The "InDesign result" is on-screen rendering: the font may look better on-screen under a particular set of circumstances, but it does not necessarily follow that it will look better in all circumstances by applying this "magic formula." Your initial "Eureka" moment may simply have been an instance of harmonious hinting or just-so antialiasing...

Well, it might have been Nick, but I did make a PDF and print that out. Looked better on the 1200 dpi laser printout, too. The real test is a 2400 dpi offset press, but I do think it'll pass that one as well.

The differences between FL and FOG are small, but making this change in a font does mean redoing the ligatures, and at least revisiting all the characters, sidebearings, and kerning. I really only want to do that once. ('Course I'd really like peace on Earth, too.)

I guess what you're saying is "just pick one and work with it . . ."

Wouldn't printing out both scaled versions of the font in InDesign at 100%-horizontal let you compare them with the original-at-99%-horizontal version? Or make PDFs and import all three into Illustrator for a more detailed comparison of the outlines.

We wound up writing an InDesign script that reports the width of a selected item (where "item" can be a "glyph"). Then I made up a font from the Font Lab scaling and another from the Fontographer scaling. Using the script to measure within InDesign, I got:

Width of capital A (in InDesign) using

InDesign's scaling: 5.6805 points,

Fontographer's scaling: 5.7095 points,

FontLab Studio's scaling: 5.6809 points.

So, with the limited test, FL5 behaves closer to InDesign for horizontal.

*

On to redoing the ligatures and kerning.