Council for German Orthography officially allows use of u+1E9E



  • FWIW in my experience it's generally best to refrain from explicit modification suggestions, and better to stick to verbal guidance; the necessary vagueness helps maintain the original designer's voice.
  • John HudsonJohn Hudson Posts: 2,107
    edited July 2017
    Adam, I know its at least partly a matter of familiarity, but I find I don't automatically read the hard-cornered form — as in the images of Jos' Antona — as an ẞ. I'm also finding that the strong upright on the left — without any of the familiar shapes of B D E F H K etc. on the right — keeps triggering 'I' phonemes when I see it on the edge of the fovea: GROI... STRAI.... The sense I have is that the gap after the I stroke has more impact than the connecting stroke at the top (beyond the fovea, we're more attuned to patterns of vertical strokes and spaces than to horizontal strokes). So while some people object to the round-topped ẞ on the grounds that there are no other letters with this feature in the capitalis Romanus, I think this is a benefit to ready recognition of this new letter. When I see that shape, even at the edge of the fovea, I know it's a ẞ.
  • During immersive reading, within and beyond the fovea we're attuned to any notan that provides information; that horizontal components tend to be thinner –often so thin as to be lost beyond some distance from the fovea– is essentially moot.
  • Katy, some horizontal parts are still too thick. Compare to E, R. The top curve and the downward end of the bowl are just a tiny tick too prominent. The start of the bowl (middle right) is a bit too hasty.
    But the overall approach is good.
  • I agree with Andreas' assessment, but would go further than Katy did and reduce the lateral motion of the right-hand part, as per the Zürich form. ;) I appreciate the high left shoulder and high top right point.

    John: Yes, exactly — a top left corner sends a decidedly non-ß signal.
  • It remains an open question how much the ẞ must remind of a ß versus belonging in its case. I lean more towards the latter because readers can learn what it is irrespective of its lowercase, but if it's visually jarring it will largely remain so.
  • I doubt ẞ will become ubiquitous enough to be actively learned as a new part of the alphabet. Ideally, its meaning should be instantly clear even to uninitiated readers, as was presumably the case for the famous East-German Duden and other historical uses. The hard top left is an obfuscation rendering recognition unnecessarily difficult.
  • To repeat: even lacking formal instruction, ideally it should be eventually clear without sacrificing broad, long-term performance. After all, the least "obfuscating" would be simply using an ß on steroids – but that's not design.
  • Christian ThalmannChristian Thalmann Posts: 1,639
    edited July 2017
    No, going from ß to ẞ is an undeniable improvement in proportions and rhythm for all-caps setting for basically no loss of readability. Adding a hard top left is at best an irrelevant, purely academic «improvement» (which I do not recognize) for a big loss of readability.
  • A ẞ that's too close to its ß is a bigger functional loss than a ẞ to which people need a handful of exposures to fully assimilate. (BTW there is no rhythm.)

    Initial legibility is not where a text face designer earns their keep.
  • Nick ShinnNick Shinn Posts: 1,716
    edited July 2017
    I expect that type designers will continue to implement various forms, as they have with the Euro, but the form(s) used in the most prevalent faces (ubiquitous through popularity or OS distribution) will become prevalent, irrespective of the merit of their Eszett.

    The default shown on the Unicode code page may have some influence.
  • Initial legibility is not where a text face designer earns their keep.
    And yet initial illegibility is a dealbreaker for a text face.
  • Thanks to context people can decipher one shape without a second thought.
  • When looking at handwritten samples around Berlin, I don’t think I’ve ever seen a round top-left. Some people use a descending stroke to differentiate ẞ from B, such as on this sample, and the right side differs a lot, but one thing that is very common is that the initial left stroke is decisive and downwards.

    I’ll do some more social experiments over the next few days! :) 

  • A descender is indeed something that can play an important role here. In a font where the "Q" (and often the "J") descend, vertical compactness is not an issue.

    Also, the Latin alphabet –especially in the caps– is based largely on the idea of a solid vertical stem plus some augmentation*. BTW this is probably the main reason people who have not been indoctrinated into ductal "logic" flip the "A" and the "M" so often.

    * See  "Canons of Alphabetic Change" (W. Watt).
  • Adam: Actually, those look like uses of lowercase ß amidst all-caps to me... note how it looks narrower than one would expect from the other caps, and it does not actually have a top right angle: Rather, it has an n-style shoulder. While such a thing is not unheard of for handwritten caps (I use that for B, for example), it is notable that that strategy is absent from other caps in the sample, such as R. I suppose this usage is based on using ß amidst all-caps in typing.

    Hrant: That sign is neither natural nor honorable, it's just really oogly. And I don't agree in the least with Latin caps = stem plus embellishment. I think of Latin caps as square or half-square space-filling shapes. There's a good reason why H an O are used as starter glyphs when drawing an uppercase, rather than an I, T, or F. Letters are things, not augmentations of things.
  • Intent is what counts. I don't love the sign itself, but I do love the fact that it exposes the unnaturalness of ductal "logic". Intuition is honorable, and trumps formalism.

    Maybe our starter glyphs are wrong. Although what's the cap that's considered archetypical? The "R"... Mondo stem + augmentation.

    A hard top-left corner undeniably makes the ẞ more of a capital. Whether that contributes to a given ideal form does however depend on context.
  • Christian, 
    I’m all in agreement about the classical proportions (square and half-square), and that ẞ should certainly be treated as a “wide” letter, just like C, D or O! 
  • Intent is what counts.
    No, results are what counts.
    it exposes the unnaturalness of ductal "logic". Intuition is honorable, and trumps formalism.
    Ductal logic is natural, it comes from handwriting. This is not handwriting, it's attempted recreation of something they don't understand well enough. These images do not show intuition, they show cluelessness. They are just mistakes.
    Maybe our starter glyphs are wrong. Although what's the cap that's considered archetypical? The "R"... Mondo stem + augmentation.
    The R is archetypal because it unites a stem, a round, and a diagonal, all of which are important elements of Latin capitals. An R is not a decorated stem.
    A hard top-left corner undeniably makes the ẞ more of a capital.
    I'll happily deny that. It makes the ẞ look more like some capitals and less like others, which is a bad thing. The ẞ should look like itself.

    Do you believe the S and O would be better capitals if they were implemented with hard top left corners? Certainly not — they would just be less S and O. The rounds are a feature, not a bug. If ẞ should echo any existing capital letter, it should be S, since ß is undeniably a conceptual, historic, and phonetic derivative of s.
  • The result of winning a bad bet is not a good basis for life decisions.

    Ductal illogic comes from writing with a hand ~15% of the population finds unnatural, using a particular tool invented because otherwise not enough ink would flow, and most crucially is detached from the purpose at hand: reading. It's a mistake.

    > The ẞ should look like itself.

    It should be itself, but one dimension of that is belonging to its group(s).

    > Do you believe the S and O would be better capitals if they were implemented with hard top left corners?

    I believe they would be better capitals if they were less like their lc forms; the ẞ can benefit from that too. But of course that's not the only factor, and it all depends on the spirit of the design at hand. I am in fact leaving room for various forms, with simply a preference for some; try it.

    Leaning on the historical derivation of a letter is pedantry.
  • Breaking a rule in exceptional cases does not render the rule invalid.

    Capitals don't need to maximize disambiguation against their own lc forms. Their size and proportions are more than enough. We don't need to reinvent the Latin alphabet from scratch.

    The historic aspect is certainly the least important, yet the letter ß is first and foremost «a kind of /s/» in the minds, mouths and ears of its users.
  • A fundamentally better rule than ductal "logic" is the true logic of reading.

    The size/proportion of capitals that are structurally ~identical to their lc forms is often not enough; consider smallcaps for one. Slow, deliberative differentiation is not what readability is about.

    The Latin alphabet, like anything else, can benefit from improvement. In fact it's been doing that since day one, with some people helping and others looking away. Over time even entire new letters have been added...

    Let's help users get over their hang-ups. In fact I see that as part of our duty.
  • Thank goodness for "SS". Großcedillahängtagalgen seems much more honest than a city name, to me.;)
  • John SavardJohn Savard Posts: 811
    While the article cited was in German, I could see its point: that the Roman long s and short s ligatures for the eszet were not, in fact, its ur-form, but were simply something handy from another language that German typesetters of a later period used.

    If the eszet were merely a ligature, there would be no more need for a capital version in German than there is for a capital version of the fi, fl, ff, ffi, and ffl ligatures in English. But even if it were a true ligature, the fact that it was based on the long s, and survived after the long s was abandoned, would have been enough to start the glyph being regarded as a different letter in its own right. An origin that was lost in the mists of time would increase that tendency.

    And once the eszet is regarded as a letter, the fact that it can also be rendered as ss would no more eliminate the need for a capital version than the fact that ü can be rendered as ue eliminate the need for a capital version of that accented letter.

    But looking at the different designs for the shape of a capital eszet, it does seem to me that it's a very difficult symbol to produce a corresponding capital shape for. One could even suggest that it would be just as well to take an ordinary capital S and put, say, a double horizontal bar over it!

    Even if it's a "false etymology", as it were, decomposing the eszet as a ligature is a starting point for constructing a capital eszet. But that process then hits a roadblock: there is no precedent for capital ligatures! The first thing that process would also suggest is that instead of an upside-down J, or a capital gamma, on the left, the glyph should begin with the normal capital S shape.

    So, think of a symbol that looks a bit like a cattle brand... a normal S, followed by a highly condensed S, narrow but normal cap height, joined by a horizontal bar like that of the H. That would look nothing like an eszet, but it would be a capital ligature, corresponding to a small ligature built from two small letter shapes.
  • Nick ShinnNick Shinn Posts: 1,716
    there is no precedent for capital ligatures! 

    What about W?
  • John, good argument with Ue and Ü. Ue was a crutch and is no longer needed now that Ü is available; same with SS and SZ. 

    I disagree with your design proposals, though. Since ß is a monolithic glyph rather than a ligature in German, the same must go for its capital. Barred S or trainwreck-SS sound alien and ungainly. 

    The now official design does retain a bit of S in its round top. The rest of it has been stretched taut into a vertical to make space for the action on the right side, as has happened to the long s in the lowercase version. 
  • John SavardJohn Savard Posts: 811
    edited July 2017
    What about W?
    After I posted that, I thought of Æ and Œ. Of course, they, like W, if considered as a ligature for V V, or the Dutch IJ, all do retain the full capital form of the joined letters.

    A ligature for SS can't follow Æ and Œ exactly, as joining the two letters by having them in contact would look messy. So a bar joining them would be a reasonable way to link them.

    Approaches involving opening up the letter S so that the open space of one letter could overlap with the line of the other could have unfortunate associations, so I would not suggest anything in that area.

    I disagree with your design proposals, though. Since ß is a monolithic glyph rather than a ligature in German, the same must go for its capital. Barred S or trainwreck-SS sound alien and ungainly.

    This, of course, was a suggestion that is relentlessly within the Roman character tradition, and so it's appropriate for conventional serif typefaces. Sans-serif ones, on the other hand, have a better opportunity to show respect for the Fraktur heritage of German script, and so something like the existing versions of capital eszet is appropriate for them.

    But I do think it's difficult to find an appropriate capital form for the eszet qua single glyph, and all the existing forms shown in this thread seem flawed to me.

    This link,
    illustrates some of the historical versions of the capital eszet noted as existing earlier in this thread. It also explains why a capital eszet is very much needed in German, which is interesting.

    The Schelter & Giesecke 1912 design could be used for a Roman version of the capital eszet instead of my suggestion, I'm happy to accept that it is better.

    is a 1912 version from the Bauer typefoundry.

    Then there's
    illustrating that again along with others.
  • Are you referring to the S3-like design? It works in these very decorative settings but is too elaborate for more serious typefaces. It's also manifestly the result of cramming two letters together, which always looks forced and contrary to "ß as a monolithic character". 

    There are many questionable implementations out there that don't integrate well, but by now there's also a growing corpus of good examples. I find my "Zürich" recommendations particularly helpful in blending the letter in with the other caps, but I know I'm biased there. :grimace:
  • Deleted AccountDeleted Account Posts: 739
    edited July 2017
    "What about W? I_J?" 

    There were already Uppercase V, I and J, when the glyphs W and IJ were added, Nick. 

    There is no uppercase, (or small cap) long S. Adding a "go"-like shape from cyrillic, or a short cap S, (shortened by replacement of the entire bottom with a straight ending), are the suggestions beyond the state recommended use of "SS".

  • John SavardJohn Savard Posts: 811
    edited July 2017
    The shapes shown in Papier und Druck as cited in that last link gave me some inspiration; I like the idea of beginning the character with a narrow S, as in the fifth illustration in the first column, but to make the second half a Z instead of an S and the character more distinct from B, as other references note is needed, I've come up with this suggestion:

    Found a larger version of the illustration in a PDF of a German version of that posting in the form of an article, and could do a better mashup from there:

    on a larger scale as well. (I've seen that word in German-language atlases.)
Sign In or Register to comment.