Re: xpl annotation

["Carl L. Distefano" CLDistefano@xxxxxxxx]

** Reply to note from Jane Van Tassel <101233.342@xxxxxxxx> Tue, 15 Apr 97 19:35:22 +0000

No doubt about it. Long, discursive LaBels slow things down; the
degradation seems to be directly proportional to LB length. This
is the basis for the protocol of short, alpha LaBels that
Robert and I adopted in Reorganize.

OTOH, comments flagged by ;*; don't seem to matter. For the heck
of it, I tried inserting 3000 lines of dummy comments in a PM.
Amazingly, there was no loss of speed over the course of 1000
iterations. In other words, 3 million comment lines were
processed in (literally) no time!

Speed is one issue; memory usage another. It seems self-evident
(without knowing the dark innards of EDITOR.EXE) that, all other
things being equal, a PM with voluminous comments must tax memory
more the identical PM without any comments. For PMs that move
large chunks (2K or more bytes) of data in and out of Save/Gets
(the routines most apt to crash for lack of memory, and bring
down the whole editing session with them), the presence of
comments might make a crucial difference. Whether this is in fact
the case, I can't say.

> On 14-4-97, Tim Baehr wrote (in part)
>
> >Also, I wonder whether the ;*; delimited comments and the comments that
> >appear in "dead" spaces after GL's are more efficient than stuffing comments
> >inside LB's. The LB's, after all, must be evaluated by program code even if
> >nothing goes to them.



--------------
Carl Distefano * * * CLDistefano@xxxxxxxx
--------------