[geeks] Solaris 9 changes

[Lionel Peterson]

> From: Charles Shannon Hendrix <shannon at widomaker.com>
> Date: 2004/07/29 Thu PM 05:38:27 GMT
> To: The Geeks List <geeks at sunhelp.org>
> Subject: Re: [geeks] Solaris 9 changes
> 
> Wed, 28 Jul 2004 @ 23:23 +1000, Scott Howard said:
> 
> > So having done this for years, to do a backflip and start defaulting to
> > 64 bit mode for the US-1's (with the removal of the 32 bit kernel)
> > wasn't really an option.
> 
> Why not?

Well, I see the logic in it, so let me try to tackle this one...

1) Sun publicly states that running an Ultra 1 in 64-bit mode is
   not advisable - it states this repeatedly, and consistently.

2) Based on number 1 (above), Sun's stated position is that all
   Ultra 1s should be run in 32-bit mode only. 64-bit mode is
   possible, but not advised.

3) Sun corporate decides that Solaris 10 will be 64-bit only.

4) Since Sun has repeatedly stated #1 and #2 above, and they are
   "putting all their wood" behind 64-bit computing, they can
   not suddenly reverse #1 and #2 (above). To do so would cause
   the user base to question the reality of any future platform
   statements - a bad thing.

5) By making that decision, they are able to drop a load of testing
   for their future releases, since they have removed one entire
   platform.

Anyway, it makes sense to me - it is the only possible outcome based on point #3 above... (Solaris 10 is 64-bit only - I assume this to be true, based on the conversations preceeding this)

I suspect your counter argument regarding your SS/5-170 relies on a less well-known problem - the Ultra 1 issue was a big deal because (IIRC) it has discovered around the time of the Pentium flaw...