Replace ß by smallcap eszett or smallcap ss?

Jasper de Waard

I would like to get responses from germans especially. An option I'm considering would be to include the smallcap eszett as a discretionary ligature.

Igor Freiberger

Jasper, I am not German but did ask a similar question some time ago in Typophile. And Frode, among others, did that here in TypeDrawers. The answer from German-speaking fellows and also specialists like John Hudson is do not include OT substitutions with Eszett and SS.

Putting simply, Eszett shouldn't replace SS (or vice-versa) because they don't have exactly the same linguistic value. Besides this, there are differences between German and Swiss German.

You can read details here, here and here (TD). And also here, here and here (TPHL). Technical information in these threads is highly valuable.

Johannes Neumeier

Aside from being the sensible version, albeit historically not practiced, the capital eszett is now being recommended by german language authorities for official inclusion in the german alphabet (page 28 here, in german).

Personally, I like it, it's a funky glyph. Any users uncomfortable with the new usage may explicitly set a double capital s. As type designer I'd advocate the soon to be amended norm and include and use the capital eszett. Caps or small caps do not make a difference in this regard.

Note that like with other ligatures in german language it is advocated not to be set for combined words' boundries, so in this case: LOSSCHIEẞEN (to dash/shoot off) with two explicit capital S between the words "los" and "schießen", but capital ß within the word "schießen". As such, german typographers can either ignore this finesse, outright disagree with it or end up having to set any and all ligatures by hand.

Nick Shinn

The smallcap glyph should function in the <c2sc> feature:

sub uni1E9E by uni1E9E.c2sc ;

(or appropriate classes)

It’s not a ligature, it’s a character. If uni1E9E has been set as text, that is a capital letter which should be rendered as a small cap, like all the other capitals, when the <c2sc> feature is applied. That’s all.

Jasper de Waard

Okay, I think I may not have been clear enough. The question is not about whether or not to include the capital/smallcap eszett, or whether to replace SS by the capital eszett.

The question is this: if a type user just typed fußball and makes that word smallcaps, would he/she probably want the ß to be converted to the smallcap eszett, or to (I think more conventional) smallcap ss. And if the latter is the case, how would I make the smallcap eszett accesible?

Igor Freiberger

Replace German double lowercase s with SS when applying all caps, small caps or petite caps formatting. From the links above, you have exactly this issue discussed in detail (starting with Chris Lozos' comment).

Hrant Հրանդ Փափազեան Papazian

Before you take somebody's opinion about the cap eszett seriously, check whether they happen to be a reformophobe who wants writing systems to be moldy museum pieces.

John Hudson

The question is this: if a type user just typed fußball and makes that word smallcaps, would he/she probably want the ß to be converted to the smallcap eszett, or to (I think more conventional) smallcap ss. And if the latter is the case, how would I make the smallcap eszett accesible?

The way to think of this is as follows:

1. The standard case mapping of ß is SS. Therefore, it follows that correct smallcap representation of ß is a sequence of two /s.smcp/ glyphs.

2. The alternative uppercase form of ß is the independent (not case mapped) character U+1E9E. The correct smallcap representation of U+1E9E is a single /uni1E9E.c2sc/ glyph.

3. Crucially, the character ß and the letter sequence ss are not orthographically equivalent; therefore, one should avoid glyph-level transformations that obscure the distinction.

4. It is up to users to determine whether they want to display all-caps (and hence caps-to-smallcaps) using the standard SS casing of ß or the alternative U+1E9E character. This is a character-level, text encoding decision, and fonts should not try to second-guess what the user wants.

Hrant Հրանդ Փափազեան Papazian

Type designers are not mere vehicles of typographic culture, but are the most qualified to guide its progress, and are hence duty-bound to do so. Following standards blindly can do more harm than good in the long term.

"All progress depends on the unreasonable man."
— George Bernard Shaw

Johannes Neumeier

Jasper de Waard said:

The question is this: if a type user just typed fußball and makes that word smallcaps, would he/she probably want the ß to be converted to the smallcap eszett, or to (I think more conventional) smallcap ss. And if the latter is the case, how would I make the smallcap eszett accesible?

ß becomes ẞ. Writing a ß to a ss is a convention that stems from computer systems and typewriters lacking the proper glyph. Although commonly understood, I don't think it is a correct substitution, as argued for the inclusion of the uppercase character to the official alphabet. If the user explicitly wants to keep the dated way of writing, they can explicitly type a double S.
Imagine someone typesetting a reference list and applying small caps to a style for some highlighting. An author Max Heß would turn into MAX HESS and the publisher from Straßkirchen would be located in STRASSKIRCHEN, both of which are semantically different.

Chris Lozos

In cases like this, where there is yet to be universal acceptance of one single method of usage, it is best to allow the user options instead of forcing our own preferred method on them.  At some point, German speakers may chose to make a decided change.  Until then, I prefer to allow for the user to decide of his or her own volition.

Jürgen Weltin

Johannes Neumeier said:

ß becomes ẞ. Writing a ß to a ss is a convention that stems from computer systems and typewriters lacking the proper glyph. Although commonly understood, I don't think it is a correct substitution, as argued for the inclusion of the uppercase character to the official alphabet. If the user explicitly wants to keep the dated way of writing, they can explicitly type a double S.

Jürgen Weltin

Sorry, this is completely wrong! Is this how history is written anew?
This has absolutely nothing to do with technical conventions. PLAIN WRONG!
Read the German DUDEN for reference first. German orthography states to put ß to SS in all caps writing. The ß existed long before typewriters, folks.

Jürgen Weltin

Plus, the ß to SS convention existed also long before typewriters.

Christian Thalmann

By that logic, shouldn't we also capitalize ä to Ae?

I'm all for using ẞ (even though I'm Swiss!).

Jürgen Weltin

This is absolutely not about liking or disliking, but about historical facts and common (German) understanding about orthography of the German language. If you want to follow logic, Mr Thalmann, then think about whether there is a handwritten origin of a capital ß.

Christian Thalmann

A historical practice is not necessarily a logical practice. The practice of writing capital ä as Ä overruled Ae since it's more consistent. Similarly, ẞ should now overrule SS when capitalizing ß, since it's more consistent. We have no obligation to stick to past mistakes.

Johannes Neumeier

Yes, I expressed that incorrectly and didn't intend to cut history short; I didn't mean it to sound like typewriters or computer keyboards introduced the double S. One of the ways the double S capitalisation of ß as a convention has been further strengthened is due to this technical limitation.

The Duden entry for the ß does acknowledge this for the lowercase instance:

Regel 160:

1. Fehlt das ß auf der Tastatur eines Computers oder einer Schreibmaschine, schreibt man dafür ss. In der Schweiz kann das ß generell durch ss ersetzt werden <§ 25 E2>.

Strasse (statt: Straße), aussen (statt: außen), Fussball (statt: Fußball)

2. Auch bei Verwendung von Großbuchstaben steht SS für ß <§ 25 E3>.

STRASSE, AUSSEN, FUSSBALL

3. Bei der Worttrennung wird dieses ss wie andere Doppelkonsonanten behandelt <§ 108>.

•   Stras-se, aus-sen, Fuss-ball

In Dokumenten kann bei Namen aus Gründen der Eindeutigkeit auch bei Großbuchstaben das ß verwendet werden.

Für den im internationalen Standard-Zeichensatz „Unicode" (ISO/IEC 10646) verzeichneten Großbuchstaben für das ß gibt es derzeit noch keine allgemein verwendete Schriftform. Er ist nicht Gegenstand der amtlichen Rechtschreibregelung.

Short translation. ss for ß when no ß available, SS for ß, but you can also use ß within caps, and the capital ß is currently not part of the official orthographical ruleset. Hence my link above about the Rechtschreibrat's recommendation for its inclusion, which is, after the unicode inclusion, the most current development to my knowledge.

Yet also on a special page the Duden acknowledges the capital eszett:

So ist die Forderung nach einem großen Eszett durchaus verständlich. Die internationale Organisation für Normung (ISO) hat nun – in den internationalen Zeichensätzen ISO-10646 und Unicode 5.1 – ein Zeichen für das große Eszett festgeschrieben: eine wichtige Neuerung, besonders für Buch- und Zeitungsverlage, aber eine, die nicht Bestandteil der amtlichen Rechtschreibung ist. 

 

(image from here)

It's really up to user taste and context. Personally I'd find an OT substitution of ß to SS bewildering. Especially in a context with of person or place names the OT rule would distort the meaning of the written names, which it never should.

Hrant Հրանդ Փափազեան Papazian

Didn't the Duden recently accept the use of cap eszett? But anyway it's just another set of guidelines, not gospel. And handwriting origins? Gimme a break.

Jürgen Weltin

As usual, these debates are getting nowhere :-(
It’s not about being for or against a cap-ß. It’s good to have one in typography. And yes, Duden accepts it now. So anyone can do as he pleases or as it is right in context. Famous word, for instance, would be Masse/Maße (first can be mass, crowd, mixture, etc./second is plural for measure). Note that in most cases the textual context won’t lead to confusion. Another thing is the spelling of names.
All i’m saying is, if there had been a handwritten capital-ß, there most probably would have been a rule to capitalize ß to capital-ß. But, ß evolved as a ligature of long-s and s. And there are no ligatures in German capitals, neither in a bunch of other languages.

Hrant Հրանդ Փափազեան Papazian

Such debates might not sway many of the overt participants, but they are invaluable to those who enjoy the luxury of being yet undecided. In fact to me it is a duty to engage in them.

Again, any handwriting [lack of] precedent is a red herring.

John Hudson

Following standards blindly can do more harm than good in the long term.

Following anything blindly will eventually end in walking into a wall or off a cliff; however...

Standards are what enable people to make choices with predictable outcomes, and a flexible standard — which is what the German orthographic standard and Unicode encoding standard now provide — enables those choices to express personal preference, regional tendencies, data-appropriate encoding, etc.

Blindly not following the standard undermines these choices by making the outcomes unpredictable, so it is absolutely not useful for font makers make arbitrary personal choices about how to map between characters and glyphs in layout features.

Nick Shinn

True, John, which makes it important to distinguish situations where there are standards from those where there aren’t. 

Hrant Հրանդ Փափազեան Papazian

> Blindly not following the standard undermines these choices

Agreed, never blindly.
We're essentially professing [qualitatively] different levels of respect for standards.

John Hudson

We're essentially professing [qualitatively] different levels of respect for standards.

I don't think it's about respect for standards per se, but about figuring how to make things that are useful, and where standards fit into that. This is why it is important to understand what a standard is and how it functions — especially in the case of a technical standard, i.e. an implementation standard. There have been times, over the years, when I've rejected a standard and encouraged an alternative implementation, simply because the standard isn't useful. An obvious case is normalisation sorting of Hebrew marks based on Unicode canonical combining class assignments: that's a standard, and one that — due to stability agreements — cannot be changed, but it's simply broken, worse than useless. So about twelve years ago I wrote my own normalisation recommendations, in consultation with Hebrew text processing experts, and was recently pleased to discover that it's been taken up by web browsers.

In the case of the eszett, there is an official orthographic standard, which is the domain of certain organisations and regional ministries in Germany. This is something more than 'just another set of guidelines', because it has legal force in numerous and significant areas such as education, official documentation, government databases, etc. Then there is the implementation standard, which for our purposes as font makers is Unicode, which reflects the orthographic standard as it has stood for a very long time, treating the standard case mapping of ß to SS, while also providing for an alternative encoding of ẞ without case mapping at the Unicode level. This latter point is important: the fact that Unicode specifies the standard case mapping of ß to SS — which also cannot be changed due to stability agreements (in software implementation standards, stability is often key to getting a standard adopted) — does not mean that systems cannot implement alternative mechanisms. So, for instance, if one were building a database of personal names in Germany, in which entries might be entered or displayed in all-caps — leaving aside whether that in itself is a good typographic decision — one might choose to implement mapping of ss to SS and ß to ẞ in order to maintain spelling distinctions during case transformations. And that's fine, and there's the possibility that such mechanisms will become common and even become more widespread in use than the standard ß to SS mapping. And maybe they'll even be enshrined in a German national implementation standard at some point, at which time software will need to decide and declare, in this particular instance of case mapping for eszett, which standard is being followed.

In all of this, the task of font makers is to understand what the standards are and how they function, so as to be able to make fonts that provide predictable outcomes when users make text encoding decisions. Whatever happens with eszett case mapping in the long run in Germany — even if Germans were eventually to follow the Swiss into an eszettless future —, what we should be very clear about is that the decision about case transformation of ß is a character-level decision. This means that glyph-level processing, such as smallcap features, needs to reflect the character-level decisions that a user might make.

Jasper de Waard

My conclusion so far:

c2sc feature:     sub Germandbls by germandbls.sc;

smcp feature:    sub germandbls by s.sc s.sc;

Otherwise, no mention of germandbls.sc/Germandbls is strictly needed.

But, and this brings me to my original question: IF a designer wants to use either germandbls.sc or Germandbls (note the cap G), they have to do so through the glyph panel, which seems cumbersome. Could it be helpful to include both as discretionary ligatures? Or would that just be confusing? I mean, in essence, it kind of IS a ligature, and it IS discretionary, right?

Chris Lozos

perhaps better as stylistic set, Jasper.

Jasper de Waard

@Chris Lozos I keep hearing that, and I'm inclined to just go along with it, but I still don't really get why.

Christian Thalmann

As for «predictability»: I would argue that switching fonts will generally change the appearance of a given text more drastically than exchanging SS with ẞ... and while using SS for ẞ can cause loss of information, sticking to ẞ never does. After all, if the user typed ß before capitalization, they meant ß rather than ss. Anyone who can read ß can also read ẞ without hesitation — if not, the ẞ is badly designed.

But then again, I'm not a typographer.

Hrant Հրանդ Փափազեան Papazian

@John Hudson Thank you for the detailed reply.

> the task of font makers is to understand what the standards are and how they function, so as to be able to make fonts that provide predictable outcomes when users make text encoding decisions.

This is where we disagree. When it comes to duly "figuring how to make things that are useful" to me predictability is not untouchable. People learn change, and type designers can direct culture.

Jameson R Spires

I type things. I then realize I'm being trolled and I delete things.