rec.autos.simulators

I mean issues like tires, drafting, variable weather,
better FF and AI, etc., and not new types of cars or
series necessarily.  In other words what kinds of stuff
still bears to be improved upon (or even invented),  
and what has been pretty much optimized?  What will
we see 5-10 years down the road? [Pun kind of
intended...]

When I watch a NASCAR telecast I get the impression
that a lot more is going on than there is in either of our
NASCAR titles, as teams feverishly try to tweak the car
back to optimum during a race. [Not that you'll get a
lot of this technical info from the announcers, just a
bunch of stuff like "Gordon is tight in the middle of the
turn, loose on exit"...].

        John DiFool

*** behaves in a very complex way, and with temps (ambient, surface &
tyre), surface type & so on so varied from one race format & track to
another there's probably quite a bit of mileage left in the perfect tyre
model. Most if the rest of a car (at least a racing car with rose joints) is
pretty straightforward mathematically speaking so there probably isn't much
that can be improved, just tweaks for things like chassis flex etc.

The main improvement I can see coming over the next few years is the damage
model, both from mechanical abuse (do a donut & pop a driveshaft etc) and
panel/body damage. There's an awful lot of variables here too & I'd like to
see a move away from look-up tables to modelled wear & damage, a bit like
Viper Racing's ace body damage but taking into account the varying strengths
of different parts of the car.

Other than that it's just a matter of simulating different racing formats
but the problem there is what sells gets made ;-)

Malc.

(this is a topic I've covered about once a year for the last 4 or5 years)
   An online pitcrew sim.  7 or 8 guys sit around and do nothing for 30
minutes, then they all jump over the wall, change the tires and jump back
over the wall.  One player, the crewchief, makes the calls for tire
pressure, wedge adjustment etc.  A Doom3 engine or Halflife would make an
excellant base for such a sim.  You could even meld in a real racing sim,
so that the driver needs to recruit the country's best online pitcrew.  
Crews could be like hired professionals,  pitting for drivers in various
leagues all over the world.  
  you'd need voice comms,  you'd need telemetry, you'd need a way to
interact with the racing sim itself...sheesh...a couple of weeks coding
right???

dave henrie
ps...due to the limitations F1 always puts on multiplayer, you could have
a bigger crew..but they could only hotseat compete against another team
at the same site.
dh

Here are a few that jump to mind.

How about non-linear client vehicle prediction for multiplayer racing.
AFAIK, every multiplayer sim out there defines client car paths as a
series of short straight lines.  This stepped car paths leads to warp
collisions and whatnot.

Papy has yet to give us dual rate shocks.

Voice recognition and intelligent pit crew answers.
You ask the computer, "Has Eldred pitted yet?"
The team boss replies "no."

Dynamic marbles, oil, and other junk on road surface.

-Greg

  don't they already have corvettes?

<teeehheeee>

dave henrie

Better FF.. better wheels.
Wheels that turn in a 1:1 ratio that the car's wheel would turn.. and be
totally linear.

Mike
http://mikebeauchamp.com

As well as tyre behaviour, I'm sure there's still a *lot* of work to be
done and CPU cycles to be absorbed on advanced aerodynamics, particularly
in complex situations like drafting.

However it may be that we're increasingly close to the point where
additional simulation can't translate well through our limited interface
to the computer. The law of diminishing returns may have kicked in for
sim developers... why bother putting in the extra programming & debugging
hours when most folk probably won't notice or feel the difference?

Actually for me that started with GPL. It's not perfect, as we all know,
but I can still imagine having it on my HD in 20 years' time. Whether I'll
be able to find it among the 24 squillion terrabytes of other accumulated
detritus is a different matter.

Still, I hope developers keep trying to push the boundaries. Sometimes you
don't know a threshold is there until someone crosses it.

Andrew McP

The whole simracing world is slowly closing the gap between their physics
and graphical engine.  The transition is long, and it'll take a while.
Around 10 years (makes sense).  Someday everything on screen, will be a real
entity and not be a bunch of polygons together, with a hidden set of
parameters for the physics engine to take in consideration.

The oil a car will let on the track, will run down it if it's banked.  A
fence will be a fence, not just a transparent polygon with a texture on it.
It'll be deformable.  Same goes for any part of a car.  They will become
more and more complex.

The only thing I hate of the whole thing, is the lack of cooperation, or
open-standardness between companies.  Heck, FIVE years ago a company like
Motorsims created the most incredible tracks I've ever seen.  Not even today
companies have devoted that much attention to them.  All that work is
wasted.  Developpers spend TONS of money recreated existing tracks.  How
many ***ing version of Laguna exist?  Now if Laguna Seca would create their
own perfect model and sell it to companies, it would always be perfect.  Or
perhaps a company that would sell track information to other developpers.

--
-- Fran?ois Mnard <ymenard>
-- This announcement is brought to you by the Shimago-Dominguez
Corporation - helping America into the New World...

That's where open-sourced stuff or a large SDK-type tool would give you the
advantage.  Weight in the millions of hours in development, why do ALL
companies waste their time redoing what others have already done...  Really
it's crazy the work they waste modeling tracks and cars when a set of
standards would make it possible to seriously advance the simulation in
other aspects while still offering the same level of quality in the final
product.

In other domains, they never do that.  Movie directors don't have to
re-invent the camera each time they make a movie.  Same for about every
other form of entertainment.

--
-- Fran?ois Mnard <ymenard>
-- This announcement is brought to you by the Shimago-Dominguez
Corporation - helping America into the New World...

Bernie Ecclestone?

Papy...gone.

PS2/Logitech already has that: 2-1/2 turns lock-to-lock.

    Not that it's to do with the guts of sims themselves, but the next big
step for me will be when you have a headset like the Track-IR but the
headset will have 3D sound in the ears and a pair of small monitors over
each eye - now that will be total immersion for all computer games! :) The
peripheral vision problem will be a thing of the past!

  Lots of good new ideas posted by others.  Stuff that still bears to be
improved (especially since some of it has taken steps backwards).
  Better/more accurate track modeling, better online collision/detection
code, better damage models, better physics, especially drafting.
  Of these four, the latter three have decreased in quality over the years,
in terms of Nascar sims.

That watchamacallit to do with being able to emulate the bumps on the
road.  Yeah,  I forget the name and I'm pressed for time to look it up.
<BG> Unfortunately, we'll need sims that run at 1000fps I'd guess.  :)
--
Pete Ives
Remove All_stRESS before sending me an email