Implementation:

doToggleButton(qualityIndex) is in jcontrol.js

toggleIsDown[slabIndex] ARRAY NOT FUNCTION
	is state of buttons in all control panels for that index.

(There is toggleQuality() but it is only in EDM.)


-------------------
WHY does the quality button take a long time on a large molecule like 1g3i (45K atoms)
but antialiasdisplay is instantaneous?

Initial quality is a preference (v 3.5).

Default quality is up/down in controls.js, but that only defines toggleIsDown[qualityIndex],
it does not set antialiasdisplay.

controls.js:	makeToggleButton("qualityOnSpt", "qualityOffSpt", Quality_txt, "down", "",
controls.js:		QualityTT_txt, cptab);

FG 3.1: Changed default to down, but if > allAtomsM1C1Count > qualityMaxAtoms (33000)
in doPostLoadActions() in top.js:
	if toggleIsDown[qualityIndex] doToggleButton(qualityIndex)
BUT the toggleIsDown is false at that point.
And neither doToggleButton nor setButtonState() work in doPostLoadActions() -- ???

Spin is slower for larger molecules. This was corrected when
allAtomM1C1Count > qualityMaxAtoms, quality is turned off at outset.

Before FG3.1, 'set spinfps 30' in V1 scripts.js.
Changing spinfps to 10 sped up large molecules in high quality, so 
set spin y 5 remained in effect.

doPostLoadActions() is called in msgcall.js at the end of
processPDBHeader() is called in
jmolMessageCallback()

BUT doToggleButton FAILS in doPostLoadActions where in fact toggleIsDown[] is FALSE (???)
even though at alert time the button is down!
STILL setButtonState works in doPostLoadActions, so I decided to set the antialias
conditionally on {model=1}.count in the V1 script (at Jmol level).