Also there was no specification right from the start and the feature set was very limited, so that everyone started implementing their interpretation of HTML and adding their extensions. Trying to combine and standardize all the things existing in the wild would necessarily be a hard task. Many early extensions were probably also quite ad hoc and even if not they most likely did not take into account interactions with extensions from others.

But in the end this could probably never have happened much differently, future developments and requirements could not be known in any detail and even if nobody would ever have spent a decade to write good specifications for a new technology which might or might not get adopted.