By Ross Carlson 887155
#495414
Robert Bressert 1139527 wrote:Ross, what about pilots who fly on IVAO too. They have the IVAO-MTL models in their SimObjectsPath. Will 2.0 be programmed to exclude them?


The IVAO models will not be in the model database, so it won't use them for automatic model matching. Users could still make custom model matching rule sets to use the IVAO models if they want ... I'm not going to go out of my way to stop that from happening.
By Pierre Ferran 1259058
#495415 Ross,

I think the changes you are proposing are really good, we need a more "all in one" solution.

A few suggestions on that topic:
- Make sure commercial models (PMDG, Aerosoft) are not spawned by AI, you don't want 14 PMDG T7 spawning during CTP...
- Maybe broadcast the registration for the model of the user, and if the same registration is available on the other end, display that. That would resolve the winglet problem for instance.
By Ross Carlson 887155
#495416
Pierre Ferran 1259058 wrote: - Make sure commercial models (PMDG, Aerosoft) are not spawned by AI, you don't want 14 PMDG T7 spawning during CTP...


Yeah, the model database used for automatic matching will not have high resolution models in it.

Pierre Ferran 1259058 wrote: - Maybe broadcast the registration for the model of the user, and if the same registration is available on the other end, display that. That would resolve the winglet problem for instance.


Not sure what you mean by "registration" here. To me, registration means tail number ... are you talking about the ICAO aircraft type? If so, the user enters that when they connection vPilot to the network.
By Pierre Ferran 1259058
#495418
Ross Carlson 887155 wrote:Not sure what you mean by "registration" here. To me, registration means tail number ... are you talking about the ICAO aircraft type? If so, the user enters that when they connection vPilot to the network.


Yes, I do mean the "tail number" by aircraft registration, N172BX, F-HEPF, etc...
By Ross Carlson 887155
#495433
Pierre Ferran 1259058 wrote:
Ross Carlson 887155 wrote:Not sure what you mean by "registration" here. To me, registration means tail number ... are you talking about the ICAO aircraft type? If so, the user enters that when they connection vPilot to the network.


Yes, I do mean the "tail number" by aircraft registration, N172BX, F-HEPF, etc...


In that case, your two suggestions seem to be at odds with each other. For example, if I'm flying a PMDG triple 7 with registration number N12345, then other pilots that I encounter that have the same PMDG T7 livery installed aren't going to want to have vPilot use that model since it would be hard on the frame rates.
By Pierre Ferran 1259058
#495434
Ross Carlson 887155 wrote:In that case, your two suggestions seem to be at odds with each other. For example, if I'm flying a PMDG triple 7 with registration number N12345, then other pilots that I encounter that have the same PMDG T7 livery installed aren't going to want to have vPilot use that model since it would be hard on the frame rates.


To clarify, i'm not talking about an absolute selection of the model based on the registration. But say a user has got a AI traffic package of all aircrafts in an Airline, he's got all different aircraft with different registrations, each might have different options like winglets, etc...

Now, another user is flying his high poly, hight end Aerosoft Airbus of that same airline, the registration is then send on the network, and used to see if any low definition/AI traffic that the user has installed has got the same registration and airline. If so, the correct aircraft with the same registration is shown.
By Ross Carlson 887155
#495439
Pierre Ferran 1259058 wrote:But say a user has got a AI traffic package of all aircrafts in an Airline, he's got all different aircraft with different registrations, each might have different options like winglets, etc...


Every AI package I've seen doesn't work that way. They just have one model for each type code, and sometimes for each livery. They don't have a separate model for each registration.

Maybe this is something VA's do if they have their own custom model set? I don't fly for any VA so I wouldn't necessarily know if this is common practice. And it would have to be a common practice in order to warrant a network protocol update like this.
By Pierre Ferran 1259058
#495444
Ross Carlson 887155 wrote:Every AI package I've seen doesn't work that way. They just have one model for each type code, and sometimes for each livery. They don't have a separate model for each registration.


Exactly, it'd just be an optional selection parameter. Certain livery in AI packages are only associated to a certain registration, provided that the AI package implements it correctly, since usually one aircraft in the whole airline fleet is gonna have that special livery.

You could even use the registration sent from the client to map in your custom database the aircraft characteristics. So if the AI package doesn't contain one model per registration, just find the best one based on what you know about that registration in your system.

What i'm basically trying to say is instead of doing like FSINN which used a unique model string, use the registration of the AC, which is already unique to do that.

I don't know if I'm making any sense, if not i'll sleep it off and explain it better tomorrow :mrgreen:
By Ross Carlson 887155
#495451
Pierre Ferran 1259058 wrote:I don't know if I'm making any sense, if not i'll sleep it off and explain it better tomorrow :mrgreen:


I understand the suggestion completely. I think we just differ on whether or not a registration number match would occur frequently enough for it to be worth the development effort.
By Nick Botica 999991
#495461 Quite often I see incorrect ICAO codes such as the case in an earlier reply; B73H, which is not an ICAO code but in fact an IATA.

While I think it's a good idea to have a free form field (4 characters), perhaps an autocorrect type box could be introduced.
For example the user starts typing B73 and several valid options appear under it B733, B734, B735 etc.
This way you can encourage people to select a valid code while leaving the user the flexibility to enter a code that isn't in the database (although could this be possible, apart from fictional aircraft where most people wouldn't have the aircraft installed anyway)

Cheers, Nick
By Ross Carlson 887155
#495469
Ned Hamilton 1215492 wrote:I know your deadline has already passed and maybe vPilot is not the place to do this:
But since the parse errors are still coming. Is it possible to have an option in vPilot to ignore (not display) parse error messages?


That's already been fixed in the current 1.1 beta version. I'll be releasing that fix in the Stable update channel soon.
By Ross Carlson 887155
#495471
Nick Botica 999991 wrote:While I think it's a good idea to have a free form field (4 characters), perhaps an autocorrect type box could be introduced.
For example the user starts typing B73 and several valid options appear under it B733, B734, B735 etc.
This way you can encourage people to select a valid code while leaving the user the flexibility to enter a code that isn't in the database (although could this be possible, apart from fictional aircraft where most people wouldn't have the aircraft installed anyway)


I mentioned above that I'm planning to show a popup warning if the type code you enter isn't a known valid ICAO code, but still allowing the user to proceed with the unknown code.

However, I like the idea of a suggestion list appearing as you type. To be clear, it would not be an autocorrect function ... it would not change what you type in. It would just show a list of matching codes that you could click on or use the arrow keys to select. You can still just ignore the suggestions and continue with whatever you type in. It would still show the popup warning if you type in something that is not a valid ICAO code. Those two things together should go a long way towards eliminating invalid codes.
By Nick Botica 999991
#495482
Ross Carlson 887155 wrote:However, I like the idea of a suggestion list appearing as you type. To be clear, it would not be an autocorrect function ... it would not change what you type in. It would just show a list of matching codes that you could click on or use the arrow keys to select. .


Ah yes, wrong terminology on my part... more like an autocomplete :D