good move Chris

[John (For the colours are many, but the light is one.) Mackin <john@ civil.su.oz.au>]

Thanks for creating this list.  One day perhaps I will have time to
do something with the new sam.  If anyone _does_ do scrolling menus
please let me know, else I will have to do it myself, in my copious
spare time.

Should an article advertising the existence of this list be posted
to comp.editors?

Now, finally, I have an audience to ask a question that has been
bothering me for YEARS.  Historical background: I first used sam
on a Jerq in 1989, hosted by a VAX-11/780 running Eighth Edition.
That version of sam (the `Pike original') didn't have what I call
`tagged REs'; you could use parentheses as you liked as grouping
operators in REs, but that was _all_ they did.  You could not use
\<digit> on the right hand side of an s command to represent `what
the n'th paren part matched.'

I have always wanted to believe that Pike designed the command language
of sam (composition of structural REs) with great precision, and I was
disturbed to find, in the V8 days, that there appeared to be absolutely
no way to accomplish the standard two-column interchange of ed: suppose
you have a file with two words per line separated by one space, and
you want to exchange them on each line, you

[[NOTE, this is an _ed_ command]]

1,$s/\(.*\) \(.*\)/\2 \1/

I found that literally _impossible_ in the Pike original sam.  I tried
hard, but always ended up with the dread "changes out of sequence".

The further history of this is that my friend and colleague Noel Hunt
(you on this list, Noel?  If not, you should be!) decided that he was
going to add tagged REs to sam, and he did it.  That code got sent back
to Pike, and it appears now in the version of sam that they've distributed.
I have always wondered (but never been brave enough to bother Pike)
what the full story is.  Challenge: can anyone find a sam command
to accomplish the exchange _without_ using \<digit> on the right
side of an s command?  (Note: please don't mail me any _theories_ about
this; I worked at it for some days, and have already tried whatever
you have theorised.  I'm only interested if you have tested it in
sam and it worked.)

OK,
John.