COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	hypert[e88,jmc]		Comments on ACM Communications special issue on hypertext
C00004 ENDMK
C⊗;
hypert[e88,jmc]		Comments on ACM Communications special issue on hypertext

p827 - Making the node a fixed size is fascism.  Bringing another
document into the hypertext system requires that someone chop it
up.

p828 - Is KMS just an overgrown text editor?  That's what I want.

p828 - I need an explanation of what links are.

p828 - E's link destinations can be single lines.

829 - rigid choice of time multiplexing  

My own ambitions for hypertext.

There can be many strings made up of substrings.  I can scan all my
files.  Some of the strings are virtual.  This is the general
situation, i.e. strings are defined by programs as in Chaitin-Kolmogorov
complexity theory.
So the key operation is string search in a virtual string.
What about tree structured objects?  Maybe we have search in
a virtual S-expressions.  We test nodes to see if we go into
them.  We test them by searching S-expressions that are functions
of them.

As long as we have an actual string produced by a human, there is
no efficiency reason not to search it linearly.