Captains,
TL;dr: We have a ton of stuff going on with our continual improvement program for the 737-700. Want to know more? Come back and read when you have time!
We thought we were going to push out an update for the 737 late this week, but we are still at work on a few pretty invasive changes and decided to sit on our work a bit longer rather than jamming a micro update out the door that winds up breaking something important. Someone once said something about discretion being the better part of valor... or was it something about the wisdom that comes from past mistakes... I forget which.
All the same, I wanted to give you an idea what we are doing so you have a sense of what is taking place in the back room!
Armen is at work on continued improvements to the sound environment. As the airplane has gone out to users, a couple of oddities have appeared and he has been working with a fantastic resource counterpart at Asobo who has been helping him to work through a few changes that will mitigate those issues going forward. Because Armen is Armen, and because professional sound engineers are normally by definition, perfectionists- he has been continuing to cull through the massive library of recordings we made with the live airplane in order to improve upon his already magnificent work.
On the model end, Vin and Jason continue to work through the taskings sent to them by our tech support team. Those taskings are made up of reports provided to our support team directly, as well as a number of observations made here by various users. Many of the details they are tweaking are very small, but the nice thing about suddenly having so many sets of eyeballs looking at the work is that very little goes unnoticed so they are doing that cleanup while simultaneously working on the 600 and 800 projects.
Code side we have a couple of victories to report. I have mentioned these in a few other threads, but they are important enough that I'll put them here too:
- Flashing artifacts on the displays: Some users have seen this phenomenon- most have not. It seems that those who do see it see it occasionally, for the most part. We did see this prior to release, but it was constrained to the EFIS/MAP version of the display layout. We brought it to the attention of Asobo back in September, but they were a bit distracted by their own agenda, we were distracted by ours, and Alex took it upon himself to engineer what we thought was a fix for it. Unfortunately, right before release when SU9 was published by Asobo it came back in even more aggressive form but again only in the EFIS/MAP mode. We made the executive decision to lock you out of the EFIS/MAP option package until we could get it resolved, but then the day we released one of our beta testers reported seeing it in the PFD/ND format. Obviously with many many thousands of times more customers than beta testers- a number of you have also seen it.
We re-floated this to Asobo and the good news is that one very enterprising developer (Thank you, Eric!) reported back the following day that the source of the problem had been identified as a problem in the GDI+/NANO layer of the sim has been fixed internally at Asobo. The fix will roll out to customers concurrent with SU10. There really isn't much you can do to trigger it or prevent it- it seems to be something that either happens or doesn't... But- they found it.
- Pause in all forms seems not to play well with the 737: We have also elevated to Asobo something else from our September list of "stuff" we reported and that is "how the airplane behaves when the user exits to a menu or pauses." When in pause, the sim doesn't currently communicate to C++/WASM based projects that it is in pause. The airplane's core logic continues to think that the airplane is flying happily along, but it cannot understand why the position has stopped changing or why the airplane is not responding to control commands- so it keeps trying. This can lead to horrendously convoluted results when you unpause the thing.
We have been asking for some time for the ability to detect that the sim is in pause- and we were informed early this week that it is being investigated. We thought we'd hear results by end of the week, but I suspect they simply forgot so I'll jangle the line early next week in hopes that this fix will also appear in SU10.
- A Few Items Related to Control Inputs: One of the invasive items we are working on involves finding a way to make the 737's rudder/ground steering inputs play well with game controllers and twist-grip style joysticks. The 737 (and all Boeings) have a very nice, but complex steering interconnect mechanism that makes flying them in crosswind conditions a joy. Changes made to the nose steering axis by asobo with SU9 broke this functionality, and we are still awaiting an updated SDK so that we can get a better idea what it is they changed. We have asked a number of times on the topic and the answers have been vague, which tells me the person getting the question probably isn't the same person who made the changes in the platform. We have tried a number of solutions that can best be described as hacks to get the game controller and twist grip stuff working properly- but haven't found anything stable enough to call a solution. Work continues.
- Choppy Animation of the control column: This one is peculiar and would be simple to solve if we had a debugger that allowed us to work with the code in real time. Unfortunately we are left here creating software tools to dump data out of the sim into text files in real time so that we can see what various inputs and outputs are doing- and this is a bit like reading year old newspapers in order to find investment advice. The date we are looking at in this environment is so outdated that it doesn't tell us what the simulator is giving to the model the control the animation until long after the event has taken place- and then we aren't entirely sure which event corresponds to which test. Sounds fun, doesn't it? My kingdom for debugging capability more advanced than what we used to do back in 1992.... It remains under investigation.
- TrimWheel Animation Improvements: An area that I am working on is improvement in the animation control for the trim wheels. We actually took high speed video of a marked trimwheel operating in various trim modes in order to correlate it it to the rigging documentation so that we can determine how fast/slow the wheel will turn and how many radians it will rotate per index of ANU/AND trim. I thought we had done this once before, but apparently we hadn't actually corroborated the results in a live airplane previously and I suspect the documentation may not be telling us the entire story. Needless to say, this change will look really nice when we give it to you. (It **might** slide to the next update if I run out of time before this update- which is possible because of my flying schedule next week...)
-Pop-Ups of Displays: We have mentioned a few times that the C++ methods required to make the displays in the 737 "pop up" have been locked out of availability, thus preventing us from having the displays pop up as we have done in all of our products going back to 2008. Mathijs and Hans, or good friends at Aerosoft, reached out to lets us know we may have missed a notation in an update cycle of the SDK a while back that provides a provision to make this possible. We don't know for certain if it works yet, as we haven't had a chance to implement their recommendations- but I expect we will get to testing it on Mon/Tue next week and will report results. We use a different display drawing method than Aerosoft, but based on what I know technically, I don't think this will be problematic.
BTW: You frequently hear me talk about how much back-chat goes on between developers and how one aspect of the developer community that I really enjoy is how much we all try to help one another un-stick problems in order to benefit the community. We have worked with Aerosoft since the year 2000 and Mathijs and Hans are both good friends to PMDG and to the dev community as a whole. It is always nice knowing another developer has your back and will ring you up to offer good advice if they see you knock your knee on something.
- A Ton of Other Items: We have a change list that already surpasses 70 items that have been changed for the upcoming update. These include a range of things from code corrections, documentation corrections, model fixes, texture improvements, clickspot tweaks and more.
DANGER: No, I didn't list your favorite "can you tell me if <something> is included because it is clearly the most important thing you should be doing."
That doesn't mean it isn't/hasn't made the list. It also doesn't mean it won't make the list later. We don't answer those kinds of questions because it becomes a never ending challenge/reply litany that takes a ton of time and doesn't really help anybody out. Your patience is appreciated.
- Timing of the update: I originally thought we would put the first of two updates into testing on Tuesday and release it to you on Thursday or Friday. We are deep into some fixes and just need more time. Right now my expectation is that we'll give the update to our testers on Wednesday for verification and then push to you on Thursday or Friday. There is some chance we might go earlier, some chance we might go later. All depends upon how it goes on Mon/Tue!
- The 737-600 and 737-800: Both are moving through development on schedule.. . We expect to have review images of the 600 shortly! (See what I did there? Shortly? 737-600? SHORTly?
)
TL;dr: We have a ton of stuff going on with our continual improvement program for the 737-700. Want to know more? Come back and read when you have time!
We thought we were going to push out an update for the 737 late this week, but we are still at work on a few pretty invasive changes and decided to sit on our work a bit longer rather than jamming a micro update out the door that winds up breaking something important. Someone once said something about discretion being the better part of valor... or was it something about the wisdom that comes from past mistakes... I forget which.
All the same, I wanted to give you an idea what we are doing so you have a sense of what is taking place in the back room!
Armen is at work on continued improvements to the sound environment. As the airplane has gone out to users, a couple of oddities have appeared and he has been working with a fantastic resource counterpart at Asobo who has been helping him to work through a few changes that will mitigate those issues going forward. Because Armen is Armen, and because professional sound engineers are normally by definition, perfectionists- he has been continuing to cull through the massive library of recordings we made with the live airplane in order to improve upon his already magnificent work.
On the model end, Vin and Jason continue to work through the taskings sent to them by our tech support team. Those taskings are made up of reports provided to our support team directly, as well as a number of observations made here by various users. Many of the details they are tweaking are very small, but the nice thing about suddenly having so many sets of eyeballs looking at the work is that very little goes unnoticed so they are doing that cleanup while simultaneously working on the 600 and 800 projects.
Code side we have a couple of victories to report. I have mentioned these in a few other threads, but they are important enough that I'll put them here too:
- Flashing artifacts on the displays: Some users have seen this phenomenon- most have not. It seems that those who do see it see it occasionally, for the most part. We did see this prior to release, but it was constrained to the EFIS/MAP version of the display layout. We brought it to the attention of Asobo back in September, but they were a bit distracted by their own agenda, we were distracted by ours, and Alex took it upon himself to engineer what we thought was a fix for it. Unfortunately, right before release when SU9 was published by Asobo it came back in even more aggressive form but again only in the EFIS/MAP mode. We made the executive decision to lock you out of the EFIS/MAP option package until we could get it resolved, but then the day we released one of our beta testers reported seeing it in the PFD/ND format. Obviously with many many thousands of times more customers than beta testers- a number of you have also seen it.
We re-floated this to Asobo and the good news is that one very enterprising developer (Thank you, Eric!) reported back the following day that the source of the problem had been identified as a problem in the GDI+/NANO layer of the sim has been fixed internally at Asobo. The fix will roll out to customers concurrent with SU10. There really isn't much you can do to trigger it or prevent it- it seems to be something that either happens or doesn't... But- they found it.
- Pause in all forms seems not to play well with the 737: We have also elevated to Asobo something else from our September list of "stuff" we reported and that is "how the airplane behaves when the user exits to a menu or pauses." When in pause, the sim doesn't currently communicate to C++/WASM based projects that it is in pause. The airplane's core logic continues to think that the airplane is flying happily along, but it cannot understand why the position has stopped changing or why the airplane is not responding to control commands- so it keeps trying. This can lead to horrendously convoluted results when you unpause the thing.
We have been asking for some time for the ability to detect that the sim is in pause- and we were informed early this week that it is being investigated. We thought we'd hear results by end of the week, but I suspect they simply forgot so I'll jangle the line early next week in hopes that this fix will also appear in SU10.
- A Few Items Related to Control Inputs: One of the invasive items we are working on involves finding a way to make the 737's rudder/ground steering inputs play well with game controllers and twist-grip style joysticks. The 737 (and all Boeings) have a very nice, but complex steering interconnect mechanism that makes flying them in crosswind conditions a joy. Changes made to the nose steering axis by asobo with SU9 broke this functionality, and we are still awaiting an updated SDK so that we can get a better idea what it is they changed. We have asked a number of times on the topic and the answers have been vague, which tells me the person getting the question probably isn't the same person who made the changes in the platform. We have tried a number of solutions that can best be described as hacks to get the game controller and twist grip stuff working properly- but haven't found anything stable enough to call a solution. Work continues.
- Choppy Animation of the control column: This one is peculiar and would be simple to solve if we had a debugger that allowed us to work with the code in real time. Unfortunately we are left here creating software tools to dump data out of the sim into text files in real time so that we can see what various inputs and outputs are doing- and this is a bit like reading year old newspapers in order to find investment advice. The date we are looking at in this environment is so outdated that it doesn't tell us what the simulator is giving to the model the control the animation until long after the event has taken place- and then we aren't entirely sure which event corresponds to which test. Sounds fun, doesn't it? My kingdom for debugging capability more advanced than what we used to do back in 1992.... It remains under investigation.
- TrimWheel Animation Improvements: An area that I am working on is improvement in the animation control for the trim wheels. We actually took high speed video of a marked trimwheel operating in various trim modes in order to correlate it it to the rigging documentation so that we can determine how fast/slow the wheel will turn and how many radians it will rotate per index of ANU/AND trim. I thought we had done this once before, but apparently we hadn't actually corroborated the results in a live airplane previously and I suspect the documentation may not be telling us the entire story. Needless to say, this change will look really nice when we give it to you. (It **might** slide to the next update if I run out of time before this update- which is possible because of my flying schedule next week...)
-Pop-Ups of Displays: We have mentioned a few times that the C++ methods required to make the displays in the 737 "pop up" have been locked out of availability, thus preventing us from having the displays pop up as we have done in all of our products going back to 2008. Mathijs and Hans, or good friends at Aerosoft, reached out to lets us know we may have missed a notation in an update cycle of the SDK a while back that provides a provision to make this possible. We don't know for certain if it works yet, as we haven't had a chance to implement their recommendations- but I expect we will get to testing it on Mon/Tue next week and will report results. We use a different display drawing method than Aerosoft, but based on what I know technically, I don't think this will be problematic.
BTW: You frequently hear me talk about how much back-chat goes on between developers and how one aspect of the developer community that I really enjoy is how much we all try to help one another un-stick problems in order to benefit the community. We have worked with Aerosoft since the year 2000 and Mathijs and Hans are both good friends to PMDG and to the dev community as a whole. It is always nice knowing another developer has your back and will ring you up to offer good advice if they see you knock your knee on something.
- A Ton of Other Items: We have a change list that already surpasses 70 items that have been changed for the upcoming update. These include a range of things from code corrections, documentation corrections, model fixes, texture improvements, clickspot tweaks and more.
DANGER: No, I didn't list your favorite "can you tell me if <something> is included because it is clearly the most important thing you should be doing."
That doesn't mean it isn't/hasn't made the list. It also doesn't mean it won't make the list later. We don't answer those kinds of questions because it becomes a never ending challenge/reply litany that takes a ton of time and doesn't really help anybody out. Your patience is appreciated.
- Timing of the update: I originally thought we would put the first of two updates into testing on Tuesday and release it to you on Thursday or Friday. We are deep into some fixes and just need more time. Right now my expectation is that we'll give the update to our testers on Wednesday for verification and then push to you on Thursday or Friday. There is some chance we might go earlier, some chance we might go later. All depends upon how it goes on Mon/Tue!
- The 737-600 and 737-800: Both are moving through development on schedule.. . We expect to have review images of the 600 shortly! (See what I did there? Shortly? 737-600? SHORTly?
)
Rolando Olmo
Danny Z.Cebis
Comment