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.


Nick Shinn's picture

Here are some more I've done recently, for Oneleigh and Goodchild.
These will be released soon in updates.
I prefer the "Leipziger" style for old-style (Antiqua) faces.

Mark Simonson's picture

That pointy part at the upper right of the Leipzig variant just looks weird to me. I can't think of any other letter that's made using a pinched vertex like that.

In spite of that, yours are looking pretty convincing, Nick. How would it look in something really bold or geometric like Soft Machine? :-0

Nick Shinn's picture

You're right about the pinched vertex being unusual.
I'm afraid I didn't rationalize it too much!
The Leipzig variant would be a challenge in something really bold or geometric, but for me it just seems appropriate for the old style, and I'll probably go with the Dresdner variant for Softmachine.

piccic's picture

I think as designers, we work on glyphs and have a knowledge of what is the essence of that glyph […]
I think Chris is entirely right. I think you should get the essence of a letterform. Since the SS is a "S-related" or "S-sound" letter, I couldn't treat it as an "abstract" form to invent.
Here's my form for Neoritmo. The UC [ß] is the same height of the [H], which has an ascending part. The interruption of the [ß]s is a choice, and the form can be seen continuous. What I think is that we may compensate for its feeling of "lowercaseness" by making the upper left corner a "supercurve" (more angular), and thus neither rounded nor angled.

piccic's picture

P.S. Nick, why did you use for the UC of Oneleigh Roman the "Leipzig" form, while the LC had the "Dresden" one?

ralf h.'s picture

many of these are hard to distinguish without context: Il1, O0o, rnm, 5S, 2Zz. ... Context helps a lot.

I totally agree to that, but the thing is: B and capital sharp S will always appear in the same context, so we need a clear distinction within the skeleton of the letter. Context can not always help:
"Die große Wurst" (the big sausage).
"Die grobe Wurst" (the coarse sausage).
I wouldn't even mind if different shapes come into use (as with single/double-storey a and g), as long as they cannot be mistaken for a B.

Nick Shinn's picture

why did you use for the UC of Oneleigh Roman the “Leipzig” form, while the LC had the “Dresden” one?

I see what you mean.
I chose the Dresdner form for the LC as I liked the way the top stroke suggested a chirographic break with the bottom, and that quirkyness seemed to suit the face.
I don't think there's a necessity to have overt shape-echoes between the cases, as the UC eszett only appears in all-cap settings.

The main thing is to have the cap Eszett look well in all-cap settings, and the LC ß look good in mixed case.

cuttlefish's picture

Here is how I've handled this character in Agamemnon, Californian Grotesque, Palormak, and Effluent:

These are all works in progress and subject to change. I've tried a variety of the suggested solutions and these seem to work best within their own context, being neither to lower-casey nor B -like, and certainly unmistakable for their lower-case counterparts.

I keep forgetting or swapping in my mind what distinguishes the Leipzig from the Dresden form, and I'm not sure how either applies to my designs, so I'll refrain from using those words for now. I'm sure there is room for improvement.

dezcom's picture

Those look good to me except the last one (the didone). The Greek Gamma plus s just looks like a strange creature or an old Hangman game?


cuttlefish's picture

I agree that one on Effluent (the didone) looks pretty odd, but the other styles weren't working at all on that one. It is the same structure Tim Ahrens used for his, though.

Nick Job's picture

Purely out of respect, don't you have to look at what German typographers have done/are doing/will do with this character? Is this not one for German designers to solve (at this 'early' stage at least)? This is a uniquely German challenge, isn't it? I can't help thinking that if there were a 'new' character used only in my language, that if someone steamed in from another country that spoke and wrote a different language and told me what it ought to look like...

If one can get rather animated and protective over the correct vs incorrect angle of the smallest accent in a particular language, how much more would an entire character get tensions running high.

I feel completely unqualified to comment on any of these designs, much as I like some and dislike others. I am far more interested in what, for example, Ralf thinks (he is the 'consumer'), especially as he has been so helpful by elucidating some of the issues involved, which would never have occurred to me.

Just because there is going to be a gap in the Unicode list, it doesn't mean it has to be filled in a given font, especially not with dross (I'm not saying the designs above are dross, by the way). The slightly irritating fact of the matter is that the trend is in danger of being set by the hastiest designers rather than considered and informed solutions which might solve the problem in a more efficient manner. Idealistic, I know. I'm sure someone will pile in and put me straight.

For all sorts of reasons, let's not regret our Eszetts.

Mark Simonson's picture

Granted. I just hope that whatever form is settled on, it will fit comfortably and flexibly into the typographic gamut.

Nick Shinn's picture

Nick, I don't think the trend will be set by the "hastiest designers", I'd credit German typographers with more sense than that.

I don't believe this is a problem non-German designers can't solve, it's not very difficult in that respect. All the background is accessible in this thread. All we need to know is summarized by Jason,

"these seem to work best within their own context, being neither to lower-casey nor B-like, and certainly unmistakable for their lower-case counterparts."

All we have to do is make that work in our own types, and you can see an example in Jason's Agamemnon--surely you don't have to be German to appreciate how nicely the Eszett is realized as an expression of that typeface, how well it integrates with the other characters?

hrant's picture

Every time I catch my breath long enough to post to
this thread, it has already grown beyond recognition!
It warms my heart - I love newborn characters.


This is clearly a tricky design problem, with many factors pushing and pulling. If German didn't use caps so much, the problem of the cap eszet looking too much like a "B" would be secondary. But that's what we have to work with.

I mostly agree with Nick on the "nativity" issue in this case. As unnerving as the ethnic intrusion might feel to some (I know from my own experience of watching non-Armenians design Armenian type), especially because this is a new character/glyph I think non-Germans can contribute a lot.

That said, of course Germans are the "customer". Even though we can't trust laymen to give reliable feedback about a proposed glyph's stylistic integrity with the whole or its non-ligatureness, the issue of diverging sufficiently from the "B" is something I think that only German laymen can help us with.

So I propose rounding up the best few candidates (for a normal text font, like maybe Georgia), preparing a Flash movie where some carefully-chosen words are... flashed for varying very short durations in front of readers on non-typographic web sites, and then ask them what they read.


Mark Simonson's picture

It's sort of a chicken and egg problem. The reason we can read something in a "flash" is because we are thoroughly familiar with the forms. If you introduce unfamiliar forms, the odds are stacked against their being interpreted correctly, no matter how reasonably the forms have been designed.

hrant's picture

If you only show individual glyphs, certainly. But if you carefully chose the words that either have a "B" or a cap eszet, I think you might start getting a feel for the boundaries. Think for example of the word "apposite" in English. It's a much rarer word than "opposite", so in Avant Garde for example you will get many more people reading it as "opposite", but in Quadraat for example more people will read it as "apposite", even if they don't even know what it means. BTW, "apposite" sort of means the opposite of "opposite". :-)

So we would need to find German word pairs like that, and try the different proposed cap eszets to gauge which ones look less like a "B".


Nick Shinn's picture

That's an interesting proposition Hrant (research-worthy), and the results would be valuable, but your example would appear to legitimize forms that are not optimal for disambiguation. If the single bowl "a" is apposite in Avant Garde, Futura, etc., why shouldn't a cap eszett that is somewhat B-ish be OK in some faces?-- and in fact I think Mark's cap eszett for Futura is like that; it's B-ish, and it seems to suit the face precisely for that reason.

Tim Ahrens's picture


If German didn’t use caps so much, the problem of the cap eszet looking too much like a “B” would be secondary. But that’s what we have to work with.

I don't understand what you mean. Can you explain that?

Nick Job's picture

Nick S

When I said 'hastiest', maybe I should have said 'most eager' and maybe being eager is a good qualification to set the pace. Maybe I was hasty, sorry.

> I don’t believe this is a problem non-German designers can’t solve...

You're absolutely right, anyone from anywhere can solve a problem (to say otherwise would be bordering on racism), but in this case isn't it a bit like telling a French, Spanish, Catalan person etc, what their cedillas should look like, when they were the ones who invented it/them? That just makes me nervous.

There remains a conflict in my mind because the reality is that no-one can really tell me what my Eszett should look like in my own font. It's no-one else's business (not even Germans) and in that sense we all have to be able to solve the problem for every font that we design.

However, I'm still intrigued because the 'problem' seems to have only come to light in recent times (hasn't it?) That makes me think Germans were not that bothered about solving it (not as bothered about it as the majority of us on this thread who aren't German) and for us to tell them that they have a problem makes me very nervous.

Finally, on a creative note, no-one can have a genuine problem with this new character being B-ish? (P and R are all relatively B-ish but voice completely different sounds) as long as the two characters cannot be mistaken in context (that is where Ralf and friends have indispensable advice). In the end usage and context may well design this character for us.

Was ever a character conceived in this 'mass committee' way before?

guifa's picture

Was ever a character conceived in this ’mass committee’ way before?

Not just a character but an entire alphabet.

«El futuro es una línea tan fina que apenas nos damos cuenta de pintarla nosotros mismos». (La Luz Oscura, por Javier Guerrero)

Jongseong's picture

Not just a character but an entire alphabet.

Actually, despite the popular Korean belief that King Sejong appointed his team of scholar-officials with the creation of hangul (the Korean alphabet), modern scholars agree that it was a personal project of the king himself. So hangul was not conceived by a mass committee.

As for characters that were designed in a mass-committee manner... I have no idea, but maybe phonetic symbols? Characters for some African orthographies?

Nick Shinn's picture

Nick, I don't see this character as a problem, but an opportunity created by Unicode and OpenType.
Unicode and OpenType, among other things, are part of present-day internationalism, and the field of font design and production are wide open to foundries worldwide.

If foreign type-designers are heavy-handed, it's up to indigenous typographers to ignore their typefaces.
Trans-national corporations like Microsoft and Adobe, that are bundling large numbers of fonts globally, have to be more circumspect, because their types will become something of a de facto standard. However, the fact that there is a variety of options available to those who choose and purchase fonts from independent foundries should provide cultural checks and balances, while allowing some "foreign spice" to make things more interesting.

hrant's picture

Nick, what I'm describing is merely an observation tool - it can only help a
designer make a decision; it cannot replace the designer's overall judgment.

Of the three factors I can discern, namely stylistic fit, non-ligatureness and non-"B"-ness, only the last can be addressed by the sort of informal field study I describe. And all such a field study could do is give the designers a feeling for how a glyph is being read. That still leaves the bulk of decision-making to the designer's judgment.

So for example you might only give up on a very "B"-like cap eszet for Avant Garde if the study results are totally disastrous.

> I don’t understand what you mean. Can you explain that?

I was hoping my "apposite" example clarified things. But let me try this: you choose a pair of German words where one has a "B" in one spot while the other has an eszet in the same spot (ideally with both words having comparable frequency; or you could weight the results based on the relative frequencies). You set them in all-caps and flash them for very short periods in front of readers, then ask them to type the word. Refinements might involve: using different durations, maybe between 1/10 and 1/4 of a second; using words that start with "B"/cap-eszet and setting the rest in lc; and putting the words in the parafovea (by asking the reader to focus on a point that's some distance* to the left of the flashed word).

* To be determined by factoring in screen
resolution and viewing distance; tricky.

> Germans were not that bothered about solving it

Well, most Germans don't design fonts. :-)
In fact many Germans want[ed] to dump the lc eszet! For shame.

> ... as long as the two characters cannot be mistaken in context

1) Sometimes they could. But the fewer words that might have this problem, the lesser the problem, and the less it makes sense to sweat it, I agree.
2) This is a new character. Changing an established character is a whole other animal!

> modern scholars agree that it was a personal project of the king himself.

Really? I didn't know that. In a way, that's even more impressive (or maybe my monarchist tendencies are at play :-).

It's worth noting here that many people (including myself) consider Hangul to be by far the most powerful and elegant writing system. To me it makes English look like the village idiot.

> characters that were designed in a mass-committee manner

A mass, anonymous committee:


ralf h.'s picture

isn’t it a bit like telling a French, Spanish, Catalan person etc, what their cedillas should look like, when they were the ones who invented it/them?

No, because were designing a completely new character here, not changing an existing one. Anyone in the world can understand the design problems and solve them. As for me, I really appreciate the ideas in this thread.

The introduction of complete new alphabets is a different thing, because everyone was forced to learn a completely new system. But here we're trying to sneak in a new character in a set of characters everyone is used to. This is not easy. It must have a distinct design but still feel like it was always there and must be easily recognized as a cap Eszett.

Arranging legibility test with system fonts like Georgia would be a nice way to check for the B problem. Speaking of Georgia: Maybe we should also try to solve this the Carter-way: The character must also work in a small copy text like this one here on Typophile. Maybe it would be a good idea to start with simple pixel grid, in which it would be easy to see if the skeleton is different enough from B and R.

cuttlefish's picture

Putting a bit more work into it, I came up with a very different solution for my didone Effluent (I'd really appreciate another name suggestion--I'm tapped out). The new one follows a configuration similar to what I used for Agamemnon, with the Gamma_s lig type from before shown for comparison. The new one is a bit more B-like, but not so much as to cause confuddlement.

hrant's picture

I think this might have been discussed before, but what about giving it a descender?* That might pull it nicely away from the "B", but hopefully not too close to looking like a lc (which would be another possible field test). The descender could either be at the left stem or the right curl.

* Except in fonts where both the "J" and "Q" don't descend I guess.
A relevant question here: are "J" and "Q" very low frequency in German?


hrant's picture

Oh, I said that 1.5 years ago. :-/


cuttlefish's picture

I think J has a fairly high frequency in German. Their word for "yes" starts with J, so it's at least an important letter. I don't think I saw the Q used very much in that semester of German class that I slept through, though.

dezcom's picture

That looks much better, Jason!


Florian Hardwig's picture

If German didn’t use caps so much, the problem of the cap eszet looking too much like a “B” would be secondary.

Yes, we like to capitalize a lot of words. All nouns, to start with. But there are no words that start with an ‘ß’.

So we would need to find German word pairs like that

Here you go:

[big — gross]

[greeting — to dig (past tense)]

[to be called (past tense) — hit/slash]

[to let (past tense) — dear]

[to tear — to rub]

[white — woman (as in shrew)]

[to pound sth. — to scatter (past tense)]

[shit — pane/slice]


hrant's picture

Florian, awesome! More, and much faster, than I expected.

Any way to get frequency numbers for those? Actually, failing that, migrating the frequency numbers from English* would be much better than nothing (since people pretty much talk about the same things no matter where you go :-).

* Which I can provide, via Kucera & Francis.

So who's gonna set this puppy up? :-)
I'm willing to help with the "interface
design" and (non-linguistic) testing.


ralf h.'s picture

Here is a quick test with a 9 pixel cap height:

It is build in FontStruct and anyone can modify it/add Eszett variants:

nina's picture

Excellent idea, Ralf. I took the liberty of trying out a few variants:

I didn't change your FontStruction but clone it to my own space (version with my added variants is here).

I agree with your post over on typografie.info that a descender on the right hand side (as in my third variant) probably doesn't make sense from a logical point of view; still, it does help to differentiate so I figured I'd give it a shot.

Jack B. Nimblest Jr.'s picture

Tim: "First, show us your solution and then you can criticise the others."
No thanks. Drawing, you know, is the first refuge of the confused.

... an uppercase long s and an uppercase s englyphed together into one?

I must apologize for any l.c. German ß I made with a leading edge crossbar on the long s. Those were stupid mistakes following stupid models. And why I ever followed the crowd on roman l.c. ß that followed the f and not the fj, I have no idea.

Should I ever make an uppercase ß either with a crossbar on the long S or via an uppercase F with the bar removed, crit me to within an inch of my life. And if I ever make one that looks like a l3 or J3 ligature, strike me dumber.

Now, would I want to make something that tries to disappear right away, or should I try to make glyphs that are correctly clear, but that may take a generation or two of readers before the glyph disappears into the text?

I await an order, that is as usual, specific to an audience, a style, a resolution and a size.

Nick has gotten the closest so far as I can agree, but just once and not quite.


Nick Job's picture

Should I ever make an uppercase ß either with a crossbar on the long S or via an uppercase F with the bar removed, crit me to within an inch of my life. And if I ever make one that looks like a l3 or J3 ligature, strike me dumber.


The trouble is, many upper case glyphs look like their lowercase counterparts but many don't - it's about half and half depending on which font you're talking about. Now, who is to say whether the upper case long s is one of the characters that looks like its lowercase counterpart or not; e.g. it might actually look like an F/gamma or even a crazy J, who is to say?

But, surely anyone can decide what an uppercase long s looks like and may the best man/woman 'win'. Aren't you (and Nick S) saying that the uppercase long s looks like half an inverted U and not a F/gamma? Why is that anymore right than the F/gamma approach?

I'd love to say that historically the long s had nothing to do with f but it kinda did, didn't it? You can't really fault anyone for going for the uppercase F as a model for the uppercase long s too.

Are you saying, first and foremost, that we need to have some consensus on what an uppercase long s looks like? If so, you are bound to get some that think that because what you're talking about doesn't really exist, then this new glyph should actually look like two regular uppercase S's ligatured. Then the next logical question is, "Did they actually need ligaturing?" I'm not sure they do, just like FI and FFL etc don't really warrant special characters.

All I want is for the logic to be good. Rightly or wrongly, having good, robust logic bothers me more than whether the character looks nice. Not many posters on this thread seem to share my concerns logically. Anyone (OK, not anyone) can design a lovely a character, a lot of the proposed Long s's in this thread look great, but don't they amount to nothing more than follies if the logic that conceived them was flawed?

Question 1 for me is "So what does an uppercase long s (if it actually ought to be made to live and breathe at all) really look like?" If one answers this then one might be on the way to defining the look of the actual ligature in its entirety. If you can't answer this question, maybe don't force a ligature to exist.

k.l.'s picture

Reading things into Mr Berlow's comment, at the risk of misreading it ...

It doesn't look like Mr Berlow said that this or that uppercase form of long-s is right or wrong, but rather that there is and never was an uppercase long-s so it's a bit of an idle game.

As regards the logic part ...

Ralf, earlier: But all of Marks designs use the skeleton of the B with a "strange thing" going on at the bottom. [...] Tim's designs are not the ones I would favour, but they clearly solve the problem. I am not tempted at all to read GROBES.

Amusingly Mr Simonson's shapes fit much nicer into uppercase letters' visual laws while Tim's, nice as they are, suffer from having four rather than the usual three horizontal elements and a 3/4 S which -- whether Mr Berlow qualified for judging by not submitting his own Eszett to the jury or not -- is a little alien. Consider Mr Simonson's versions as long-s reminder (the stem) plus a diacritic mark similar to connected semicolon, which construction has some resemblance to early "eszetts".
Still I cannot help considering the uppercase Eszett as a bastard. The lowercase version already contains a letterforms that by itself does not exist any more (long-s), and any uppercase version needs to come up with an uppercase long-s which never existed and which must be unintelligent by design. The only solution is, more or less, to uppercasify the lowercase version. But this solution actually means violating the typographic rule to not use an eszett in all-caps setting, but hiding this a bit by dressing up what structurally is a lowercase letter so it looks more like an uppercase. "Uppercase Eszett" itself is logically not sound.

Nick Shinn's picture

...the look of the actual ligature...

Two models emerged for the "double u", one with kissing v's, one with overlapping v's. Both legitimate.
There are variants based on the shape of "u" as well as "v".

Look at the different design solutions that have been proffered for the Euro.

It's wide open, that's why it's so exciting to take a crack at.

hrant's picture

> “Did they actually need ligaturing?”

What is "need"?
I believe that ligatures are good for you.

> I cannot help considering the uppercase Eszett as a bastard.

All the lc glyphs are bastards (as Cassandre duly noted).
But they rule the earth now, so we better work with them.

The interesting twist with the cap eszet is that -apparently- no
words start with it. So its harmony with other caps seems to be
more important than with the lc. (This is also makes it quite
unnerving that it took so long to take it seriously.)


guifa's picture

hrant: I believe that ligatures are good for you.

While I'm in total agreement, I could have sworn you were in general against the use of ligatures. Maybe I'm just remembering some old threads wrong.

«El futuro es una línea tan fina que apenas nos damos cuenta de pintarla nosotros mismos». (La Luz Oscura, por Javier Guerrero)

hrant's picture

I've been liking them for at least four years,
but maybe before that I had said something that
I no longer agree with. Or maybe I had said that
ligatures that are only there to make text pretty
are no good. I mean text text.


Mark Simonson's picture

They have imaginary numbers in mathematics, indicated with an italic lowercase i. Since the long (cap) S is an imaginary character, we could do the same and write the imaginary long S as iS. Following this logic, the capital Eszett would then be written iSS. However, it would be a bit difficult not to read the i as a letter, so we could just put the dot from the i over the first S to indicate the imaginary long S. The cap Eszett would then be written SS, with a dot over the first S.

cuttlefish's picture

Are there any words in ANY languages that begin with "ss"? Would these languages benefit from a spelling reform to replace that digraph with eszett (and by extension Eszett)?

Just throwing some more ideas against the wall to see if anyone salutes.

hrant's picture

Mark, that doesn't sound anything like you! :-)

Jason, that's a major long-shot, but
certainly the presence of a cap eszet
makes it much more likely! The thing
is, a lowercase eszet would probably get
spliced in first.


Florian Hardwig's picture

Are there any words in ANY languages that begin with “ss”? Would these languages benefit from a spelling reform

Don’t know of ‘Ss’, but Hungarian has a lot of words starting with ‘Sz’. According to an entry in the Decode Unicode Wiki, some Hungarians like to replace ‘sz’ with ‘ß’ in text messages – because so they can save a character. To me, that doesn’t sound very plausible, in times of T9 dictionaries.
Same is said about the Swiss: SMS efficiency made them raise the ‘ß’ from the dead. Again, I’d say this mainly has to do with the built-in dictionaries: When typing ‘Grus…’, the dictionary proposes ‘Gruß’ [greeting] – which is gladly accepted.
Just leetspeak.

nina's picture

Actually in Switzerland, among young people, the ß is even widely replaced for ss in text messages where it doesn't make sense at all from a grammatical / orthographical perspective; I assume it has less to do with T9 than with, like you said, saving a character. Stuff like "wißen" for "wissen" (to know), which has never been written like this; makes me cringe, but it does appear to be a new use for the character.

nina's picture

Add to that: There still aren't any words in German that begin with either ß or ss. Which is, the way I understand it, part of the reason why the ß did not previously exist as a capital letter.

ralf h.'s picture

Still I cannot help considering the uppercase Eszett as a bastard. ... any uppercase version needs to come up with an uppercase long-s which never existed and which must be unintelligent by design.

This sort of logic is common, but I don't think it is right. Are German umlauts bastards because they are originally build from a lowercase letter (e) placed on top of a capital letter? Is an A a wrong character because the ox was turned upside down? This sort of logic, that is based on a character's history or the history of parts of the letters doesn't lead anywhere.
Letters are nothing but tools and they should be judged according to this and nothing else. Capital Umlauts were created because there was a need for them. And now is the time for the capital sharp S. We just need to figure out the right shape(s) of this tool. And this design process should be based on the question of what works best for the user/reader, not on the history of Eszett (which is unclear till today anyhow).

k.l.'s picture

Hi Ralf, I was addressing Nick Job's post which spoke about logic.
I do not understand what you mean by "this sort of logic". If you design a letter, you need some kind of logic to make sure that the letter can be identified as that. A letter must fit into the context of other letters in terms of particular design style and, more important, fit into the rest of the alphabet in terms of structure (construction) and features -- to make sure a tone [mere shape on a page] can be identified as a token [instantiation] of a type [this or that character]. Beware of people who (want to) make history without knowing it. Look at IPA monsters to see what happens if history -- repertory of signs and marks that already exist -- is left aside: "tool approach" at its best.
Problem with uppercase Eszett is that either one tries to make it visually sound, then it looks like Mark Simonson's or maybe like a lowercase eszett tweaked to look more like an uppercase, or structurally correct, which option is not even available because there is no uppercase long-s.

hrant's picture

Florian, that's wonderful to hear.
Human needs should trump modernist bureaucrats all the time!

> part of the reason why the ß did not previously exist as a capital letter.

So Germans never set all-caps? :-)

> This sort of logic, that is based on a character’s history or
> the history of parts of the letters doesn’t lead anywhere.

Heartily agreed.

> you need some kind of logic to make sure
> that the letter can be identified as that.

Yes, but historical forms are circumstantial. What matters is what's in use (and what should be in use). Historical fidelity is generally promoted by people who simply enjoy history; but users don't need their enjoyment.

> Beware of people who (want to) make history without knowing it.

I would say that ignorance of history isn't the problem there.
The problem is simply poor craft (which does include observation however).

History leads us to where we are, but it's never really with us.


Syndicate content Syndicate content