Reinoud Koornstra <kahan@ibb0021.ibb.uu.nl> writes:
> > 'vi' in particular is one that tends to drive people nuts on Linux.
> > For vi, we use 'nvi', the clone written by Keith Bostic (one of the
> > Berkeley CSRG guys).  He went out of his way to be bug-compatible with
> > traditional vi.
> 
> Plz correct me if i am wrong, but i thought vi is written by the CSRG and appeared in 2BSD (the Berkeley PDP-11 system) and is definitly not a gnu tool. I could be wrong. Plz tell me if i am. Bye,

The original 'vi' was done at berkeley, but is pretty much irrelevant
to free software systems (from a "can we use it" standpoint) because
of the licensing issues.  However, it remains the standard to be met,
when talking about the functionality of other implementations of vi.

I'm not an expert in vi history, but from what i understand it went
something like: The original vi was fairly well glued into the
structure of 'ex,' which was (or was based on) licensed software and
not redistributable.  So it couldn't be taken and just given away, and
apparently the effort involved in untangling the "new to Berkeley"
bits from the proprietary software would have been not worthwhile, so
Keith did 'nvi' from scratch.

I wasn't saying (and didn't say!) that 'vi' was a GNU tool.  My point
was that what passes for 'vi' on linux tends to be one of the
non-'nvi' vi clones, all of which seriously deviate from original vi's
"normal" (it really is hard to describe most things about vi as normal
8-) behaviour in a lot of ways.  This tends to drive people who are
used to real vi nuts (e.g. i find it just about impossible to edit
using vi on liux).  This is as compared to using nvi, which is really
very close to real vi (except for the lack of 'open mode').  A lot of
work was put into making it bug compatible.


cgd
-- 
Chris Demetriou - cgd@netbsd.org - http://www.netbsd.org/People/Pages/cgd.html
Disclaimer: Not speaking for NetBSD, just expressing my own opinion.