Subject: Re: integrating PAM
To: Dan Melomedman <dan%dan.dan@devonit.com>
From: Harry Waddell <waddell@caravan.com>
List: current-users
Date: 01/22/2003 16:22:58
On Wed, 22 Jan 2003 18:49:08 -0500
Dan Melomedman <dan%dan.dan@devonit.com> wrote:

> Harry Waddell wrote:
> > Your probably right about these things not being needed most of the
> > time. Unfortunately, the exceptions that come to mind, e.g. academia or
> > government, are the kind of users that may not use NetBSD because of
> > these"deficiencies". Based on my [and Bill's] past experience, these are
> > also the kind of places that occasionally fund/contribute substantial
> > work on netbsd. That said, this is probably more of a non-critical
> > advocacy issue than a technical one.
> 
> This is MS mentality - add every possible feature anyone might need.

Gee, you could have just compared me to Hitler. :-)

> Unfortunately bloat costs even more in human time spent maintaining it.
> And when it becomes too bad, resulting in a complete rewrite of  the
> bloated piece from scratch.
>

I'm not saying PAM is the solution, since I'm only just beginning to look at
the xsso docs, I not entitled to an opinion yet. [although is seems a bit
ponderous, even at first glance] I am saying that things like token based
auth foo should be supported _somehow_ if we care about being inclusive of
some of the places that regularly "fund" R&D. Just because something isn't
commonly used, doesn't mean it's not important. 

> Good software is written following the KISS principle - Keep It Simple,
> Stupid.
> 

I like simple and orthogonal tool-sets as much as the next guy. 

Anyway, as entertaining as all this squabbling is, until such time as
an impending change is announced, doesn't this whole thread belong in
tech-userlevel?

-- 
Harry Waddell
Caravan Electronic Publishing
---