Re:Doug Engelbart’s design for knowledge-based organizations...
by zozbot234 on Saturday September 03, 2022 @03:45AM
(#32699676)
> The main limitation here is there's only one possible tree, and it's the privileged view.
Mainly because inferring a different "view" over gigabytes or even terabytes of data (current collections of publicly available "general purpose" datasets can run in the terabytes) is really, really hard and computationally expensive.
And our current data models don't even include satisfactory ways of representing the kind of "common sense" data that would be reflected in a HN comment, an e-mail thread or really any kind of textual source. You can kinda get there by marking "named entities" of sorts but that's the tip of the iceberg, all the structure and semantics is still missing. It's something that's very much being worked on, and has been for decades (with "frame structures" and the like). But it's still hard.
Mainly because inferring a different "view" over gigabytes or even terabytes of data (current collections of publicly available "general purpose" datasets can run in the terabytes) is really, really hard and computationally expensive.
And our current data models don't even include satisfactory ways of representing the kind of "common sense" data that would be reflected in a HN comment, an e-mail thread or really any kind of textual source. You can kinda get there by marking "named entities" of sorts but that's the tip of the iceberg, all the structure and semantics is still missing. It's something that's very much being worked on, and has been for decades (with "frame structures" and the like). But it's still hard.