I think the thing is more the TIME it takes to install that 83K+ of 
     data. I have noticed that on my machines I have installed NetBSD on 
     (IIci, IIcx, Q630 and PB540c), one of the most time consuming parts of 
     the install was installing the timezone files ;-) It's not the size, 
     it's the fact that each is a seperate file (hard link or not) and 
     takes a certain amount of time to install. I think an 83K file would 
     be quicker to install than however many hundred different time zone 
     files totalling ~83K.
     
     That's just probably the way that the installer extracts and writes 
     the files... and one reason why a native NetBSD install is so much 
     quicker. 
     
     Sure, it's picky, but if it's really that important to people I'm sure 
     there would be a way to consolidate all of the data from the various 
     timezone files into a single timezone file... which would be alot 
     quicker to install ;-)
     
     Ryan


______________________________ Reply Separator _________________________________
Subject: Re: Installer proposal (timezones)...(My waste of Bandwidt
Author:  Non-HP-kenn (kenn@echna.or.jp) at HP-Vancouver,mimegw1
Date:    4/15/98 9:29 AM


On Wed, 15 Apr 1998 08:42:57 +0100,
Greg Evans <mr_krak@televar.com> wrote:
> On 4/15/98 4:21 AM, havenerk was rumored to have said:
> >I, too, thought all the timezone info was a little bit drawn out at 
> >install.
[...]
> 
> Regarding all this timezone stuff, call me crazy, call me stupid, but 
> what is the big deal with time?
     
Nothing.  Deleting *all* zoneinfo files from the base set would save 
only 83K+ out of 9M.
     
> why don't we just make it so everyone has to learn GMT and then make it 
> so that the GMT can be adjusted in 1 minute intervals? <g>
     
Hm, it doesn't sound so crazy to me.  Installer could ask the user for 
her timezone specs (offset from UTC, whether or not to use DST, etc.), 
then generate an appropriate /etc/localtime file on the fly, if it's 
so important to save less than 1% of the base set.
     
Ken