From: Tom Sawyer <transami@tr...> - 2002-09-14 11:24:37
so i have one more question.
by the way clark your answers have been fantastic and would make a good
start on a yaml book (hint hint)
i'm curious about the difference between serial tree and random-access
graph. now i understand the techinical differnce. that's not what i'm
getting at. and i am also asking this indiffernt to our previous
conversation about the serial tree retaining mapping key order and
duplicates, and the graph not doing so. assume that's not invloved.
i am asking on the level of use. what does the serial model give me that
the graph does not, that is of real advantage and such that it simply
would not be worth doing otherwise using the syntax and/or the graph?
this has to do with my analysis of the semantic chain. the only semantic
factors of tree vs. graph are that aliases become referent and types
resolved. correct? when would i ever really need to see them as not
being so? i am trying to find reasonably common use cases here and i'm
not seeing it. everything i can think of has very limited use that can
be delt with at the graph level and/or at the systax level with minimal
what's the concensus here?
tom sawyer, aka transami