[indiana-discuss] scope

[Keith Bierman]

On Aug 9, 2007, at 9:20 AM, Shawn Walker wrote:

>
> How is printing a warning indicating that a script should specify its
> actual shell instead of assuming the shell bash "denial"?

You are issuing the warning the poor sod running the script; not the  
author of the script (99.9% of the time). That is just irritating,  
unhelpful and past evidence demonstrates it only drives people away.

>
> How is that anymore denial than the whole issue of seg faulting when a
> developer tries to print NULL instead of printing "(null")?

It is denial in that it was a common idiom, well accepted, and  
allowed programs to work harmlessly (so it wasn't Standard compliant;  
the standard doesn't require failure, once you step beyond the bounds  
of a standard all bets are off; while you can start WWIII that's  
hardly going to win you customers). And in that case, it was accepted  
by the same vendors system for the preceding many releases, so the  
change was viewed as particularly irritating ... especially to poor  
users trying to run existing binaries.

The key bit of denial involved, is that Solaris is not in a position  
to be arbiter of style. We are the minority. If we seek to gain  
users, we need to be nice to them, not annoy them from the onset.

Discipline the right people, at the right time and they thank you.  
Discipline the wrong people, or the right people at the wrong time  
and they shun you.


Keith H. Bierman    keith.bierman at Sun.COM   |  khbkhb at gmail.com
Strategic Engagement Team                   | AIM: kbiermank
<speaking for myself, not Sun*> Copyright 2007