On Mon, 9 Jan 2023 at 03:45, Ethan Dicks via cctalk
<cctalk(a)classiccmp.org> wrote:
I have a memory of installing Windows 95 on a
monochrome 386SX laptop
w/4MB of RAM in August, 1995 at McMurdo because that's the equipment
we had on hand when Win95 arrived on the continent. It was
unpleasantly slow but it did run.
Way better on a 486 w/8MB.
Oh my word yes.
But the surprising thing was that it did work, my careful
benchmarking, using MS Office, Photoshop and some other real apps,
automated with macros, showed that MS' optimization work had gone in
the right places.
WfWg 3.11 with 32-bit disk access and 32-bit file access had a fast
disk subsystem, but it wasn't able to adjust cache sizes on the fly.
You set min/max sizes and that was that.
W95 could shrink them to next to nothing if it needed.
Result: W95 started slower and felt slower on a very low-end machine,
such as a 386 with 4MB, the min spec. WfWg 3.11 started quicker and
was much more responsive.
But put both through the same set of demanding exercises in real apps,
doing a lot of work, generating documents, outputting info over OLE
into other apps and things, and W95 ran the whole benchmark suite
quicker.
It _felt_ slower but it actually traded off responsive feel for doing
big demanding jobs faster overall.
In comparison, an OS that went the other way was BeOS, which was tuned
to feel maximally responsive at all times... and for the most part it
didn't _have_ big demanding apps that could be scripted into long
heavy workloads, so BeOS felt much massively quicker on
turn-of-the-century PCs.
--
Liam Proven ~ Profile:
https://about.me/liamproven
Email: lproven(a)cix.co.uk ~ gMail/gTalk/FB: lproven(a)gmail.com
Twitter/LinkedIn: lproven ~ Skype: liamproven
UK: (+44) 7939-087884 ~ Czech [+ WhatsApp/Telegram/Signal]: (+420) 702-829-053