Re: Tame (was: XY and Win2k)

["Robert Holmgren" holmgren@xxxxxxxx]

** Reply to message from "Martin J. Osborne"  on
Sun, 01 Dec 2002 16:11:48 -0500


>> One other detail, confirmed by David Thomas to be still pertinent (I
mentioned
>> this when Tame discussion first arose), is to disable idle detection, by
going
>> to Properties==>Misc==>Idle sensitivity, and pushing the slider all the way
to
>> the left (to "Low").

> That's only relevant for a PIF, not for a LNK, right? At least, I can't find
> "Idle sensitivity" anywhere in the Properties for my LNK to XyWrite.

Correct. Idle detection is M$'s response to the problem that legacy DOS apps
think they own the machine. The slider establishes how much slack NT is going
to cut the app in this regard -- how fast the OpSys is going to pull the rug
out, when the DOS app is hogging all resources but actually doing nothing.
Implementation is half-assed, as usual, otherwise we wouldn't have any of these
problems. The point here is, that Tame is doing the work now, so with the PIF
we want to tell M$ to step aside. With LNKs, apps are assumed to be better
behaved, and memory will be managed dynamically by NT -- no idle detection is
attempted -- which is what we want anyway, to let Tame do the work.

-----------------------------
Robert Holmgren
holmgren@xxxxxxxx
-----------------------------