Gameplay has also expressed some interest in our systems recently, which is pretty cool. It's really tough to communicate just how much of a productivity boost our code autogeneration systems provides us. That 10 to 1 code compression factor only covers some of the benefits. So no, the clobbering of code by the autogeneration system has nothing to do with any UI regression bugs.

[Karl Berg]

Monkeystador, on 03 April 2014 - 08:56 AM, said:

Hello, i have been kindly pointed out this post and it's reference to TrackIr. I am interested in the implementation of this. Mainly though because i like to have this pave the way for any headtracker. My main interest is support for the coming Oculus Rift. I wrote another post outlining the most simple way to get it integrated: * x,y orientation for player head in the controller assigment list, make this absolut. controller x/y some value like 0 is straight ahead. * make head control active all the time * the hard: decouple the crosshair from the player head view vector. Unlike the freelook currently implemented. could you comment on that?

The branch I'm working in is titled 'Rift'. Several of us also own G1 rift dev kits. Take from that what you will..

More usefully, I've gone through Rift integration start to finish with another source base already, and generally found it to be a reasonably straight-forward process.

There are a few challenges for MWO however that prevent easy integration right now. Without extensive UI support and/or refactors, we lose the front-end and all HUD. Our render performance is also .. not optimal right now, so maintaining 60 hz with double the draw-calls would be a significant challenge. I'm also massively susceptible to motion sickness, so my ratio of working to lying flat on the couch with a terrible headache and nausea is about 3 to 4 minutes of work to 4 hours of recovery.. Getting correct IPD behaviour for eye convergence was a real experience. To the point where I was probably suffering from 'post-traumatic-rift-syndrome' for a time.. Regardless, it's ridiculously cool tech, and MWO is almost a perfect showcase product for this tech.

[Karl Berg]

Quick update on TrackIR: great success has been achieved, NaturalPoint has been extremely supportive and helpful, so many thanks to Seth and the others over at NaturalPoint! The branch is now in the hands of the Gameplay team.

[Loc Nar]

Karl Berg, on 03 April 2014 - 12:59 PM, said:

Quick update on TrackIR: great success has been achieved, NaturalPoint has been extremely supportive and helpful, so many thanks to Seth and the others over at NaturalPoint! The branch is now in the hands of the Gameplay team.

Simply wonderful news, thank you so much.

[Krinkov]

Since you already render the mechbay at start up, have you considered letting us walk around it and interact with scaleform based in engine menus? It would make for amazing rift immersion to be able to walk around and see our mechs and there would be no more in your face disorienting views of the main screen in rift. Honestly, even without rift I would love to be able to walk around the mechbay. I firmly believe MWO is one of the best option for conversion to VR due to the in cockpit view eliminating motion sickness for most people.

[Karl Berg]

Krinkov, on 03 April 2014 - 03:07 PM, said:

I firmly believe MWO is one of the best option for conversion to VR due to the in cockpit view eliminating motion sickness for most people.

Yes, this exactly, I agree.

[Alaskan Nobody]

Krinkov, on 03 April 2014 - 03:07 PM, said:

.....eliminating motion sickness for most people.

Wait?!

So other people have that issue with other 1st person viewpoint games to?

I could never find much about that!

[p4r4g0n]

Rebas Kradd, on 02 April 2014 - 12:38 PM, said:

Karl, this is the kind of insight that would really help flesh out the company to your customers. Not that shutting up the belligerents is or needs to be a priority, but still, I dig this stuff.

You should have your own weekly "Karl's Korner" or something equally cheesy, where you just casually answer the question "What did you work on this week?" Far more interesting than a monthly Vlog halfway devoted to stuff that was already answered in greater detail in a Command Chair post.

QFT - most informative discussion I ever had on this forum was in a thread about Elo (pre-release) & matchmaking was when Karl popped in and offered some insights about some of the under the hood mechanics.

[Rebas Kradd]

p4r4g0n, on 03 April 2014 - 04:02 PM, said:

QFT - most informative discussion I ever had on this forum was in a thread about Elo (pre-release) & matchmaking was when Karl popped in and offered some insights about some of the under the hood mechanics.

I'd love a link for that!

[Krinkov]

Shar Wolf, on 03 April 2014 - 03:43 PM, said:

Wait?!

So other people have that issue with other 1st person viewpoint games to?

I could never find much about that!

Yes, the static frame of reference provided by the cockpit nearly eliminates motion sickness. The motion sickness of 1st person shooters is magnified when using the rift. Also, MWO is great for VR because the movement of our mechs is closer to the type of walking movement our brains are used to. The human body can't suddenly change it's direction of travel like your avatar in a FPS. Changing from walking forward to walking backwards instantly over and over while in VR will give most people motion sickness.

The other advantage of our cockpits is it adds to the feeling of presence while in VR. Since both you and your MWO avatar are in a sitting position, it will feel more like you are actually in the game rather than feeling like you are viewing a game. I hope the Devs are also able to add support for magnetic field position sensors to the game. With these sensors you could track your hand and upper body position and map it to the pilot. If they add this support I plan to set up a control system that mimics the layout of the in game cockpit so that I can truly get lost in the game. You could even push buttons in the cockpit at start up with magnetic field sensors.

[p4r4g0n]

@Rebas Kradd Was a year or slightly more ago and content search only goes back to April 2013. I'll see if I can find it but since they reorganized the forum since then, it may be a little difficult ... finally had to resort to google to find the thread.

http://mwomercs.com/...__fromsearch__1

Note that with only 5 posts in that thread, Karl cleared up so much confusion and contradictory speculation that was going on in the forums in a multitude of threads at that time about the forthcoming Elo. Time well spent IMO from a developer's perspective.

Apologies for OP for OT post.

Edited by p4r4g0n, 03 April 2014 - 05:00 PM.

[Loc Nar]

[Alaskan Nobody]

Krinkov, on 03 April 2014 - 04:27 PM, said:

......

I had known all of that -had to work it out on my own since my web searches were rather fruitless (fortunes of the randomness of the world sometimes?)

Feel kinda silly that it seems to be a fairly well known phenomena

[Tekadept]

Karl Berg, on 03 April 2014 - 10:27 AM, said:

I'm also massively susceptible to motion sickness, so my ratio of working to lying flat on the couch with a terrible headache and nausea is about 3 to 4 minutes of work to 4 hours of recovery..

Off Topic, are you able to play a solid MWO for a solid period of time since cockpit glass? I am highly succeptible for motion sickness, and cockpit glass makes me feel unwell,when it was first out id last 10-20 seconds, since it was tweaked id go 3-4 minutes max. I have the same issue with "motion blur" on most games, but i can at least turn that off in other games.

And I don't ever see myself wanting to even try an occulus

On Topic, say a new programmer comes on board, what kind of learning curve would they have in regards to the automatic generation you have implemented? ie good example, taught not to do manual edits etc
and do you find smaller bugs have a greater impact?

I remember reading a book many many moons ago caled the pragmatic programmer who had a section on writing code that wrote code.. But i figured thats how the whole terminator future started

Edited by Tekadept, 03 April 2014 - 10:12 PM.

[Monkeystador]

Thank you for your answer. As you stated i see the same issue with the framerate too. Oculus stated that the customer version could possibly reach 85hz. Take that for a goal.
Rift HD DevKits are due in ~4 month.

[Karl Berg]

Tekadept, on 03 April 2014 - 07:20 PM, said:

Off Topic, are you able to play a solid MWO for a solid period of time since cockpit glass? I am highly succeptible for motion sickness, and cockpit glass makes me feel unwell,when it was first out id last 10-20 seconds, since it was tweaked id go 3-4 minutes max. I have the same issue with "motion blur" on most games, but i can at least turn that off in other games.

And I don't ever see myself wanting to even try an occulus

I've had almost no time for games at all for the last few months What little time I've had in MWO though, the cockpit glass hasn't really bothered me too much. What gets me nearly all the time is head bobs in FPS's. I can't last more than 15 to 20 seconds without starting to get a headache. Oversensitive mouse controls really bother me too; a couple rapid / instant orientations and I'm pretty much done and have to stop.

Oculus was just way too cool. Once it's calibrated correctly, I can use it for a good 15 or 20 minutes without too much ill effect. The major problem for me was getting correct IPD behaviour. Work cycle was tweak projection matrix code, compile, start program, put on headset with eyes closed, quickly open eyes to see if it worked, quickly close eyes when it didn't, repeat.. This maximized work time for me, but I also ended up pretty ill for a while.

Tekadept, on 03 April 2014 - 07:20 PM, said:

On Topic, say a new programmer comes on board, what kind of learning curve would they have in regards to the automatic generation you have implemented? ie good example, taught not to do manual edits etc
and do you find smaller bugs have a greater impact?

I remember reading a book many many moons ago caled the pragmatic programmer who had a section on writing code that wrote code.. But i figured thats how the whole terminator future started

I joke that we could let design 'write code' in the backend now. It's a joke because we'd never let them I guess, but a good chunk of backend work is now done in XML. You mess up the XML, the build system does a static analysis pass and generates compiler errors in the XML files themselves, with nice little descriptions of what you did wrong. I had a screenshot of an example build error in the talk. I think the ramp time is pretty low if you aren't trying to figure out how it all works. If you're the type that needs to know how everything works, then the ramp time is pretty extreme since it is a highly distributed, asynchronous, event driven server platform written from scratch in C++.

As for the code writing code bits, I think of it more like the automation system in Star Trek III. The one that Scotty set up to control the whole ship from the bridge, which worked great up until they hit inadequately specified state (combat), at which point the entire automation system exploded. That feels more like a code auto generation system Only with less fire and exploding things.

[GalaxyBluestar]

really wish this stuff was in a more prominant place on the forums, such info from a dev/pgi worker is MOST invaluable. great communication happening here.