On Fri, 08 Jan 1999 02:02:37 GMT, par### [at] mailfwicom (Ronald L.
Parker) wrote:

>(a bunch of mumbo jumbo that only two people understood)

Whoops, I was wrong on that one.  Turns out you have to perform the
local transforms globally IN REVERSE ORDER.  So the hypothetical
(G1 G2 ... Gn)(L1 L2 ... Ln) is in fact done properly as
(LnG ... L2G L1G)(G1 G2 ... Gn).  I didn't see it because all the
examples have been symmetric.  Hope I didn't confuse you with my
mindless prattling.  The result is neater, though: you don't have to
worry about which ones are global and which local: just take all the
local ones, in reverse order, then all the global ones, in the
original order, and you're done.  Even if you started with 
G1 L1 G2 L2.

Post a reply to this message