Re: [Flightgear-devel] FSWeekend 2012...

[Renk T. <tho...@jy...>]

Thanks for the responses, I think I understand the issue much better.

> The eye catcher is not the gem, but things you see, and notice, from a distance,
> like a huge flag with "free" writen on it.

I guess it'd just be a matter of setting things up. Sure, on a laptop things tend to get lost. But just a simple example - on googling 'FSX sunset' images I find that FSX doesn't seem to have anywhere Mie scattering on clouds. We do, and it looks pretty darn impressive when clouds illuminated from behind  light up in low light. So all it'd take is to get a weather situation in which this happens set up and project it on a sufficiently large screen, and voila - here's a potential eye catcher.

Of course, overall we're still going to lose that game in the long run though. Almost by definition, a commercial product is going to focus all resources on the things that people notice first. Say to get a good FDM takes five full working days, to get the last 10% takes another 100 days. But only a tiny fraction of users will notice the last 10%, and only after some trying out, but they make your product 20 times more expensive - and thus a commercial product will develop things on average only to a break-even point so that it makes most users happy at minimum effort. 

And that's where Flightgear is very different, because now and then people put an insane amount of work into something that will never be noticed by vast majority of users - because for someone a particular detail is important and so it gets lots of attention, way more than would ever be economically viable.

We just don't care too much about appearances because the problems are so simple :-) - I think for instance consistently removing unrated, non-functional or unfinished airplanes from the main download page would go quite a way in making Flightgear look more professional in the eye of the casual user who is just curious and wats to try -  if we ever wanted to do so - but we're not going to do it, because it's a boring administrative problem, and nobody likes those when he can tackle an interesting problem (me not being the exception here...).

And we're losing out on resources. I don't know how much manpower a commercial sim typically has, I think I remember MS Flight had a team of 20 or so. That looks about like the number of active major contributors here, except that if my coding time is representative, I get about 1/10 to 1/20 of the workload of a full-timer done, so we're losing by more than an order of magnitude. It doesn't always help us that with an open source product we have a potentially unlimited  number of contributors, since for some tasks you need a highly coordinated workforce.  

If I think what I would need in order to make the sky visuals better than any FSX screenshots I've seen, it's mainly down to things like raw data and image processing - e.g. I lack aerial images which tell me what I am aiming at. For instance, my worst problem is - how does a high altitude scene look like when the sun is at the horizon and I look *away* from the sun? You can simply forget googling images for that, because everyone points the camera the other way. I've twice sort of seen it by being on the right side of the plane during a transatlantic, but it's not the same as having an image which you can use to sample rgb values. I also know some of the cloud textures could benefit from a better extraction procedure and better raw images - if I had a graphic artist which can do these things, we'd be good, but if it's down to my 12 year old digicam pointing at the sky and my GIMP skills, there's a limit to the quality of the final product. Lots of things don't require coding but just patient testing of parameter values - I remember getting a different (rgb) triplet for the base sky color from someone, which made things look much better - but that simple color info is  worth for me several days of tests. If I had an artist adjusting parameters to get scenes right, we'd be in much better shape.

But even with a potentially infinite pool of contributors, we don't get people to do what we need when we need it - because volunteers work on what they want when they want. So, also in terms of resources, commercial products have an edge here. 

I think generically, we can only win in areas where people are really obsessed over details, and the fact that FGFS is superior in that particular detail will never be eye-catching. I also think if we really wanted, we could do a lot to make a more professional appearance to new users - well defined standards on the download page, a consistently designed GUI maintained by a co-ordinated task force rather than everyone adding as he thinks fit, standards on documentation updates, all the nasty things,...

Well, that's my 2 cents at least...

* Thorsten