YES. This is how proper standards are made if you give a crap about proper
engineering.
You don't make the standard first then make a product around it and expect
it to work properly. It is going to be nearly impossible for you to
properly anticipate and take into account the full scope of the problem at
hand. What you'll end up doing is putting a lot of effort into focusing on
things that are going to end up being minor issues and ignoring things that
are going to be major issues.
By having multiple implementations of a standard in draft form and then
being willing to modify and improve the standard multiple times before the
final draft is published then it allows you to identify the problem much
more accurately.
This is why things like HTML5 take so freaking long.