I think that if hypercard had been the engine of the web instead of html/css, we would have had the current state of the web (ie. with client-side dynamic applications) a lot sooner.
Wether Hypertalk would have been a better choice than javascript to program the web, this is another story :)
For the curious, the Hypertalk entry in wikipedia contains a lot of code examples: https://en.wikipedia.org/wiki/HyperTalk
But is this a good thing? After all, the web started and succeeded as a simple platform for passive document-browsing rooted in markup (SGML) technologies. There never was a need for a new general-purpose app platform (because that existed, and still exists, in the form of operating systems). Arguably, seeing the web as primarly an app delivery platform is what ruined the web even.
As to HyperTalk, it's an odd language for sure (did a HyperCard app for budget control in an agency around 1992), but listening to Bill Atkinson is still very inspiring.
There's a whole stream of forgotten technologies (server push images? Image maps? Html layers?) that show this.
Don't believe this rose tinted view of the old web. There was lots wrong with it and the current web stack is much, much better.
(Also SGML was never deployed as part of the web)