Uppercase germandbls is coming to Unicode

twardoch's picture

Andreas Stötzner and the German DIN committee submitted a proposal to the ISO 10646 working group that uppercase ß (germandbls, eszett, sharp s) should be added to Unicode/ISO 10646.

U+1E9E is the envisioned codepoint.

The proposal can be viewed at:

It is important to note that according to the proposal, even after adding this character to Unicode, the standard uppercase mapping for "ß" will remain "SS". This encoding effort is not about changing existing application or even spelling rules -- it is simply an effort to encode a character to be used in an "alternate" spelling which some people use (and currently have problems with properly encoding the text). It is an observed fact that "uppercase ß" exists, even if the official rules don't envision it.

I believe it is an interesting effort, and it would be reasonable to discuss what the best possible shape for the new character would be.

Some links in German:

http://www.signographie.de/cms/signa_9.htm (published by Andreas Stötzner, I recommend reviewing all the PDFs published there.)


Some German type designers posted some of their design proposals for an uppercase ß at:

I find many of these design proposals structurally flawed -- they don’t look like uppercase letters. They look like lowercase letters enlarged to match uppercase. The graphical structure of the Roman uppercase is very different from lowercase. If one were to invent a new uppercase letter, it would have to stylistically match the Roman uppercase. If Unicode really decides to encode uppercase ß, type designers should imagine what the uppercase ß would have looked from the very beginning, rather than trying to work out of the existing lowercase ß form.

Note that the history of "ß" is somewhat surprising. The letter developed in a two-wise way: as a ligation of long s and round ("normal") s, and as a ligation of long s and z. The German language adopted unified spelling rules only in 1901. Before that, both in the middle ages and in the humanist period, German spelling differed much. For example, "Thor" and "Tor" were equal variants of spelling the word meaning "gate".

Sharp s was denoted by different writers differently (as ſs or ſz, which looked like ſʒ). The graphical shape of the ß ligature developed independently in these two ways.

This dichotomy still shows itself in a small minority practice of uppercasing ß as "SZ" rather than "SS". Incidentally, this practice is understandable for most German readers (though not actively practiced), i.e. "GROSZSTADT" or "MASZGEBLICH" is understandable as the uppercasing of Großstadt or maßgeblich.

See http://www.flickr.com/photos/adamt/490566363/ for an example.

One interesting issue is that in the 1996 spelling reform the status of ß as a single letter has been finally confirmed. In the previous spelling, the general rule was that short vowels are denoted by following them by doubled consonant letters while long vowels are followed by single consonant letters. So writing "met" always indicates a long "e:" while "mett" indicates a short "e".

In case of "s"/"ß", it was confusing. Following a vowel with a single "s" always denoted a long vowel, following a vowel with a doubled "ss" indicated a short vowel, but following a vowel with "ß" did not give clue whether the vowel was short or long. So "Ruß" was actually pronounced "ru:s" as if the "ß" stood for a single consonant letter, but "Nuß" was pronounced "nus" as if the "ß" stood for a doubled consonant letter.

The 1996 spelling removed this uncertainty by changing the spelling of all "ß" into "ss" when the preceding vowel was to be pronounced short. Today’s spelling of "Nuss" or "dass" underlines that the vowels are to be pronounced short.

The uppercasing of "ß" as "SS" but also as "SZ" defeats this clear rule. If I uppercase the word "Rußpartikel" into "RUSSPARTIKEL" or even "RUSZPARTIKEL", suddenly the natural way of pronouncing the "U" changes from short to long, so the reader is confused. The confusion is even bigger now, after the reform, because the special "undefined" treatment of "ß" no longer exists, so readers are used to "ß" being always treated as a single consonant letter, not as a ligature of a doubled consonant.

To remain logical, consistent and reader-friendly, "ß" needs (at some point) to assume a single graphemic shape in the uppercase.

I strongly feel that uppercasing "ß" as "SS" is now -- especially under the new rules -- a temporary anachronism. "ß" is a single CHARACTER (as per orthographic perception). It has functionally liberated itself from its historical background (which was a ligature of ſs or ſz).

Today, "ß" is no more a ligature of "ſs" than "ä" is a ligature of "ae". The transition process from "ae" to "ä" has been completed about 200 years ago, and the transition process between "ſs" to "ß" is happening now. Encoding the uppercase "ä" as "A ZWJ E" (or something like that) would make as little sense as encoding the uppercase "ß" as "S ZWJ S".

I believe that "SS" is an anachronic, still-in-use but slowly-to-vanish poor man’s solution to write the uppercase "ß". I believe that it should be an exciting task for type designers now to come up with a new form. In my opinion, this issue is definitely not one that is completely solved. We’re in the middle of a slow transition period for "ß". The 1996 reform started it and showed the direction.

I myself once had the idea that Scedilla (U+015E, Ş) would be most appropriate for denoting uppercase ß.

After all, Ş is historically an S with a subscribed z (that at this time looked like ʒ). Since ß is a ligature of either ſs or of ſʒ, uppercasing it as Sʒ, or, effectively, Ş, would historically make sense.

Using this notation, "Gauß" or "Roßberg" would be uppercased to "GAUŞ" or "ROŞBERG".

Similarly, the umlaut in "ä" or "ö" is historically a superscripted "e", so historically "ä" and "æ" are two different ligations of "ae", and "ö" and "œ" are two different ligations of "oe".

Since German readers are currently used to uppercasing ß as SS, i.e. they write "GAUSS" or "ROSSBERG", I even thought of a compromise: the SS remains doubled but for added distinctiveness, a subscribed z (i.e. a cedilla) is added after the first S. In other words, "Gauß" or "Roßberg" should be uppercased as "GAUŞS" or "ROŞSBERG".

Historically, this would make sense. The cedilla would here have a similar function to the trema in Spanish or French: "GAUŞS" would make clear that it comes from "Gauß" while "GAUSS" would make clear that it comes from "Gauss".

"ROŞSBERG" does not look very awkward to a German reader. The addition of a diacritic does not dramatically change the reading pattern but still adds a distinctive mark that is, indeed, needed. If I were to design a glyph that should go into U+1E9E, it would probably look like ŞS, or perhaps just SS, depending on the style of the typeface.

An alternative approach is to look at the existing uppercase-to-lowercase relations within the Latin alphabet and try to derive a shape for the uppercase ß which maintains the same relations.

In most of the middle ages and the period up until the 19th century, the long s ("ſ") and "f" were closely related, "f" being simply a "ſ" with a stroke going through. The same, very primitive graphic relation exists between the prototypic shapes of the Greek letters gamma (Γ) and digamma (Ϝ). Since the minuscule "f" always has been a "ſ" with a middle stroke, then the capital "F" might also be considered an uppercase "ſ" with a stroke going through. Of course an uppercase long s never existed, but this relation may be helpful when constructing the uppercase ß.

Because I think that *if* the Latin alphabet ever used or needed another capital S, the preferred shape could be that of a gamma (Γ). This is a simple, effective shape that maintains a stylistic relation to the lowercase long s that is typical of other uppercase-to-lowercase relations.

If we look at the relations between Aa Ee Ff Mm Pp, we will notice that sharp, edgy connections in the uppercase are related to more smooth, round connections in the lowercase. If "F" developed into "f" in a cursive hand, then it is very easy to imagine that a cursive rendition of the "Γ" shape might, indeed, look very much like "ſ".

This is an important observation when thinking about the shape of an uppercase "ß": I assert that the shape of uppercase "ß" must be "edgier" than the lowercase. In short, I think that the left part of uppercase ß should be "Γ".

What about the right part? Here, I would call to exploit the double origin of "ß", which developed paralelly as a ligature of "ſs" as well as of "ſz" (where the "z" historically used the "ʒ" shape, so "ſʒ").

These days, the lowercase "ß" is typically derived from the ligated form of "ſs". For visual dissimilation purposes -- to strongly set apart the lowercase and the (new) uppercase "ß" I would derive the uppercase "ß" from a ligation of the hypothetical uppercase "ſ" (i.e. "Γ") and the shape of "the other" origin of "ß", i.e. of the historical "Z" shape.

In short, I believe that the best graphical rendition of an uppercase "ß" would be be a well-designed ligature that incorporates these shapes: "ΓƷ"

I have made a small simulation using Garamond Premier (please excuse my poor drawing abilities):


The first line shows what the historical origin of ß looks like, i.e. long s followed by a round s. The second line shows the current shape of ß as we know it. The third line shows what a hypothetical uppercase long S might look like ("Γ"), which is just a mental exercise. The fourth line is my proposal for the uppercase ß shape.

Andreas Stötzner has proposed an elaborate document that tries to explore all possible combinations of drawing an uppercase ß:


My proposal corresponds to the scheme A1-B2-C1, which I has the most "uppercase" appearance of all those presented there.

On a related matter, at the exhibition "Neue Baukunst. Berlin um 1800", which is on display at the Alte Nationalgalerie in Berlin until May 28, I have discovered a fantastic calligraphic lowercase "ß" shape, in which the "long s" part connects to the BOTTOM and not to the top of the following "short s". Please take a look:


This got my imagination going.


dberlow's picture

ABCDEFGHIJKLMNOPQRSTUVWXYZ pretty clearly expresses itself as dominated by second-side terminal logic. So, when something is expected to be, by the sound unvoiced though it may be in a reader's mind, first-sided terminal logically, I think it is pretty important that it is, just to keep der dear reader downset. That a barless F looks more like an uppercase character than a J, is interesting, and may be appropriate in some places, but maybe just easiest way to make something look uppercase, and is not specific enough. Most who have experimented here have gravitated to serif designs, the ones with little feet things attached which makes it blend in all nice, but... I think in humanist sans, and of the sans experiments here, I think few approach success. So, relying on the stored biological memory of the original, pre-laziness lowercase form, and the space-saving intent of its creators, I go that way — clearly not a B, Beta, F3 ligature, SS, with no confusion at the waist and no gigantic open counter and with first-sided terminal logic.


aszszelp's picture

Cagri, could you please me point to those two medieval examples of capital long S?


piccic's picture

Drawing, you know, is the first refuge of the confused.
@David: why?

Beware of people who (want to) make history without knowing it.
[…]If you don’t have a proper understanding of what you are supposed to design, what are you designing then?

@Karsten: most important remarks. Altaira is also right: see historical precedents, if the form has in some cases been written in "more uppercase" form (there is no fixed distinction about the two, save the one matured in typographic tradition).

A great specific example here is the ampersand. It comes from “e” + “t”. Whoop-di-doo. When you take that too seriously you end up with a form that causes reading problems for many people*, like in Poppl Laudatio (which I otherwise adore). A well-designed ampersand is based on what people need; what it used to be is just fodder for pedantic discussion.
@Hrant: Downright pretentious, because the "and-per-se-and" is, and actually remains an "et".
I would have had problems to understand its being an [e] just *before* someone told me it was an "et", which – as an Italian – comes as a perfect logical succession, or better as the same thing.
It has happened some people asked me to change Rotis' "Et" into a more common [&] form, but that's because people in Italy still think of the [&] solely as a "commercial e" (in Italian it's called "e commerciale"), and they do not know it's an "et", out of ignorance.

Read both these entries:
http://it.wikipedia.org/wiki/Ampersand (Italian)
http://en.wikipedia.org/wiki/Ampersand#Usage (English-International)

As much as Wikipedia is nonsense, these entries show how the evolution happens by a combination of planning and usage, despite all the thinking it may be done before.
The world may have become "anglocentric" (in terms of shared tongue), but – as you see – this happens out of necessity, otherwise we wouldn't be allowed to be here to talk together in a shared experience…

Now, my big question (since it is still not clear to me): is the [ß] a simple double s? Of course not, since it's a "sharp" z-sound. So, why people keep using it as a substitute for "ss"?
Besides, how an "ss" ligature should be drawn?
And [AE] or [OE]? They are not ligatures, they address sounds, right?

AtoZ's picture

Followers of this thread may be interested in an article on the history of ß that appeared a year ago on the Typefoundry blog: Esszet or ß. Interestingly, there is a reference to this Typophile thread in the Sources section of the article -- is this an example of circular reasoning? (Smile.)
         When going from A to Z,
         I often end up At Oz.

Nick Job's picture

Any takers?

k.l.'s picture

Nope. This is anything+J.

(The bottom bowl of the lowercase eszett is too wide. I think you can move it toward the left stem -- there is no reason why the internal distance is wider than that between eszett and e, or e and n.)

paul d hunt's picture

It looks that there will be some spotty font support for this character in Windows 7: http://blogs.msdn.com/michkap/archive/2009/07/28/9850675.aspx

eliason's picture

I thought it was interesting that at the "crit session" at TypeCon, the panel of experts seemed pretty skeptical of the whole idea of the glyph.

dezcom's picture

Yes, I was in the hot seat at the time when the wrath of the Fonting Dutchman fell upon me for including it. I take their comments on this with a grain of salt. There is a unicode point for it and there were enough German speakers supporting it to make it accepted as a legit glyph. Even though it will rarely be used, I just include it for those who need it. I don't care to reopen the battle Teutonic Knights. I will let the Germans who want it have it, and those who don't can just ignore it like the florin.


paulow's picture

I dont see, into now, samples of this uppercase "germandbls" in calligraphic styles. Someone here can me show this kind of samples? Where I can find?

Igor Freiberger's picture

Hi Paulo. This article from Andreas Stötzner shows samples for script Cap Eszett. You may find two samples at the bottom right of this PDF.

Additional sources of research:

All these links are found from Signographie.de. Andreas Stötzner is the great specialist on this character and also the one who proposed it to Unicode. You may also note there are different historical forms according to German region the shape comes from. For me, Dresdner is the most interesting one.

paulow's picture

Thanks, Freiberger. I am thinking to add (althought I dont know where...) this character in my new Penabico font. Thanks ever. abraços gaudérios.

paulow's picture

So, Freiberger and people, I create my version of the Uppercase germandbls, to my Penabico font. What you think about the aesthetic solution

by the way, I have a topic forum in the Calligraphic forum about the first tests with Penabico in deutsche language texts (with Lowercase germandbls, but without the Uppercase) - here is http://typophile.com/node/74389

and, a last question. How I add the Uppercase germandbls at FontLab. Just create a 1E9E unicode gliph?

twardoch's picture

Yes, the Unicode codepoint should be 1E9E and the glyph name should be "uni1E9E".

paulow's picture

Thanks Adam.

Igor Freiberger's picture

Paulo, my favorite is the first Eszett, although all them are beautiful. Anyway, I think the design with rounded upper may cause confusion with B. A straight upper (the Ezh-like form) probably would be preferable to avoid confusion. Hope to hear some German typophilers about this.

Arno Enslin's picture

@ Paulo

No word begins with an Eszett. So I wonder, if your script Eszett is intended to be used in the begin of words.

paulow's picture

Arno, you are right. The uppercase Eszett can be used only at the middle of the words. I show here my lowercase versions - you can see the lowercase eszett in german testing in the pdfs here http://typophile.com/node/74389 -

Unfortunately, Penabico (like many copperplate scripts) dont work good using only the versals, but only using versals followed by lowercases. The use of all uppercase letters to write words need by avoided. In the case of the uppercase germandbls is really a problem. But, I am working in this. This is a new experience for me, in any case.

paulow's picture

Freiberger, you can be right. So, I will design a version (the Ezh-like form) to try using your advice. However, in this font, the B and Eszett are very distincts, like you can see here.

In time : the design of this font is from my wife, Iza W. I am only the art creation director and the programmer at the FontLab.

quadibloc's picture

Since the character ſ is an alternate form of the lowercaſe S, one could claim that it has no business in an uppercase character at all. And so perhaps the uppercase ß (for use by people who wish to use this character, which traditionally has no uppercase form, and so words that contained ß and which could alternately be spelled with ss, for which ß is, as pointed out, a ſs ligature - and, therefore, in printing, the word Große becomes GROSSE when in all caps, and current Unicode mappings reproduce this - although since no Unicode code point was previously defined for "uppercase ß", this means information is lost, which can cause problems when changing back to lowercase, it's not as if an uppercase ß codepoint just prints as two S glyphs) ought to be derived from SS - just print the two uppercase S's close together and overlapping.

As long as they're normal curved S shapes from the font in question, and not in any way... runic... a certain disturbing comparison ought not to arise as an issue.

quadibloc's picture

Here is what I am thinking of:

It looks like the uppercase of the things the lowercase eszet is made from, rather than like an eszet turned into an ahistorical character that sort of looks like a capital letter. Of course, as I'm neither German nor a professional type designer, I can't claim to be qualified to make a suggestion to be taken seriously... but, then, this is such an unusual notion that perhaps anyone can play.

dezcom's picture

It looks very dark and calls too much attention to itself. It also reminds me of the Section

quadibloc's picture

It is just a rough sketch on my part, though. I see that at least one person here did already suggest something based on the capital S but in a different way (elongated middle bar, top and bottom curves including a pointed cusp to suggest two 'S's). The middle stroke of the two 'S's could always be made lighter.

But you are right that something with two parallel strokes separated by a small space "calls too much attention to itself", as it looks more like a logo than a letter, even if the "too dark" part is easily solved.

Mark Simonson's picture

The other problem is that, as far as anyone who cares about this is concerned, the train has long ago left the station.

paulow's picture

Freiberger, concerning the Ezh-like form, like you advice me, I did make these variations. But, I pretend add to the font all the six versions (three earlyers plus these ones) I think

ralf h.'s picture


To me as a German reader, I still see them as Bs.
It doesn't matter, whether your real B look different in this typeface. Reading is too automatic to make such comparisons "on the fly".

But Nick Shinn has already solved it outstandingly:

paulow's picture

Thanks, Ralf, very clear the distinction between the shapes. I think now I understand.

Igor Freiberger's picture

Paulo, I missed your question. Anyway, Ralf already give an answer –BTW, much better than I could. Hope to see Penabico released soon.

paulow's picture

Thanks, Freiberger. I think which, maybe, Penabico will be released to public in few days, althought I will work for many months making updates, into the final project to be complete. I am attaching here a new design to the Germandbls uppercase, following the advice from Ralf (thanks, Ralf). I know which the shapes are not good yet, but, the overall shape is running to the Ralph/Nick Shinn concept, what you think?

John Hudson's picture

Paulo, can we see this in some kind of context? It isn't possible to judge these shapes without seeing them next to other letters to compare proportions and fit.

I suggest going back through this discussion and find some sample German words involving the sharp-s to set in all caps.

paulow's picture

Yes, John. Here is a sample with two words. A great problem is which this kind of font is projected to works with uses of uppercases followed by lowercases. The dozens of versals variations works well in this way, but are confused writing all in uppercases.

twardoch's picture

UPDATE: In 2010, the German Federal Agency for Cartography and Geodesy (“Bundesamt für Kartographie und Geodäsie”) published the 5th revised edition of the document “Toponymic guidelines for map and other editors for international use Federal Republic of Germany”.

Section 1.1.3 of this document says:

1.1.3 Special letter ẞ ß
The special letter ß (“strong s”) existed hitherto only as small letter (minuscule). In 2007 DIN and ISO accepted the capital letter ẞ which is rendered on position 1E9E of the Unicode character tables. In official spelling it must not be substituted by any other letter combination, as e.g. SS, ss. In the alphabetical order ẞ, ß is treated like SS, ss. Since a typographical implementation of the upper case letter ẞ in the various character fonts will take some time, it may be temporarily substituted by SS, ss.
NEW: Due to the new regulation of German spelling the letter ß is after a short (stressed) vowel now replaced by ss. The letter ß remains after a long vowel or a diphthong.

Examples for names of geographical regions:
— now: Hassberge (short and stressed vowel a), until now: Haßberge
— will continue: Meißner (diphthong preceding ß), MEIẞNER
— will continue: Großer Feldberg (long vowel o preceding ß), GROẞER FELDBERG

Example for an officially approved name of a municipality
— will continue: Haßfurt, HAẞFURT (though a short and stressed vowel a is preceding ß)

This is, to my knowledge, the first official German government body that not only approves but actually prescribes the use of uppercase ß (U+1E9E, ẞ), and only accepts the use of SS “temporarily” (“since a typographical implementation of the upper case letter ẞ in the various character fonts will take some time”).


Andreas Stötzner's picture

the capital letter ẞ which is rendered on position 1E9E of the Unicode character tables. In official spelling it must not be substituted by any other letter combination, as e.g. SS, ss.

This is a little bit astonishingly put. That an ẞ “must not” be subsitutet (»darf nicht«) would mean that WEISSENFELS is not acceptable because WEIẞENFELS is right. In practical terms, both versions are right.

paul d hunt's picture

what i really want to know is when are uppercase Latin ligatures coming to Unicode? i need my FFI!

Té Rowan's picture

The day after the Rapture, at the earliest. IOW, it's the PUA purgatory for them.

twardoch's picture


well, I think this body (permament council for geographical names, Ständiger Ausschuss für geographische Namen (StAGN), at the Federal Agency for Cartography and Geodesy) is concerned with uniquely identifable names, so there are no ambiguities. For example, the name of the town I used to live in is officially "Frankfurt (Oder)". In the past, there were various spellings ("Frankfurt (Oder)", "Frankfurt/Oder", "Frankfurt-an-der-Oder", "Frankfurt a.d. Oder" etc.), but it seems that the Germans like to have everything normalized, so, well, they also decided to normalize the geographical names.

Which I think makes sense. After all, these days geographical names need to be stored in databases, processed digitally etc., so it's useful to have *one official* spelling of a geographical name.


ralf h.'s picture

what i really want to know is when are uppercase Latin ligatures coming to Unicode? i need my FFI!

Not sure if that was supposed to be an argument or a joke, but just for the record:
The ß does not belong to ligatures like fi and fl, it belongs to ligatures like w, which (might have evolved from a ligature but) are a an official letter of the alphabet with a distinct phonetic function. And this is no matter of opinion.

dezcom's picture

At least we have the double lambda :-)

Nick Shinn's picture

Paul, you could present the case for FFI to Unicode, using René Chalet's work in the 1960s as an important precedent.

hrant's picture

As Ralf said though, there's a difference between
basic members of an alphabet versus ligatures.


paul d hunt's picture

where can i find the ‘official’ version of the German alphabet?

ralf h.'s picture

Here is the official and mandatory, most recent version of the German orthography from 2006, defined by the Council for German Orthography:
Rules: http://rechtschreibrat.ids-mannheim.de/download/regeln2006.pdf
Words: http://rechtschreibrat.ids-mannheim.de/download/woerterverzeichnis2006.pdf

Page 15 lists the official German letters:
a b c d e f g h i j k l m n o p q r s t u v w x y z ä ö ü ß

I also must stress that no one should use the term ligature as something intrinsically different than normal letters of the alphabet. This misinterpretation must be based on the fact that we usually just think of the typical f-ligatures.
But “ligature” just refers to a formal connection of some sort:
From Middle English, from Middle French, from Late Latin ligātura < Latin ligātus, past participle of ligāre (“to tie, bind”). http://en.wiktionary.org/wiki/ligature

A ligature can serve a typographic function (fi, fl, ...), a decorative function (ct, ...) or it can be an official letter of the alphabet like for example w, œ, æ, … and ß.

hrant's picture

Pragmatically speaking, I think it's confusing to call something like the ß a ligature.


Nick Shinn's picture

Three observations on German special glyphs:

1) fi, fl, ff, ffl, ffi, and st were grandfathered into Unicode as Alphabetic Presentation Forms, but not fz and tz.

2) German ligatures that aren't: ch and ck.

3) Ä, Ö and Ü — with lowered accents in many mid-20th century Antiqua fonts.

ralf h.'s picture

For more information on the phonetic importance of the letter ß also check out Adam’s post here:

3) Ä, Ö and Ü — with lowered accents in many mid-20th century Antiqua fonts.

By the way: The German umlauts had the same development from a lowercase only connection to a full letter of the alphabet. (ß is just a little bit behind, because in blackletter there was no need for it, because blackletter wasn't set in uppercase.)

Here is a book from 1774. The e had moved on top of a, o, and u:

But there was no uppercase version yet: (Ueberlegenheit instead of Überlegenheit)

This stayed this way for some time, but the uppercase umlauts still emerged, because like Adam said: it’s the only logical thing to do. Now we are at the same point regarding the ß. It just took a little longer.

twardoch's picture


not only good examples but also some beautiful shots of them, thanks! :) Do you have them somewhere in high-res? I'd love to use them in a presentation if you don't mind (crediting you of course).

Syndicate content Syndicate content