Re: Why people still want to encode precomposed letters

From: Karl Pentzlin (karl-pentzlin@acssoft.de)
Date: Sun Nov 23 2008 - 11:43:56 CST

  • Next message: Doug Ewell: "Re: Why people still want to encode precomposed letters"

    Am Sonntag, 23. November 2008 um 17:24 schrieb Doug Ewell:

    DE> ... The important point is that the algorithm would not have to be
    DE> custom-fitted for each combination of base letter + combining mark(s).
    DE> Some of the font experts are saying or implying that the custom-fitting
    DE> is necessary for virtually all such combinations.

    This depends on the level of typographical quality which the user
    expects. E.g., if you see an A with a ring above its peak with a
    certain distance, you will undoubtedly recognize an Å, while judging
    the typographical quality as poor. If you see the ring with the same
    distance over an U+0245 Ʌ or an U+A72E Ꜯ, you also will recognize the
    intended combined letter, but possibly be less concerned about the
    typographical quality. Maybe you would even be irritated if such
    unusual combinations try to look graphically too much sophisticated.

    In fact, a "real working" algorithm has to regard the limited set of
    "established" combinations for which special design is appropriate.

    Thus, custom-fitting is not "necessary" for virtually all
    combinations. It is a fine tuning for typographical quality, the
    necessity of which depends how much typographical quality the user
    needs in comparison to the quality which is accomplished by the
    algorithm, which in turn depends of the complexity of that algorithm
    itself.

    Maybe, a future font formats allows "supplemental glyph sets" for
    established fonts like Times New Roman. Then, font creators can
    provide such elaborated sets for user of special languages or
    applications (like linguistic areas) to gain the maximum typographical
    quality for their specific subject.

    - Karl Pentzlin



    This archive was generated by hypermail 2.1.5 : Sun Nov 23 2008 - 11:46:36 CST