Saturday, February 24, 2007

Which sim(s) to develop for?

Back in October not long after FSX was released, with some grumbles around the forums in respect to performance issues in FSX and many users saying they would remain with FS9, I posted a poll at FSDeveloper.com asking developers whether they intend still support FS9.

Four months down the track and people are still voting, with 156 developers voting to date. The results are interesting and percentages have largely settled down during that last 2 months.

44.23% of developers intend to develop for FSX only.
23.72% intend to develop for both FSX and FS9.
21.79% FSX and Fs9 with FS9 being for a limited time.
10.26% FS9 only.

As developers complete their existing FS9 projects I expect the majority will soon be developing for FSX only, especially if FSX SP1 delivers the goods with performance. It will be interesting to see if intentions change in the 23.72% who intend to develop for both sims and the 10.26% FS9 only. Maybe another poll in a couple of months?

Developing for FSX is the way forward. To date, I have no real grumbles with the SDK as far as aircraft creation goes. Rotational animations such as wheels and props could certainly be improved in the sim. Some developers appear to be unhappy with the implementation of virtual cockpit lighting in FSX though appear to find solutions. I will encounter the challenge of cockpit lighting in my next project, the Sopwith Snipe. Given that I use 3D gauges and lighting in WW1 aircraft is somewhat primitive compared to more modern aircraft, lighting in the Snipe was always going to be a challenge. My experience of the FSX export tools for 3DS Max is that they are robust and produce excellent results.... in seconds. Compare that to the limitations in the FS9 makemdl utility, vertex count limitations, model resolution, and the time it takes to compile an aircraft using makemdl.

Having completed the FS9 version of the Bristol F2B, in addition to the FSX version, I'm not exactly enthusiastic about creating further FS9 aircraft. Though I have not discounted the possibilty completely. I guess, in some respects I have benefited by being a relative 'noob' to FS design, coming from CFS3. Many FS developers appear to be set in their ways and have experienced difficulties moving forward. Maybe I'll be the same when it comes to FS 11? ..... LOL.

I always believe that it's best to design an aircraft for a specific sim. FSX & CFS3 share the ability to create cockpit models of a higher resolution thus one can use the same base mesh. FSX offers a superior range of material properties to both CFS3 and FS9. Thus one could possibly use the same base mesh for both sims. However to make effective use of textures and uv maps a somewhat different approach would be required for FSX to CFS3. Likewise a mesh for a FS9 would have to be reworked significantly from one which took full advantage of FSX (and CFS3).
As one of the posters to the poll thread points out, users will blame designers when the FS9 version does not look as good as the FSX version.

The Sopwith Snipe was orginally designed with CFS3 in mind, and also FSX, even though FSX was an unknown quantity at the time. The exterior model of the Snipe was UV mapped with CFS3 in mind. I use 2x1024 texture maps in my CFS3 WW1 models. The interior is only partially mapped at this time and will be completed with FSX as the primary design goal. It will be interesting to see what additional work is involved in the CFS3 version e.g. damage modeling and maybe re-jigging uvs. Unfortunately it's unlikely that I will produce a FS9 version of the Snipe. The Snipe may also be my final project for CFS3..... we'll see.

I am looking at the download stats of the different versions of the Bristol F2B (part1). In just over 3 weeks the FSX version has been downloaded 5,181 times (probably a higher number due to Simviation stats being updated about once a week) . The FS9 version which has been available for download only 6 days has acheived 1,935 downloads but has now slowed to just to 2-3 per day. So far it's a 3:1 ratio in favour of FSX. I expect that to change to further favour FSX, as time passes. Whether these ratios would apply to a payware release I do not know. The real surprise for me were the figures for the CFS3 version - 6,105 over 10+ month period.... not bad for a CFS3 WW1 aircraft hidden away in the web.

I resume work on the Snipe in a couple of weeks (it will be freeware). After that I will be reviewing my options as to projects though I tend to think I will focus on FSX, though with Train Simulator on the horizon and possibly other future products from Microsoft Aces who knows? Life will be interesting anyway. :)

cheers

Rob

No comments: