You have to look at what each version offers and see if there is anything in that version you may require (such as script commands for example). As each version rolls out, additional script commands are added so that helps to make a difference. Each version of FPSC is unique in its own way. While we have each version running in a VM to test, they all have their strengths and weaknesses.
V1.18 introduced water, variable system, and additional scripting features. V1.19 saw the integration of WASP mod commands and additional features for culling, weapons, camera commands, etc. Also added was scriptable post effect shaders. V1.20 added a few more features and had a lot of fixes to be done. I do not have a compiled list of what exactly was added.
The point is, each one has its pros and cons. If you do not need water, then v1.17 is more than sufficient. If you want to use the variable system, then v1.18 minimum is required. Each version runs differently than the next- you may experience some fps drop or increase depending on what you are doing. Only you will know what feels right for your work.
I would not recommend going lower than v1.17 as v1.16 and lower could not build more than 5 levels. One thing to keep in mind- when you pick your version and start a development,
stay in that version. I cannot begin to count how many "issues" have been reported because people are working on a development in one version and then decide to switch. While you can switch
technically, you will have undesired results with some media or scripts, or saved data. That is because each version has changes in the save/load routines that are the main
cause of these issues, among other things.
Quote: "If Blackfox shows up here, he might be able to tell you more about 1.17. They mod it and I'm sure he knows more about its "gears" than I do.
"
There were some key reasons why we chose to remain in v1.17. First- the GUI X9 garbage. While I can understand the intent, it was totally useless. The documentation was vague at best, half the features did not work properly. I never saw one development use it. We attempted to use it and it was more of a pain to get it to work. So rather than trying to remove it from the source (it was introduced in v1.18 and it is everywhere in the source), we picked v1.17 because it was not introduced in this version. Second- ragdoll was introduced. Third- the start of DarkAI was introduced. We managed to expand on it ourselves but it was nice to have a start point. Fourth- other fixes from the older versions were made, including the ability to build games beyond 5 levels. The big one for us- this version gave us the *best* fps level compared to the rest. We do run on two laptops and two desktops and v1.17 seems to like the four machines. Even with porting water features, scriptable shaders, video textures, and other features did not affect our performance. Plus it was easier for me to rewrite the entire source code, organize it properly and add features we wanted with ease. We also do not seem to get the "collision" or other issues that have been reported in v1.19+. Mind you, I am the only one that codes our source so that makes a big difference right there. Multiple people coding can cause issues if they have their own style, etc.
In the end, it is developer choice.
There's no problem that can't be solved without applying a little scripting.