whoa, he said this really? even adding these as a user parameter would’ve been greatly appreciated. but i guess they gets to decide their target audience. i feel so dirty now. lol.
inNative’s tracking issue for implementing the Fixed-Width SIMD proposal is Implement Proposal: Fixed-width SIMD · Issue #43 · innative-sdk/innative · GitHub if you want to keep an eye on it or contribute (inNative is licensed under
Apache-2.0). Asobo will still have to update their vendored inNative & enable the proposal’s flag once it’s implemented, but I’d expect that to happen soon after the functionality is stable.
Thanks for the link!
Do I understand correctly they intend to break some of the SIMD speed advantage in floating point operations:
The floating-point operations in this specification aim to be compatible with WebAssembly’s scalar floating-point operations. In particular, the rules about NaN propagation and default NaN values are the same, and all operations use the default roundTiesToEven rounding mode.
AFAIK, one particularity of floating point operations in SSE is that they are not abiding by the IEEE rules. This permits some operations (with NaN and/or other specific cases) to fall through and still returning “sensible” results (better suited for multi-media) instead of error results (better suited for maths).
I’m not saying it isn’t going in a good direction as a standard, because in effect this will help leveraging more than 80% of the actual silicon in our CPUs (the transistors in there handling the SSE opcodes). It would be much more simple and immediate, and risk free for the Market Place and the Flight Simulator.exe process and its address space, let alone for its WASM stack and JS/HTML layer, to actually just adding a few new APIs in the SDK to supporting shared memory and sockets.
Actually there is a loophole in their implementation which I won’t document yet, which allows loading any 3rd party DLL and hacking the process (this is not a system wide hook). This is a higher security risk than the gauges.
I’d recommend looking around for answers to this question at Issues · WebAssembly/simd · GitHub , or opening an issue with your question there if you can’t find an answer. You’ll get better answers (and chances for improvement) out of the people actively working on the spec than me.
If you know this to be a vulnerability in inNative and not MSFS2020’s embedding, there are invite links to its Discord guild on its site or in its README, and you can PM Erik McClure about the vulnerability from there. Thank you for your responsible disclosure.
I’m talking about a FS2020 loophole, defeating the whole point of sandboxing 3rd party devs for security.
In other words as of today, and most likely during the lifespan of FS2020 given the nature of the loophole, you can inject a virus in FS2020 and stole customer credentials via the loophole, but 3rd party devs can’t even access the bare bone metal for faster and better add-ons.
I understand the reasons behind WASM and this is a good technology for what it is good at, but it shouldn’t be used in my opinion to make customers fearing 3rd party vendors are going to stole their personal information if they are not sandboxed thanks to Microsoft and Asobo… Otherwise in my opinion this is a fallacious argument, telling the 3rd party vendors are thieves and customers must be protected from them.
- DLLs can carry unsafe code and they can’t have unsafe code in add-ons in the Market Place. Sandboxing is good for having safe add-ons in the Microsoft store.
- Sandboxing is also a good way to port legacy addons in C++
Eric – When you ship a DLL, you never know what it is going to do! It has access to personal informal information and result in having security issues. Using WASM, we are sure that 3rd party content only accesses designated files!
NB: you’ll find my other comments about this and the SDK from the link above.
Doesn’t FS2020 also use Wasm to enable publishing of addons on console, where there are restrictions on JITing external code? If you properly ship Wasm that doesn’t rely on Windows-specific bodges and is self-contained, that can be compiled on the Marketplace side and shipped to consoles as native code without changes in behavior. Additionally, you gain easier portability between desktop OSes.
I agree 200% with you, and like I said above: “I understand the reasons behind WASM and this is a good technology for what it is good at”
Nevertheless, there is no technological reasons to forbid better constructs and access on the PC, just because you can’t on the Xbox, especially if there are means to extend the PC capabilities in terms of in-container/out-container IPC with safe technologies like Shared Memory. I mean what is the risk with IPC really?
- On the WASM side, it is all array-bound checked
- At the OS level, you can add guard pages and trap
- At the external exe level, you’ll be also guarded by the OS.
- You can’t inject code via IPC (but you can with WASM using similar techniques as those used with Spectre)
- You can’t run code from IPC (actually you can run code planted there from the external app side of things, but the WASM side will just treat this as bytes of data, not code).
In short, for those not knowing what this is about, it is sharing a portion of memory between 2 processes. Each side can read and write this memory and if you read or write outside its range, Windows will guard and trap the faulty application (safe guarded at the lowest level).
And even if later there is a way to plant opcodes in the IPC memory and make it run from the WASM side (doubtfully), it would be so FS2020 SDK specific that I can assure you there are already easier way to do so right now.
Otherwise, you’re explanation and the official one are differing a lot.
I don’t know what customers are understanding in all this technology centric discussions:
- Even if that would been the case, there is no need for a poison-the-well tone or trying to belittle…
- …and that is quite an odd take on it, it’s not what he said and rather taken out of context to be honest:
Developer Q&A Twitch Series: Top Community Question [01.27.21] - YouTube
Developer Q&A Twitch Series: Top Community Question [01.27.21] - YouTube
I reckon you refer to the when he talks about it in the second part…
I think it’s enough to watch this, to understand that they are working on some serious simulation models, where some parts are not implemented in Xplane as an example.
Feature Discovery Series: Aerodynamics Part 2 - YouTube
Is it perfect? No, nor they claim it to be, they keep working on it and say there are improvements to be made.
this is the exact moment I’m referring to (at 10:45):
There is no doubt I agree they are improving things everyday, and I’m not entering a XP11 vs FS2020 debate because this is not the point.
As a 3rd party developer there are conflicting messages and communications between a “sim” and a “game” orientation (and this terminology is debatable but I believe we’re all intelligent enough in this forum to understand the “gross” distinction posters are referring to in general).
In my opinion you can’t expect other users not questioning whether some of the actual flight simulation centric aspects, which are wind effects (it would be tide/wave effects for a boat simulator by analogy) are dumbed down for the sake of making it a better simulator, or something else.
I also believe there is a difference between not understanding what this is about as a developer of a simulator (wind effects), and knowing what this is about but restricting it. The former would definitely tell they have no clue what they are doing, the latter means not only they know, but they are voluntarily limiting it. Which begs the question and illustrates one type of “conflicting messages and communications” I’m referring to.
PS: here are some other examples:
What does it have to do with the discussion that you are a third-party dev? I fail to see what it has to do with what you read into what Sebastian said in the stream and how you see conflicting messages in that?
I think you read too much into certain things and also have a rather old-fashioned view of the scope of a software product, for me MSFS can be used the way you want to. Whether you enjoy flying around with all assistance on or casually look into things from the drone view or if you enjoy to dig into the manual of the CRJ before doing some more serious flights with it. But to use your scope of definitions, MSFS is without a doubt a sim first and foremost.
If you believe that they have not gone all in for some specific draft scenarios and hence cater for “gamers” seems like an odd take on it, I think getting stuck in such details makes you miss the greater picture of this project.
It’s a bit like if you do e-sport with the F1 game, then it’s ■■■■ serious, but then you can also drive around with your Xbox controller, using all assists (auto gears, brake assistance etc) and having a beer, it’s what you make out of it really and same goes with MSFS, which is not a weakness but a ■■■■ strength.
OH man, all this dumbing down talk
This is very concerning to me! I wanted and hoped this was going to eventually be on par with or close to X-plane, I thought it was going that way as well, now I am reading all this…shame really. I guess it kind of confirms my fears since hearing about the XBOX plans which I feel is huge mistake in the sims current form. Think I am going to hold off any more investment in this sim until there is a clear picture of where it is going.
too bad eh?
If things like this are not going to be possible
Then it’s kinda pointless to the more hardcore flyers.
I’m not arguing with you at all, I’m reminding you some facts.
Please rewind the topic from the beginning.
@EvidencePlz is expressing his opinion on the simulator:
To which you’re judging his comments in the following way:
I’m helping you understanding why he might think so, and this is because Asobo is explaining themselves they are voluntarily dumbing down the simulator because users couldn’t otherwise understand what is happening to their aircraft. This are Asobo own’s comments, not mines, not @EvidencePlz either:
Now of course you can have an opinion like anyone else in these forums, and more generally in life, nevertheless, you shall respect others to having theirs, and to some of them, there is a conflict of terminology or destination when you’re dumbing down a flight simulator root component of flight.
No one has to agree with any one else here, every one is just sharing his/her opinion, but there is no reason to judge others opinions because they do not match yours in my opinion (no pun intended). Instead I believe (a belief, not a fact) the more we are sharing our opinions, however diverse they are, the more we can understand each other. And even if we don’t, the very fact a single subject could have so diverging opinion, should be enough of a signal to Asobo to reconsider the way they are implementing some of the aspects of the simulators, or the way they are communicating about it. Otherwise they’ll never make it clear to the larger audience, and divergence of opinions will remain.
Sadly and honestly, some on here are really resisting any additional complexity. The sim could suite all, but it does not seem that way. I’m not going to buy anything else now for this, until I know 100% its going to be a true sim with options to dumb it down if that’s how it is to be.
I love the graphics, they are amazing, XP11 cant compare (Graphics, weather is now on par or close enough with XEnviro / SkyMaxx), however its not all about those Pixels lol. Realism for me is key!
I’m sorry to say but these are Sebastian own words: “We have dumbed down up drafts otherwise people wouldn’t understand what is happening with their aircraft”. This tells it all about the subject matter to me, but I might be understanding it wrongly too.
He is not only admitting to dumbing down certain things, he also seems quite uncomfortable talking about it, remaining very fuzzy and not communicating any reliable information here. Not what will change and not when. It seems that this a hot topic within the dev team and they are not seeing eye to eye there, or maybe Microsoft wanted them to dumb things down.
All pure conjecture of course, but you can’t betray body language and something is happening there.
But the most important thing here is that they take away the learning experience of simming to not disgruntle their customers, which disgruntles costumers. So what happens if people don’t understand the physics of updrafts? Are they throwing their XBOX controller at their TV out of frustration or will they start researching?
The latter is what brings people into this hobby, trying to understand what is happening to rise to the challenge and beat it. I have no idea why the goal is to take this away from players. Why not make it optional in the first place?
Create a section in the option solely dedicated to wind simulation. No more 50% wind reduction under 1000ft AGL, no more reduced drafts, actual windshear and wake turbulence. I really like MSFS but it’s sometimes hard to defend against people who attack the weaknesses.
He was talking about how you can get 10,000 feet per minute updrafts in severe thunderstorms, and toning that down not saying they ‘dumbed down’ anything.
Surely ‘real simmers’ would never notice since they wouldn’t fly into a thunderstorm anyway, that would go against procedures and all.
To the OP, I think its a mix of their strategy and vision for this game, reach as wide as a audience as possible, and most people arent hardcore flightsimmers, and also I think it suffers from the eyecandy first everything else second issue that Cyberpunk had. Most of the juice have gone into eyecandy, scenery, blackshark ai tech and so on, so it seems like the systems that allows for all the details hardcore simmers wanted was just not pri1 or 2… or # At least thats how this whole “sim” feels to me.
I think it becomes a problem when even casual simmers like myself and my friends do not really see a reason to play MSFS because of how, lets call it simplified, it is. Keep in mind we dont hear the voices of all those who chose not to play this game, because they simply arent here. So the voices we hear are from people already invested into this product and vocal about it on these forums, and if you keep playing it chances are you really like what it is. Point is that it gets very biased and not very nuanced, and Im not sure if thats very healthy for simulator that clearly have some issues regarding details and realism. And lets be honest, the majority of the players use msfs because of the eye candy and taking photos, which is fair - to each their own. But if you were hoping for the next generation of groundbreaking flight simming, its difficult to justify the flight dynamics and physics or the lack of in msfs.
I have been trying to get a casual simmer friend of mine to play msfs with me, but he refuses because the default aircrafts are so stripped down and the flight dynamics and physics in general makes it feel like they are on rails. I managed to get him to try the fbwA32NX thinking it would be the best of what msfs could offer and it literally hurt my eyes and soul when I watched him stream his experience. All assistances set to hard, stormy weather, and he was intentionally trying to get a feel for the realism of the physics by yanking the airliner around, stalling, turning off engines midflight, trying to nosedive up and down left and right like a rollercoaster, yet the huge airliner was gliding around smoothly like it was some scripted glider with infinite power (way way too forgiving, but I guess that is partially why it is so popular too). It hurt watching the lack of well everything. He said he had no feeling of the air or aircraft responding to anything at all, it was like it was on rails - some arcade casual whatever. It just hurt my soul watching, since we both spent alot of hours in xplane with the zibo737.
I have no clue about the technical stuff and the spesifics like you guys discussing here in this topic. But at the end of the day its clearly a issue that stops even casual simmers from playing it, well it is only a issue if its not intended, which is what I get a feeling it is with statements made regarding limiting this and that.
I mean, most people use it to take pictures, its referred to as a scenery- and photo-simulator. It has a UI in external view for a reason, its being ported to xbox for… well several reasons. Look at the ratio of topics on this forum, bugs, graphics, 3rd party stuff and screenshots are the majority. Theres not a lot of attention regarding realism in systems and/or physics, flight model and dynamics. Heck even most of the updates to the sim itself are scenery stuff, graphics stuff and bugfixes. In development for many years, beta for a year and retail for half a year - to me, this is just not cutting it in 2021.
What if they just relax a little with the “dumbing down” and rather set their limitations as optional, lets say, in the assistance menu?
I feel the exact opposite. To me you can not use it how you want to, it very much has a upper limit/bar which is in my personal opinion set way too low. On the other hand the lower limit is so low even a child with 0 experience could manage just fine, heck just slew mode around and spam the screenshot key if you want. I too LOVE the CRJ, but even that seems to be limited to the very physics and flight dynamics msfs has to offer, or lack of. But me as a casual+ simmer think it is sad when I dont feel there is any challenge to learn or fly in this game even in the currently most “realistic” payware airliner (regarding airliners), to me its all in the details. The details is and will always be what makes a sim a sim, so thats not really a discussion at all. One guy once told me here on the forums that as long as it is a aircraft in the sky, it is simulating flight, therefor it is a simulator. And boy do I have to disagree with that.
Wow a storm in a teacup. You are over-inflating a statement that did not mean they are “dumbing down” the sim.
He said that they put a limit to some phenomena until they get their modelling better, and they intend to remove the limit as well.
For me that is something that can be appreciated rather than condemned.
The sim has far more serious areas to improve than an extremely rare scenario you shouldn’t encounter in the first place.
You make a very good point here! But the fear of all this or the “storm in the teacup” is not without merit given the condition of some of the areas of the sim, the duration they have been in that condition and seemingly resistance to fix some of the basics, the USR/USER waypoints being a huge one, that destroy your flight plans and disrupt your approaches etc…