zyISETUpdateNotesMay2024.txt

WVF NOT IN OI ANYMORE (in dev and devPSF branches)

1) It is now an error to call 
	oi = oiSet(oi,'wvf', wvf);
If you do, you will get an error message telling you to change
it to
	oi = oiSet(oi,'optics wvf', wvf);
which will help make it clear in calling code where the wvf actually lives.

2) We currently continue to allow
	oi = oiSet(oi,'wvf paramname',wvfparamval);
but this now does the set on the wvf that is in the optics structure.
It is an error to try to set if there is no wvf in the optics structure.

We do not exectute a wvfCompute on the updated wvf. Nor do we enforce
that the value set into the wvf is consistent with any redundant values
in the oi or other parts of the optics. So, any user who
is mucking with a wvf that is buried in an oi better know what they are doing.

It is possible we should simply not allow these sets.

3) We also refrained from implementing 
	opticsSet(optics,'wvf param',val)
	opticsGet(optics,'wvf param')
as synonyms for getting the wvf, setting one of its params, and then
setting that back into the optics.  We think we want to limit how
much we encourage anyone to think of the wvf attached to the oi/optics
as a dynamic thing.  (See 2 just above.)

4) We looked for but did not find any actual calls to oiSet(...,'wvf',...)

OICOMPUTE UPDATED (in devPSF branches)

1) Add a computeMethod field to the oi.  Allow sets and gets of this.
	Options are 'opticspsf', 'opticsotf', and 'humanmw'
	
2) Adjust oiCompute to look at and use this field.

3) Find all instances of 'opticspsf' and 'opticsotf' in calling code and
change to use new field rather than the old one.


WVF [DONE ON devPsf]

1) Make wvfCompute respect the 'customLca' field of the wvf, rather than
passing 'humanlca' at compute time.

2) Fix up all calls to wvfCompute so that we're sure the 'customLca' field
has been set to match what the flag was doing previously, and disallow the
flg.

3) Take a close look at how lca is handled as oi's are created.  We think we
sorted all this out previously, but we should review that the new scheme
doesn't break the logic.

1a) Think about default for padding in oiCompute.

3) Think about how we want to create wvf fields, and in particular when we
should make it use human lca at create time.

4) Think about whether something should happen automatically with respect
to the human lens, when building an oi/optics from a wvf, as determined
by how the lca field is set.

5) We should go through code and stop using 'human' and be explicit
about 'human wvf' or 'human mw'.