Mostly it's due to requiring a consensus on things.
With XHTML for example, Microsoft didn't provide support for it in their web browser. So even though many good and interesting things were possible with XHTML, they became stalled in regard to general public web browsers.
As a result of that, XHTML was forced in to a strangely munged mess, due to needing to be interpreted as if it were HTML instead.
So a consensus is important to achieve, and as a result is a time-consuming activity.