Re: [Axiom-mail] Data structure for object definition

[root]

> On a slightly related note, how many "old new worlds", are there in the
> intepreter?  I've been reading the source codes, and each refers to
> "old" something, "new" something, "Davenport's new world", etc.
> Without proper documentation it is all confusing :-(

funny you should mention that.  'new' became a standard joke within the
scratchpad group.  we were all developing things rapidly and at each
meeting we talked about the 'new version' vs the 'old version'.  if
there were several changes we talked about the 'new-new version' vs
the 'new version' vs the 'old version'. after a while this became a
point of humor.

there was no tagging of versions to any great extent. 
we created and extended function in the main line on a daily basis.
every part of the system has 'new-new', 'new', and 'old' pieces.
sometimes 'new' became 'old' within a day or so.

i added an automatic timestamp mechanism so we could distinguish
which version was being run down to the second.

we were not trying to create a commercial product. we were doing
research. so you'll see layers of software (charybdis is a 1960s
2D printing routine that axiom still uses for 2D terminal output). 

the most rapid changes were in the parser as we thought of new ideas.
Bill Burge did extensive research in parsing techniques. the last
surviving technique uses what he called a "zipper parser".  the zipper
technique was never published as far as i know.  it basically builds a
stack of functions and then unwinds the stack to finish the parse.

so the term 'new' is without meaning and should be ignored.

t