🚀 go-pugleaf

In article <264@moene.UUCP> toon@moene.uucp (Toon Moene) writes:
>In article <giDLiB2w165w@midiline.la.ca.us> dylan@midiline.la.ca.us  
>(Comrade D) writes:
>
>[ ... essentially ... ]
>
>> the problem.  The shortest summary is, "Yes, it's a problem of varying 
>> extents and it's not one NeXT plans on fixing any time soon."
>                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>> For virtual memory, NeXTStep uses a "stack-oriented" scheme that has 
>> terrible side-effects for me.  (See NeXTanswers sysadmin.569 on this 
>> scheme.)
>
>[ ... describing NeXTStep as adding new paged out pages only at the
>      end of the file ...]
>
>This is horrible.  As the NeXTanswers item says (paraphrased): "if a long  
>running process [like the WindowServer - TM] is adding another page to the  
>swap file, it can not be truncated EVEN IF ALL INTERMEDIATE PAGES ARE  
>FREED".  Well, there you have it.  And we thought we got a hot new OS  
>architecture at hand in our lovely black boxes.  EVEN VMS CAN DO THIS  
>BETTER.   
>

This is in fact horrible.  And I was doubly horrified to hear that 3.0
was not intending to address the _Number_One_Weakness_ of the operating
system as perceived by the users here.  Having to reboot your machine
several times a day under what my users consider to be "normal" conditions
is, to them, a considerable inconvenience.  Why NeXT hasn't put major
effort into this achilles heel is beyond me.

Steve


-- 
 #====#====#====#====#====#====#====#====#====#====#====#====#====#====#====#
 #  Steve Boker           #             "Two's bifurcation                  #
 #  smb3u@virginia.edu    #             but three's chaotic"                #
 #====#====#====#====#====#====#====#====#====#====#====#====#====#====#====#