The orientation of the films used in masking depends on what one is intending.
Sometimes one desires more than one result. For example, in making a CRM by exposing the masking film through the base one can enjoy both sharpening and contrast reduction effects; a degree of sharpening is caused by the distance between the emulsions of the original negative and the emulsion of the masking film. A thicker base on a masking film gives more separation and more sharpening than a thinner base. Using a diffuser film when exposing the negative/masking film sandwich will also increase sharpening.
So the comment I read in the discussion above is incorrect: "Jilg and McNutt wanted to reduce contrast with no edge effects." In the original articles on our research we discussed both.
In my work I have used about every combination of orientations in exposing and printing with masks to achieve different purposes--in exposing a mask I sometimes put the masking film below the negative--that is with both emulsions down--and make the exposure through the masking film base (I know this sound counter-intuitive but there is a logic to it.) To make other masks I use an emulsion-to-emulsion orientation.
So the notion that masks must always be exposed in a specific orientation does not conform to my experience.
The variety of masks and the ways of making them are numerous, and therefore somewhat complex. The brief discussions of masking in this forum often over-simplify and occasionally mis-inform.
To my limited knowledge the best printed resource on masking is found in Lynn Radeka's Masking Kit. The kit includes the original three articles on CRMs, SCIMS, and HLMs that Mark Jilg and I wrote in 1989. It also includes additional articles on other masks written by Lynn Radeka.
To learn more about the Radeka Masking Kit, just do a quick Google search.
I love masking and can't imagine living without the unique powerful controls its give me when printing. It seems that learning masking by using only online resources would be challenging, to say the least.