WebAssembly

[n4gix 4,947]

8 hours ago, domkle said:

Have Asobo and MS to adapt their toolset and architecture to accommodate the programming skills and pet peeves of the 3PD and their strong desire to release fast older products in their present difficult cash situation or do the 3PD which want to board the FS20 flag ship should acknowledge that it is a whole new environment and adapt ? 

Perhaps I haven't been clear enough. At this juncture I'm not at all confident that any 3rd party developer will be able to code any truly advanced avionics (such as ProLine 21 and Collins FMS) even from scratch, much less "convert" existing assets.

I freely admit that by the time MSFS is ready for release, the issues currently blocking such may have been resolved, but that will nonetheless extend the time required for coders to have product ready for use.

[mSparks 2,086]

1 hour ago, MatthewS said:

Maybe you'd like to quote his exact words and also give the minute/second mark he says this?

the timestamp I posted, plus 10.24

Quote

my dream is to have

everything ready at the get-go but you

guys don't realize that's just not gonna

happen but we're trying to do everything

that we can to build a good base let's

give you the world the weather let's

give you the right aircraft the right

conditions to fly in and let's build

keep building on that when you guys tell

us we're getting things right now they

think it's the next step

Perhaps you could find any hint in that video or anywhere of the opposite.

As I said from day one - that got me so much hate from the MSFS astroturfers - to the point of wanting to destroy my third party content - That Microsoft guys ENTIRE idea of 3rd party content is posting videos and pictures of what we don't like.

On topic - because XBOX - hence Webassembly.

At this point I just hope PMDG will be handling the commercial simulation, and will be in need of an FAA/ICAO approved ATC course they can drag and drop into their sim.

Edited April 6, 2020 by mSparks

[GlideBy 324]

10 hours ago, domkle said:

Asobo is obviously at a crucial point in the development of the platform. Should they focus all their energy to stabilize it and respect their deadlines or focus on improving an embryonic SDK to cater to the wishes (even legitimate) of the 3PD community ?

What deadline?  They never said it would ship in 2020, just that they hoped it would.  I want to point this out so that people don't think they missed a deadline if it ships next year.  And also so they don't feel pressure to rush it out before it should go out.

[Dominique_K 10,139]

3 minutes ago, GlideBy said:

What deadline?  They never said it would ship in 2020, just that they hoped it would.  I want to point this out so that people don't think they missed a deadline if it ships next year.  And also so they don't feel pressure to rush it out before it should go out.

Did I say anything about the release  year in my post ? Don't think so.

The wish that "they don't feel pressure to rush it out before it should go out" is, excuse my French 😉,  mushy forum gibberish. It is not like that the real world works. A contract between a corporation and a subcontractor has always deadlines, intermediate and final, with penalties when not respected. I would be greatly surprised if it was not the case here.  Asobo needs to focus on their deadlines to maximize their revenue.  

Then, on a pure selfish user point of view, I'd like better that they focus on getting  their sim out  ASAP than on polishing the SDK. 3PD addons will follow in due time. As always.  I can wait a little longer. For the record, I said many times and still do that I look forward to the SDK to make my own little sceneries. But the priority for me is to see  the sim out.

 

[Lorby_SI 2,038]

2 hours ago, n4gix said:

Perhaps I haven't been clear enough. At this juncture I'm not at all confident that any 3rd party developer will be able to code any truly advanced avionics (such as ProLine 21 and Collins FMS) even from scratch, much less "convert" existing assets.

I freely admit that by the time MSFS is ready for release, the issues currently blocking such may have been resolved, but that will nonetheless extend the time required for coders to have product ready for use.

Why all the doom and gloom - why not try a different pitch: "We don't know much about the actual implementation yet, the SDK is still in development. But thanks to our many years of coding experience we are confident that we will be able to deliver highly sophisticated products in a reasonable time frame." 

For example: if you want access to the local file system, Asobo could provide an API for that inside their WASM. Or they provide a socket API so you could communicate with an outside data provider module. Or some stuff will be provided as (web-?) services. Etc. 

Unless things have been tried, tested and proven to be technologically impossible, I like to think that there always is a way. 

Best regards

[GlideBy 324]

5 minutes ago, domkle said:

Then, on a pure selfish user point of view, I'd like better that they focus on getting  their sim out  ASAP than on polishing the SDK. 3PD addons will follow in due time. As always. 

You keep saying deadline.  What deadline?  You are assuming they have a deadline.  Often creative works don't have precise deadlines.  And game deadlines depend on the developer.  In other words, many of the best games had no deadline.  And many of the games that had a deadline, turned out badly, and never recovered.

I know people look at the sim, and say it looks great, it looks done, just let me use it now.  But we aren't seeing all the negative feedback in the forums.  

Devs do listen to users.  There have been many times where devs listened to users demanding the game be released, and the game bombed.  Devs listen to users, because those are the customers.  The devs know the issues, but based on user feedback they assume that users won't mind too much, or that sales will be great any way.  They release it, it gets bad reviews, sales are not great, the devs don't have money to fix it, and both the users and devs move on.  And a game that had great potential dies in the "what if" dust bin.

Please don't pressure the devs to release before it is ready.

 

 

[Dominique_K 10,139]

@n4gix

A  honest answer. Thanks. That deserved to be said before we see an early wave of desinformation born in this thread become a tsunami. The issue is about very specific complex avionics and not the proverbial "Study Level Aircraft" in general. I consider my  A2A T-6 or Comanche as a study level even if they don't have the Collins wizardry. This is a feeling you have considering what you know at this time of the SDK, not something which would be carved in marble

@GlideBy Obviously you don't understand what I said and keep repeating the same mushy nonsense. No need to continue for me.  

@Lorby_SI Thanks for putting some optimism in the debate. I find it reasonable optimism because this is exactly how I have seen flight simulation evolves over the decades.

 

Edited April 6, 2020 by domkle

[JB3DG 225]

I see several posts asking what is/isn't possible with WASM and pointing to the advanced displays already developed.

So a couple things to point out, since glass is my specialty in my job, with the caveat that I have been heads down on work and not paying much attention to the latest FS updates.

1: (See caveat) So far the only "advanced" display is the G1000. So for starters, try comparing the terrain and weather radar displays of the 737NG, the A320 family, the PL-21, and a wide range of other advanced glass displays. Spoiler alert: They are VASTLY different from eachother.

2: Compare the navigation symbols for waypoints and flightplans etc. You will find commonality and diversity.

3: Compare the synthetic vision of the Garmin displays with the Proline Fusion, the various HUD SVS displays, the Honeywell Primus Apex (I think that's what Dassault or Gulfstream use in their jets, I forget which). Again, considerable differences not only in terrain rendering but also airport markers.

4: On the topic of airport displays on navigation displays, again lots of variety.

5: Unless provided access, no IR/TV cam views for enhanced vision displays on HUDs, or taxi cams in the cockpit.

Now a quick break before I continue on other points. What do all of the above have in common? Advanced raster image generation. Not possible with a simple 2D vector graphics library. If, and this is a big if (Those who know, can't say due to NDA, those who don't, don't know for certain), If Asobo has designed their background graphics engine to interface with the HTML or WASM or whatever they use specifically for the G1000, then it is tantamount to absolutely useless for most of the common avionics suites in use in other aircraft, beyond the most basic of functions (magenta line, text, compass rose or ADI/VSI/ASI/Alt/Engine instruments etc). Forget about custom designing advanced map and navigation displays. The current advantage of even the old FS9/FSX display system is that we had direct access to the raw image buffer for each gauge, and could use whatever rendering method we liked for advanced raster image generation. With P3D's PDK it got even better because we no longer needed to implement obscure weird low level CPU/GPU interfaces to get images that had been rapidly generated on the video card into the CPU buffer we had access to for the gauge image, which then went back onto the video card for display in the sim. Instead we got direct access to the DirectX texture surfaces which are mapped to the 3D model or displayed on the screen in 2D popups. This allows use of custom shaders for extremely efficient and rapid raster image generation which is so critical for navigation displays. It also allows considerable streamlining of rendering arrays of small objects like VOR/airport/Waypoint symbols. All of this disappears when locked down to a 2D vector library with no access to the underlying 3D rendering engine objects. 

The mention of lack of thread support is another concerning item, even if a sandboxed file access is granted. Sometimes files are large enough that they need to be loaded on the fly, or conducting searches/indexing in them once loaded in memory needs to be threaded to run in parallel or turn the sim into a slide show every now and then.

Next, there are a good many windows and other external libraries used by developers that may not necessarily be available in WASM, unless specifically granted access.

And another item: Debugging. Visual Studio has an exceptionally great debugger allowing the use of breakpoints to intercept code execution wherever we like and examine the state of variables and objects and even raw process memory (a critical item. I have lost count of the number of occasions I have been able to trace corrupted memory via direct views of raw memory and working through ASM registers etc). I have my doubts as to whether WASM will support such low level debugging, which gets more and more critical the more complex an aircraft becomes.

Oh, and one last item: Forget about using the F1 or RXP GTN/GNS/G500/G600 gauges which use the actual Garmin trainers. That involves extremely low level C++ inter-process communications which WASM is explicitly designed to *prevent* with its sandboxing. 

Another last edit: From the comments about speed of execution, forget about external flight models, where speed of execution is absolutely *critical* for best performance.

Edited April 6, 2020 by JB3DG

[JB3DG 225]

On 4/4/2020 at 6:16 AM, devgrp said:

My understanding was that wasm was being developed to port older fsx code to the new sim. I find it hard to believe that asobo is calculation rainbows, pressure etc and not being able to make complex study level avionics. If that was the case why did pmdg of all devs get on board and is all in? 

When PMDG made their announcement, there was literally *zero* information on what the SDK would look like beyond some vague statements about SimConnect. My personal opinion is they assumed that not giving developers low level C++ access would be suicidal to the success of the project. Where they are now, I do not know. They may, like most devs, including myself, be taking the position of "we will wait and see". If the rumblings mentioned by other devs here continue till the public release, I would say expect a reversal on PMDG's part.

Edited April 6, 2020 by JB3DG

[JB3DG 225]

And lastly, some comments on the hope that devs will "find a way", or "hack" etc.

Forget it. WASM is explicitly designed to *prevent anyone* from doing just that. Those who use it, are entirely at the mercy of the authors of the backend driving it.

[GlideBy 324]

1 hour ago, domkle said:

mushy nonsense.

I have been respectful, so I am not sure why I deserve this comment. 

And then to imply I don't understand, and just walk away.  I read your comment a few more times.  I am sure I know what you wrote.  

1 hour ago, domkle said:

It is not like that the real world works. A contract between a corporation and a subcontractor has always deadlines, intermediate and final, with penalties when not respected.

Let's assume that they do have a contract with a deadline.  Do you think it is possible for the corporation to extend the deadline?  Do you think they could base this decision on user feedback?

[Milviz 919]

I am 100% sure that these decisions are not made with a timeline in mind.  They are made because MS has a plan and that plan is being put into play.  What that plan is... remains to be seen.  That said, to us, anyway, it's obvious that this doesn't bode well.  And we're not alone in thinking this... I just don't mind (too much) poking the bear... The bear needs to be poked in order to remind it that it's US that helps THEM to be better at what THEY do...

[Murmur 3,941]

I don't particularly care about the so-called "study level" aircraft, so provided that simulated aircraft allow realistic-ish avionics and nav techniques, I'm ok.

I understand though that many simmers look for super-realistic avionics. But I would be really surprised if MSFS won't allow that. Certainly would be a peculiar choice in the history of the franchise. Despite the latest news on WASM, I'm still optimistic.

 

[mSparks 2,086]

2 hours ago, JB3DG said:

And lastly, some comments on the hope that devs will "find a way", or "hack" etc.

Forget it. WASM is explicitly designed to *prevent anyone* from doing just that. Those who use it, are entirely at the mercy of the authors of the backend driving it.

Exactly this.

1 hour ago, Milviz said:

What that plan is... remains to be seen.

The XBOX and PC version are [v likely] identical - "no exclusives", so anything built for MSFS2020 PC has to run identically/be identical to MSFS2020 for XBOX. even first party devs can't access the XBOX filesystem post production.

WASM likely the best they can offer.

Edited April 6, 2020 by mSparks

[MatthewS 761]

6 hours ago, mSparks said:

Perhaps you could find any hint in that video or anywhere of the opposite.

No sorry, I can't find any hint.

But I do agree with you that FS2020 is turning out to be a real disappointment 😥

The video is very interesting... IMO, the 5 Asobo guys  👨‍💻  👨‍💻  👨‍💻  👨‍💻  👨‍💻  are wanting to make FS2020 the best it can be BUT what a surprise, look who is sitting on the right, it's their Microsoft-Overload-BeanCounter 🤑, his posture and presentation is very arrogant and he's not going to let anybody's hardcore study-sim wet dreams rain on his development budget and eventual sales targets (and probably his bonuses too). 🤣

Edited April 6, 2020 by MatthewS