Re: Incompatible seq behaviour

[Brian Ginsbach <ginsbach%NetBSD.org@localhost>]

On Mon, Oct 20, 2008 at 09:28:57AM +0200, Peter Bex wrote:
> On Mon, Oct 20, 2008 at 03:15:36AM -0400, der Mouse wrote:
> > >   * It's better not to be compatible at all, than to to be
> > >     incompatible in non-obvious ways.
> > 
> > If people want (some of) the functionality but don't want to be
> > anything like compatible with other seqs, I have a count(1) I wrote
> > long ago which subsumes integer uses of seq (and has a lot more
> > features besides), but with a completely different interface - my count
> > makes no effort to be compatible with anything, and until this
> > discussion arose I wasn't aware seq existed at all.
> 
> Why do we need seq(1) in the first place? We also have jot(1).  Seems
> like gratuitous duplication of functionality to me.
> 

It is but it isn't.  In the original discussion about adding seq(1)
[see tech-userlevel January 2005], I pointed out that jot(1) !=
seq(1).  Yes, with work seq(1) functionality could be folded in to
jot(1) or vice versa.  However, in the end there may not be much
shared among them as they approach the problem domain differently.
Also jot(1) does not have seq(1)'s equal width feature.  Consensus,
at the time, by core members, if I recall correctly, was that it
was ok to add seq(1).