• State Verified Answer
  • Replies 4 replies
  • Subscribers 106 subscribers
  • Views 2325 views
  • Users 0 members are here
Options
Related

Defining very font family specific font "fail overs" via faked FAST/FV

Thomas Mangan

I'll try to explain this the best I can, here goes:

One if the questions asked after Bart's CSS One Year Later session touched on creating a new FV record that referenced multiple FASTs. The example of why someone would do this focused on combining Times with a Times version that included cyrillic glyphs.

The thought here is that if one foundry's version of Times didn't contain a specific glyph, I could attempt to fail over to another foundry's version - and hopefully minimize any visual differences.

An example would be that if Times STD didn't have a specific bullet type, but Times PSMT did, XPP would take the bullet from Times PSMT instead of failing over to the more general glyphs contained in the secondary FAST entry.

Any chance someone involved in the conversation could provide a bit more info (our at least where to look in the doc to find out more about this)?

Kind regards
-Thomas
Rate translation
×
Suggest better translation
×
  • 0

    Thomas, the doc you are looking for is the Fonts doc that comes with your system. It is a bit of a read with its 276 pages.

    Now let's suppose that your Times STD has been read into your XPP system as pts 30. And that your Times PSMT has been read in as pts 330.

    Here is one way of getting what you want (there are other ways that would work as well):

    • in the fgs 30 for the Times STD you add pts 330 as the last entry
    • recreate your FAST 30

    and that is it.

    You will now have a FAST 30 that contains all the characters of your Times STD in combination with all of the extra character (like you bullet) that exist only in the PSMT font.

    If you only want to include only the one bullet you will have to do some extra work...
  • 0 in reply to Bart Terryn

    Thanks for your reply, including the entire second font would seem somewhat safe (better to have all characters possible than miss one by customizing); any idea if doing so would introduce a performance cost?

    Ideally, I would only want XPP to use TimesPSMT for the characters that TimesSTD (or other, similar, Times families) is missing. 

    Initially I'll test things out with the all-characters present scenario prior to doing a bit of reading and trimming things down a bit.

    Kind regards
    - Thomas
  • 0 in reply to Thomas Mangan

    Thomas,

    It seems to me that what you need here is a PI FAST that you would call as a secondary FAST in the FV spec.

    The process to built such a FAST is a little bit too complicated to be explained here, but you can find all explainations in the "fonts" XPP doc (Bon courage !)

    Best regards,

    I. Durand
  • 0 in reply to Thomas Mangan
    Thomas,

    There would be no performance cost associated for compose. The character would be looked up in the FAST that you created just like any other character of that FAST. And the lookup in the FAST would return the width of the character and the font (and the position of the character in that font) it belongs to.
    The only thing that would happen is that your PostScript file would be come a bit bigger as now you would have a font change.
    Whether those things slow down distiller or not, it beyond me...:-)