Theoretically, if two or more people research to a point near a round number, they'll end up with the same version as long as they research for the same number of hours, so that's a moot point. In other words, I don't see how it makes any difference except for those of us with obsessive compulsive disorders which make things like round number rather cool to us.
No, thats not what I meant. I understand what you saying. But for example: I have a V3.456 hider and you have a V3.456 hider. (we both bought it from dealer), then I researched it with 6 task for 8 hours each and you did 13 task for 11 hours. I would have a V5.678 and you would have V6.987. And after we continue to research odds are we will never have the same version software again.

I hope you understand that...

I suppose with parallel research it is possible that the steps would not advance the same thereby ... Okay, I get it.
Maybe, we should have, in addition to the top scores (power rating, or whatever) there should be an anonymous scoreboard for software versions including the highest version and a mean/average version. That way we can get an idea of what people got but still have to go hunting for it. I dunno, it seems a bit hackneyed, though. It's just a thought.
That way we could get rid of the software dealer or limit the software available to the mean/average version. Versions in the upper end of the bell curve would be either available for a steeply increased price or not at all.
Sort of a compromise, I know, but I'm brain storming.
I kinda like the average but maby make it the average version of the top 100 players or top 200 players...

That would eliminate the drag caused by inactives, but the median value could be used instead and would account for the inactives while keeping the formula very simple. Averaging the top 100, given the span that usually encompasses, would probably also be quite close to the median.