Visually Networked GSX Ground Services

[DaveCT2003]

Umberto,

First, may I thank you for the most recent update to GSX. It is a very nice work, and the added de-icing is a very nice service to have.

I have one question, and respectfully I have one suggestion that I'm confident would be a game changer for this type of addon.


QUESTION:  Do you ever plan to have the deicing of the aircraft fuselage?


SUGGESTION:

The number of online users (VATSIM/IVAO) continues to grow and recently took a leap forward, possibly due to the closing of other networks.

There is a rather easy way to implement a way for fellow GSX users to see the other aircraft's GSX vehicles when logged onto VATSIM or IVAO, without having to go through the VATSIM/IVAO network. This has already been done, very successfully, for runway skid marks left from aircraft landing ahead of you.

Being able to see the vehicles called by another GSX user would bring any and all airports to life, and I dare say it would result in non-GSX customers (including AES users) flocking to GSX! It would be a game changer!

I can provide a working example of how to implement this, using very little resources.

I hope you'll consider this, as it would be revolutionary to the flight sim community!


Thanks for your time.

Dave

[justin48]

+1 on networking

[virtuali]

QUESTION:  Do you ever plan to have the deicing of the aircraft fuselage?

It's not really something that makes any difference. The FSX structural deice command, which is what GSX use to deice the airplane, it's just one, for the whole aircraft, so it won't change compared to how it is now.

There is a rather easy way to implement a way for fellow GSX users to see the other aircraft's GSX vehicles when logged onto VATSIM or IVAO, without having to go through the VATSIM/IVAO network. This has already been done, very successfully, for runway skid marks left from aircraft landing ahead of you.

Static objects are criminally easy to do, because once they have been created, they don't have too many state changes.

GSX objects have a constant flow of data exchange between them and FSX and, in order to have them working in sync, all this data should be transmitted over the network. Animations, for example, runs at 20-30 fps, and one vehicle might have several of them running in parallel, and there might be more than one animated vehicle on screen at any given time.

It's not "easy" as you think, and some members in our team worked in the past on AAA  online multiplayer console games, and I can tell you that fast animations over a network with unpredictable speed are a NIGHTMARE to program, you must compensate and anticipate for lag, using predictive methods, and still it's a real mess, even with game engines far more advanced than FSX, without even mentioning that, BECAUSE users are flying online with any number of other networked addons, we couldn't even be sure of how much data is transmitted, because we can't even be sure to own all the bandwidth.

Being able to see the vehicles called by another GSX user would bring any and all airports to life, and I dare say it would result in non-GSX customers (including AES users) flocking to GSX! It would be a game changer!

The very fact that GSX works with any airport (even those whose developers are *opposed* to AES) without requiring to purchase airports individually, has already achieved that purpose. Most of AES users have bought GSX anyway.

I hope you'll consider this, as it would be revolutionary to the flight sim community!

There are too many other features, all planned of course, that will greatly improve GSX, before we would start doing full network support like that, and those will benefit *everybody* not just those flying on IVAO/VATSIM

Without even mentioning, that with the upcoming FSX re-launch on Steam by DTG, they'll going to put back the default online flying system which will both REDUCE the effect you mentioned of "users flocking to VATSIM/IVAO because Gamespy is dead", and would also be ANOTHER entirely network that we would eventually have to support.

[DaveCT2003]

Umberto,

Thanks very much for your reply, it is sincerely appreciated and I understand exactly what you're describing.

Respectfully, I should like to offer a different way to skin this cat...

Instead of networking the exact vehicles called and what and when they are doing for the other aircraft, simply have them run from off the other aircraft based on a time offset from one's own vehicles.  Also have a selectable on/off for this feature as well as a selectable range that one would see vehicles, so people could adjust this based upon their custom frame rates.

The positions of the other aircraft, as well as aircraft types can be pulled directly using servinfo for VATSIM and IVAO.

 
It would be a most excellent feature, attracting even more people!


Thanks very much for your time.


Dave

[virtuali]

Instead of networking the exact vehicles called and what and when they are doing for the other aircraft, simply have them run from off the other aircraft based on a time offset from one's own vehicles.  Also have a selectable on/off for this feature as well as a selectable range that one would see vehicles, so people could adjust this based upon their custom frame rates.

No, they must synchronized correctly, because there's a LOT of random variations that happens when you see a service, in order not to have them looking robotic, almost every small time pauses between service steps are randomized, the vehicle starting times are randomized, their departure parking is randomized too, the time delay when there's two vehicles operating is randomize, even the speed of the animations is slightly different.  And of course, when there are multiple matching candidates with the same score, they are selected randomly, both as operators, but even in different variations of multiple matching models.

Under these conditions, the result would be similar as if the networked users would just decide to call GSX services locally on their own, and that would already work without even having to use a network. The end result would be the same, since the services would look different anyway, once they start.

And that without even considering that, if there are many users on the same airport, and each one would be able to see each other vehicles servicing, we'll end up in the same issues that is the main reason why we are not considering to add GSX service to AI airplane: the impact on fps would be very bad, and we are not even sure if FSX could even HANDLE that many objects at the same time, because there's not much else we can stuff in the already very limited memory space we can use with 32 bit.

MAYBE, IF and WHEN P3D will be released in 64 bit, we might START considering something like this. Obviously for AI FIRST and THEN for network flying, because AI will obviously benefit 100% of users, while not everybody flies online.

[Sprose]

+1 for networking the GSX!

[virtuali]

+1 for networking the GSX!

Have you actually read the whole explanation why it won't happen ?

Note that what has been suggested here, is to have users online at the same airport to see each other's GSX vehicles. So, for example, 10 users at an airport will mean 10x the amount of GSX vehicles to be seen, without even mentioning that FSX doesn't really handle well many instances of objects each one with its own set of private variables. There IS a reason why GSX doesn't support more than 2 objects of the same class, so it won't be any easier if the number would increase tenfold. P3D is entirely different, but we are not yet prepared to do a P3D-only product, surely not before P3D will be 64 bit.

And this is assuming we would put a cap of 10 users maximum, which will obviously result in an uproar from online user that they wouldn't be able to organize large fly-ins, which kind of defeats the purpose of having a network in the first place. Without a cap, we couldn't even sure how much data and memory would be needed, we don't even know if FSX CAN handle that many animations at once, because nobody tried that.

And I'm not referring to the issue of transmitting data over the network: just the fact that instead of ONE airplane serviced, you'll see MANY of them at the same time on your local system (those of the others), it's enough of a problem on its own, without even being to start discussing the network issues.

Regardless how many "+" it will get, it's just too much for the current agin FSX engine to handle, and we won't waste our (very limited) resources to do something that companies hundreds of times larger than us, with budgets proportionally larger, have managed to get it right without years of bugfixing, and they HAVE control of the game engine, we don't even have that luxury.

So, please, be realistic, the FSX environment is already WAY overstretched, far beyond it was designed to do, and before we'll considering wasting resources on networking, we'll prefer adding new animations, or new stuff like visible passengers, or support for AI airplanes, because those will work for everybody, not just those flying online.