Home› Technique and Theory
Options

Forced discretionary ligatures

Tobias KvantTobias Kvant Posts: 20
in Technique and Theory
Does anyone know if it's possible to force discretionary ligatures in the feature code?
Tagged:

Comments

  • Options
    Peter ConstablePeter Constable Posts: 199
    Please clarify what you mean. If they are forced, how are they discretionary?
  • Options
    Johannes NeumeierJohannes Neumeier Posts: 363
    You can put ligatures you want on by default into the "liga" feature, or ones that you want to not be able to be disabled by the user into the "rlig" feature. However, use with care, this should stem from necessity, not preference.
  • Options
    Tobias KvantTobias Kvant Posts: 20
    Johannes Neumeier said:
    You can put ligatures you want on by default into the "liga" feature, or ones that you want to not be able to be disabled by the user into the "rlig" feature. However, use with care, this should stem from necessity, not preference.
    Thanks a bunch!
  • Options
    Tobias KvantTobias Kvant Posts: 20
    Peter Constable said:
    Please clarify what you mean. If they are forced, how are they discretionary?
    I ment ligatures outside of the standard liga feature. I will use rlig instead!
  • Options
    Thomas PhinneyThomas Phinney Posts: 2,759
    Just to clarify:
    - there is a specific feature called “discretionary ligatures” (code: "dlig”)
    - when talking about OpenType feature code, “discretionary” is used to mean features that are off by default, and turning them on is intended to be an optional user choice (“at their discretion”).

    If we had it to do over again we might have called them “optional” or some other more obvious phrase. Even the average native English speaker probably isn’t familiar with that word usage. That’s the problem with geeks naming this stuff and not thinking about their global audience at the time!   :p
  • Options
    Nick ShinnNick Shinn Posts: 2,148
    I use <rlig> for the pseudo-random effect, because it continues to function beyond the narrow range of tracking observed by <liga>, <calt> and <dlig>. 
  • Options
    Ray LarabieRay Larabie Posts: 1,380
    <calt> is affected by tracking?
  • Options
    John HudsonJohn Hudson Posts: 2,985
    beyond the narrow range of tracking observed by <liga>, <calt> and <dlig> 
    Probably worth mentioning that disabling of these feature beyond a certain range of tracking is app-specific—notably in InDesign—and not part of a general OpenType Layout specification.

  • Options
    Nick ShinnNick Shinn Posts: 2,148
    @Ray Larabie
    <calt> is affected by tracking?
    In InDesign, at least, which is what I always check my fonts in.
  • Options
    Igor FreibergerIgor Freiberger Posts: 253
    edited April 5
    Thomas Phinney said:
    If we had it to do over again we might have called them “optional” or some other more obvious phrase. Even the average native English speaker probably isn’t familiar with that word usage. That’s the problem with geeks naming this stuff and not thinking about their global audience at the time!   :p
    I love your frank comments, Thomas. 😁
    In Portuguese, we have the same word (discricionário, discrição), but it's used with this meaning only in legal realm. Very similar situation with the Spanish (discrecional, discreción). I still remember the very first time I saw the feature name and didn't understand it.
    Maybe in a revision of OT specifications, the name could be changed. If it is done now, probably we can see it adopted in less than 30 years! 😉
  • Options
    Russell McGormanRussell McGorman Posts: 261
    You could just call them ligatures.

  • Options
    jeremy tribbyjeremy tribby Posts: 220
    Thomas Phinney said:
    If we had it to do over again we might have called them “optional” or some other more obvious phrase.
    "optional" would have been a great choice. evidence that "discretionary" is confusing includes figma referring to dlig as "rare ligatures," which is pretty inaccurate
  • Options
    John HudsonJohn Hudson Posts: 2,985
    edited April 6
    .
  • Options
    Nick ShinnNick Shinn Posts: 2,148
    My fave: “Quaints”
  • Options
    Mark SimonsonMark Simonson Posts: 1,654
    edited April 6
    I always took it to mean discretionary as in "to use at your discretion."
  • Options
    Thomas PhinneyThomas Phinney Posts: 2,759
    edited April 6
    That is absolutely what it means.

    But the average non-native speaker does not know it—and the related word “discretion” has a second meaning, arguably more common, relating to secrecy, lack of publicity, and keeping something from being widely known.
  • Options
    Johannes NeumeierJohannes Neumeier Posts: 363
    Now I'd like to see an app calling them "Private ligatures" :dizzy:
  • Options
    John ButlerJohn Butler Posts: 250
    edited April 8
    IMO historical ſi ligatures should be in normal liga next to fi and the like, and the ſ rules, whether implemented dumb or contextual, should be in the hist feature. Visually ſ is just an uncrossed or half-crossed f, and the need for its ligatures is visual and not historical.
    Mrs Eaves and Dalliance come to mind immediately, because I’m biased toward them, but there are many other designs where the quaint ligatures à-là tied ct and st have multiplied, and some are froofier than others. Stuff like tt, gy, zy, ggy, zzy that are not as exuberant could be put in dlig, and I propose a new “froo” feature tag for froofier ligatures. The endgame would be some kind of variable froof axis.
  • Options
    Peter BakerPeter Baker Posts: 180
    IMO historical ſi ligatures should be in normal liga next to fi and the like, and the ſ rules, whether implemented dumb or contextual, should be in the hist feature. Visually ſ is just an uncrossed or half-crossed f, and the need for its ligatures is visual and not historical.
    This is what I do with both Junicode and Elstob. I even have a footnote in the documentation for Junicode explaining that hlig is for sequences that should ligature only in particular historical contexts (even if the elements are modern), while ſ should form a ligature everywhere (even though one or more of the elements is historical).

Sign In or Register to comment.