On 3/16/04 at 9:50 PM Curtis L. Olson wrote:

>In the next couple weeks I'd like to start moving seriously towards our 
>next release.  There are probably many things that could stand to be 
>tweaked and polished but two related items jump out at me tonight.
>
>1. I'm growing less enthused with our aircraft alias naming system.  I 
>don't mind that we have the capability, but it becomes annoying to have 
>8 names for the same aircraft, even 2 names for the same aircraft.  What 
>would people say to nuking all the alias entries for the upcoming 
>release?  Now that we have the "fgrun" launcher, name length or 
>complexity is not really an issue.  Even for us hard core unix folks who 
>refuse to give up the command line, we can learn to copy/paste to avoid 
>unnecessary typing.  I think getting rid of the aliases would make 
>aircraft browsing and selection a lot less mystifying to new comers.  
>We've got 20 (!) varients of the C172.  Maybe for this release it would 
>be nice to just have one or two???
>

I agree with the fact that there is a problem with the multiple names, but
not with your proposed solution.  Please don't ditch the aliases.  Or to be
more specific, please don't ditch the short names.  Typing --aircraft=737
is so much better than --aircraft=737-jsbsim, and ditto for most of the
planes.

IMHO, the solution lies in the presentation of the available aircraft in
--show-aircraft.  At the moment, for instance, for the cub we have 4 lines
for one aircraft:

   j3cub                        Alias for j3cub-3d-yasim.
   j3cub-3d                     Alias for j3cub-3d-yasim.
   j3cub-3d-yasim               Piper Cub J3 Trainer (YASim)
   j3cub-yasim                  Alias for j3cub-3d-yasim.

This should simply be:

j3cub            Piper Cub J3 Trainer (YASim)

No-one (outside of the developers) need know if 'j3cub' is an alias or not.

As far as I can see, there are two general solutions.  Either have one
alias for each unique model (in this case ditch the -3d and -yasim aliases)
and show the alias, but not the full name, in --show-aircraft, together
with the proper description, as in the line above.  That's probably the
best short term, and possibly long-term, option, IMHO.  Or, ditch the long
names, ie. in this instance the full name would be j3cub, and then pull
details of available 3d/2d cockpit options, and fdm options, from the xml
if required.  This would work for the accepted 'best' version of models
with more than one implementation, so for instance 'C172' would be the name
of the default C172, wheras C172-yasim would be the non-default, which
might not make it into the release anyway judging by your next comments.
Then, one could type --aircraft=C172 --2d to try and get a 2d cockpit if
available (would fall back to default if not), and likewise --aircraft=C172
--3d (ditto for fallback), and a lot of names would become superfluous.

Anyway, the above would roughly cut the size of --show-aircraft by half,
since most planes have

name   Alias to description
name-fdm  description

and we'd be left with

name description

The C172 and C310 would still give probs though, since there *are* multiple
implementations of these.

   c172                         Alias for c172p.
   c172-3d                      Alias for c172p-3d.
   c172-3d-yasim                Alias for c172r-3d-yasim
   c172-610x                    Alias for jsbsim version.
   c172-610x-jsbsim             Cessna 172 with a full screen, hi-res panel
   c172-610x-null               Cessna 172 with a full screen, hi-res panel
   c172-ifr                     Cessna 172 in IFR configuration
   c172-larcsim                 Cessna 172 (LaRCsim, 2D panel).
   c172-yasim                   Alias for c172r-yasim
   c172p                        Alias for c172p-jsbsim.
   c172p-3d                     Alias for c172p-3d-jsbsim.
   c172p-3d-jsbsim              Cessna 172P (JSBSim, 3D cockpit)
   c172p-jsbsim                 Cessna 172P (JSBSim, 2D panel).
   c172r                        Alias for c172r-jsbsim.
   c172r-3d                     Alias for c172r-3d-jsbsim.
   c172r-3d-jsbsim              Cessna 172R (JSBSim, 3D cockpit)
   c172r-3d-yasim               Cessna 172R (YASim, 3D cockpit)
   c172r-jsbsim                 Cessna 172R (JSBSim, 2D panel).
   c172r-yasim                  Cessna 172R (YASim, 2D panel)
   c172x                        Cessna 172 flight dynamics testbed

Some of these could get ditched from a release.  The rest could lose the
long names, to give

c172            c172p (jsbsim)
c172-610x       c172p with a full-screen, hi-res panel (jsbsim)
c172-ifr                c172p in IFR configuration
c172-yasim      c172r (yasim)
c172-larcsim    c172 (larcsim)

That's much more manageable, still gives 3 separate jsbsim (default) c172
configurations, and 3 different fdms.

On an (almost) totally unrelated note, I think it would be a good idea to
test unknown options against aircraft names, so that, for instance,
bin/fgfs --T38 would work to bring up the T38.  Would patches to add this
be accepted?



>2. We have a *lot* of aircraft in the base package.  Some of these are 
>really nice, some of these are not even close to basic functionality.  
>Most are probalby best considered "works in progress."  (Now this is 
>perfectly fine, and is what CVS is for.)  However, for the next release 
>I would like to just include a subset of the available aircraft, picking 
>and choosing the best or most interesting ones.
>

Agreed.  It might be worth zipping up an 'experimental aircraft' package of
the rest that drops staight into the base for interested punters without
CVS.

>I think it's worth including the heavies (737, A320, 747) even though 
>their cockpits aren't finished.
>

Agreed.

>We definitely want the light civilian aircraft just as the Cub (which is 
>one of the few aircraft that are legitimately finished) :-) as well as a 
>C172 (C182?) and maybe include the C310 here, even though ours has navy 
>markings?  And definitely the pa28-161.  

Absolutely!

>It would be nice to get a 3d 
>mixture control on the default C172 some day.
>
>There are some vintage aircraft we should include such as the Wright 
>Flyer and ComperSwift and maybe the dc3 although I have no idea what's 
>going on with the cockpit there.
>
>The P51 and the Hunter should definitely be included, those are ones 
>that also could be considered "finished".
>
>Lee has built a nice fleet of various aircraft which are worth including 
>such as the AN-225 (huge russian transport), a b-52, a seahawk, a tsr2 
>and a yf23.  All of these fly well, have beautiful external 3d models, 
>are very well animated including the gear and suspention, but all lack 
>3d cockpits and instead have a functional, but non-realistic 2d cockpit.
>

I don't think that lack of a 3d cockpit should be a show-stopper, where a
2d alternative is provided.

>The a4 is nice, although I don't know how complete the cockpit is???  
>The J22 and F16 both show promise.  We should also include the X15, but 
>that takes some startup parameters to get flying.
>
>And the Ornithopter should definitely be included.  And we should toss 
>the ufo in as well.  I'll list the remaining aircraft below.  These I've 
>had less experience with or non-optimal results when I have tried them.
>
>What is the status of the helicopters?  They've seemed very crude when 
>I've looked at them.  I don't mean to be anti-helicopter, but if we are 
>trying to cull some of the less nice stuff out of the official release, 
>I'm not sure in their current form they would make the cut.

I think that the most developed heli (bo105 ?) should make the cut.  Some
user feedback might stimulate interest and further development, and I was
under the impression that initial feedback on the flight model was quite
favourable?

Cheers - Dave



_______________________________________________
Flightgear-devel mailing list
[EMAIL PROTECTED]
http://mail.flightgear.org/mailman/listinfo/flightgear-devel

Reply via email to