🚀 go-pugleaf

In referenced article, phil@pepsi.amd.com (phil ngai) writes:
>
>Well, guess what. This driver simply grabs 64 Kbytes of conventional
>memory, calls it the memory buffer, and implements EMS by going into
>protected mode and shuffling data back and forth to extended memory.
>....
>With standards like this, who needs standards?

You are not describing a deficiency in the *standard*, you have found what
might be initially considered a poor or restricted *implementation* of the
standard.

However, you are describing a *software only* implementation that gives you
EMS on a machine that has no hardware to support it.  In effect it is an
*emulation* of EMS memory using 286 extended memory, and as such can be
extremely useful.

A similar implementation of EMS can be done (and has been) on an XT machine
using swapping to disk.  I may be wrong, but doesn't another popular but
multi-tasking operating system also use hidden memory/disk swaps?  (:-).
[Yes I do understand that automatic paging a-la-UNIX is infinitely more
program friendly than the simple but awkward to use application controlled
memory paging of LIM 4.0]

The *speed* of an EMS implementation, is of course something that an
application program can do little about, however the other complaint, the
fact that it used a maximum of only one 64K page, can of course be detected
by the application which could modify its behaviour accordingly.

The fact that the standard allows a wide variation in the window and
physical memory availability is, of course, because of the wide variation in
the physical hardware on which the standard is defined to run.
-- 
Ray Dunn.                    | UUCP: ray@philmtl.philips.ca
Philips Electronics Ltd.     |       ..!{uunet|philapd|philabs}!philmtl!ray
600 Dr Frederik Philips Blvd | TEL : (514) 744-8200  Ext : 2347 (Phonemail)
St Laurent. Quebec.  H4M 2S9 | FAX : (514) 744-6455  TLX : 05-824090