Well it was the 286 I meant; DOS 3.2 on the 286 stopped at 640K, I used
ramdisk, no problem with SVr4. That's all. I'm not blaming M$ for
everything. I'm sure DOS would have been hard to keep backward
compatible, the unix didn't have to be, it was ported to the 286. DOS
caught up later.<br>
Peter<br><br><div><span class="gmail_quote">On 11/23/07, <b class="gmail_sendername">Mark Hahn</b> <<a href="mailto:hahn@mcmaster.ca">hahn@mcmaster.ca</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
> XP?) but used all 1.5 MB naturally with the unix (making it clear to<br>> me, btw, that the "640K RAM Barrier" was MS's fault, not Intel's).<br><br>I'm the last to defend msft, but this 640k thing irks me.
<br><br>640k is the physical hardware address of the graphics buffer;<br>IBM put it there. on the 8086/8088, there is no kind of memory<br>virtualization, so this is a hard limit on contiguous user memory.<br>the 286 is very, very different in memory model (16:16 segmentatation -
<br>you don't want to know more about it unless you like pain)<br>so this location no longer matters to user-space. of course, code<br>assuming only the 8086/8088 model is still 640K-aware.<br></blockquote></div><br>