nbriggs
commented
1 year ago There are two coordinate systems in an Interpress master, the master’s coordinates (which are arbitrary and defined by the emitter) and the device coordinates. The renderer translates between them with the information it is given in the master and the information that it knows for the device it is rendering on. PostScript works the same way.
From page 42 of the Interpress standard, http://bitsavers.org/pdf/xerox/interpress/610P72582_XSIS_048404_Interpress_Electronic_Printing_Standard_V2.1_198404.pdf http://bitsavers.org/pdf/xerox/interpress/610P72582_XSIS_048404_Interpress_Electronic_Printing_Standard_V2.1_198404.pdf
"For example, the creator might choose to represent all coordinates in units of 1/10 printer's point. or 1/720 inch. For an 8.5 X 11 inch page. coordinates would lie in the range 0<=x<=6120, 0<=y<=7920. The transformation from master to ICS would scale by 0.0254/720.
A particularly convenient unit for master coordinates is the mica. equal to.10^-5 meter. For an 8.5x11[inch] page, coordinates lie in the range 0<=x<=21590. 0<=y<=27940, which can be represented by a Short Number (§ 2.5.2). The mica offers sufficient precision for most routine typographic needs.”
On Oct 4, 2022, at 1:45 PM, rmkaplan @.***> wrote:
Tedit has a feature of display a paragraph using fonts with widths that are scaled to what they would be on a corresponding hardcopy stream but with glyphs that are taken from the display font. The result is ugly, often with overlapping characters, but you can see how the hardcopy lines would justify and be laid out.
There is some confusion in the code about how this is implemented, in particular, what scale should be used at various points so that margins and selections take account of the scaled widths.
Since display dimensions (and display fonts) are measured in points and hardcopy fonts are measured with much greater precisition, the general idea is to scale hardcopy widths to points, and to scale the resulting high-precision line-lengths also back to points.
One confusion in the code is what exactly to use as the scaling factor. Postscript fonts appear to have a scaling factor of 100 (i.e. specify widths in 100's of a point). But the code has a lot of places where it uses the constant MICASPERPT (=35.27778, i.e. scaled in terms of 10 microns).
I presume (vaguely remember) that Interpress used micas, and the code therefore has inconsistencies when we go to hardcopy-display mode for postrcript.
The hardcopy menu item in the paragraph menu doesn't let you specify your target file format, I think it just assumed Interpress. But however it is specified (maybe the preferred format of the default printing host, which would now be postscript), it seems that all of the scaling and unscaling should be with respect to the scale of fonts for that device, and that the code should no longer make reference to mica-based constants.
Does that seem right? Alternatively, I suppose we could pick micas or 100ths as our own uniform internal measurement precision, and convert the font-widths when we read them in.
Any thoughts?
— Reply to this email directly, view it on GitHub https://github.com/Interlisp/medley/issues/968, or unsubscribe https://github.com/notifications/unsubscribe-auth/AB6DAWN33YLSZ34NRRAHOSLWBSJPBANCNFSM6AAAAAAQ455BLE. You are receiving this because you are subscribed to this thread.
rmkaplan
D-Van-Buer
Tedit has a feature of display a paragraph using fonts with widths that are scaled to what they would be on a corresponding hardcopy stream but with glyphs that are taken from the display font. The result is ugly, often with overlapping characters, but you can see how the hardcopy lines would justify and be laid out.
There is some confusion in the code about how this is implemented, in particular, what scale should be used at various points so that margins and selections take account of the scaled widths.
Since display dimensions (and display fonts) are measured in points and hardcopy fonts are measured with much greater precisition, the general idea is to scale hardcopy widths to points, and to scale the resulting high-precision line-lengths also back to points.
One confusion in the code is what exactly to use as the scaling factor. Postscript fonts appear to have a scaling factor of 100 (i.e. specify widths in 100's of a point). But the code has a lot of places where it uses the constant MICASPERPT (=35.27778, i.e. scaled in terms of 10 microns).
I presume (vaguely remember) that Interpress used micas, and the code therefore has inconsistencies when we go to hardcopy-display mode for postrcript.
The hardcopy menu item in the paragraph menu doesn't let you specify your target file format, I think it just assumed Interpress. But however it is specified (maybe the preferred format of the default printing host, which would now be postscript), it seems that all of the scaling and unscaling should be with respect to the scale of fonts for that device, and that the code should no longer make reference to mica-based constants.
Does that seem right? Alternatively, I suppose we could pick micas or 100ths as our own uniform internal measurement precision, and convert the font-widths when we read them in.
Any thoughts?