superscript ordinals

[Jaakov]

Dear all:

Leaving aside historical considerations, ease of typesetting, space
limits, and usage frequency, is there any other reason for using or not
using superscript ordinals in English?

I.e., I'm asking for aesthetic, technical, and psychological reasons for
why reading or writing
1^{st}, 2^{nd}, 3^{rd}, 4^{th}, ..., n^{th}, ...
(where ^{xy} means that xy is typeset as a superscript)
should be more or less pleasant, rewarding, or just quick than reading
1st, 2nd, 3rd, 4th, ..., nth, ...

Thanks in advance,

Jaakov

--- news://freenews.netfront.net/ - complaints: ne...@netfront.net ---

[tlvp]

In bookbinding, it's just the convention: "quarto" or "4^o", not "4o";
"duodecimo" or "12^o", not "12o"; "octavo" or "8^o", not "8o".

In mailing addresses (in USA, at least) it's writer's choice: 500 5^th Ave.
or 500 5th Ave. or 500 Fifth Ave., all are equally acceptable.

HTH. Cheers, -- tlvp
--
Avant de repondre, jeter la poubelle, SVP.

[Jukka K. Korpela]

14.4.2016, 1:55, Jaakov wrote:

> Leaving aside historical considerations, ease of typesetting, space
> limits, and usage frequency, is there any other reason for using or not
> using superscript ordinals in English?

I wonder whether the question is about fonts. It might be off-topic in
comp.fonts (and on-topic in some group or forum devoted to discussing
the English language).

> I.e., I'm asking for aesthetic, technical, and psychological reasons for
> why reading or writing
> 1^{st}, 2^{nd}, 3^{rd}, 4^{th}, ..., n^{th}, ...
> (where ^{xy} means that xy is typeset as a superscript)
> should be more or less pleasant, rewarding, or just quick than reading
> 1st, 2nd, 3rd, 4th, ..., nth, ...

This does not sound like a question about fonts.

Fonts are relevant in some considerations, though, but mainly in the
technical side. Not all fonts contain glyphs for superscripts (either as
primary glyphs for Unicode characters with SUPERSCRIPT in their name or
as secondary, selectable glyphs for normal digits or letters). In fact,
I think glyphs for superscript letters are relatively rare.

Note that “typeset as a superscript” is a vague expression. It might
refer to a superscript glyph designed by the typographer that created
the font, or it might refer to a normal glyph in reduced size and in
elevated position. The latter produces poorer (often much poorer)
results typographically. It is what you get e.g. if you type “1st” in
Microsoft word and you have allowed the automatic change that makes “st”
a superscript – of a kind – in that context.

--
Yucca, http://www.cs.tut.fi/~jkorpela/

[Jaakov]

On 14.04.2016 14:58, Jukka K. Korpela wrote:
> 14.4.2016, 1:55, Jaakov wrote:
>
>> Leaving aside historical considerations, ease of typesetting, space
>> limits, and usage frequency, is there any other reason for using or not
>> using superscript ordinals in English?
>
> I wonder whether the question is about fonts. It might be off-topic in
> comp.fonts (and on-topic in some group or forum devoted to discussing
> the English language).
>
>> I.e., I'm asking for aesthetic, technical, and psychological reasons for
>> why reading or writing
>> 1^{st}, 2^{nd}, 3^{rd}, 4^{th}, ..., n^{th}, ...
>> (where ^{xy} means that xy is typeset as a superscript)
>> should be more or less pleasant, rewarding, or just quick than reading
>> 1st, 2nd, 3rd, 4th, ..., nth, ...
>
> This does not sound like a question about fonts.
>
> Fonts are relevant in some considerations, though, but mainly in the
> technical side.

That's why I asked here.

> Not all fonts contain glyphs for superscripts (either as
> primary glyphs for Unicode characters with SUPERSCRIPT in their name or
> as secondary, selectable glyphs for normal digits or letters). In fact,
> I think glyphs for superscript letters are relatively rare.
>
> Note that “typeset as a superscript” is a vague expression. It might
> refer to a superscript glyph designed by the typographer that created
> the font, or it might refer to a normal glyph in reduced size and in
> elevated position. The latter produces poorer (often much poorer)
> results typographically. It is what you get e.g. if you type “1st” in
> Microsoft word and you have allowed the automatic change that makes “st”
> a superscript – of a kind – in that context.

I see, thanks!

[Jaakov]

> In bookbinding, it's just the convention: "quarto" or "4^o", not "4o";
> "duodecimo" or "12^o", not "12o"; "octavo" or "8^o", not "8o".
>
> In mailing addresses (in USA, at least) it's writer's choice: 500 5^th Ave.
> or 500 5th Ave. or 500 Fifth Ave., all are equally acceptable.

Thanks!!!

[tlvp]

I rather suspect Jaakov may have been thinking particularly about the
superscript ^a and ^o (used in Spanish? or Italian?) serving as feminine
and masculine ordinal-number endings (the HTML entities "feminine ordinal",
ª or ª and "masculine ordinal", º or º ), and their
analogue in French, superscript ^ême or the like.

Those, like some of the examples in my first response, really do prefer to
be set as superscripts.

As to the technical means by which to produce them -- via HTML <sup> tags,
as dedicated entities, or by other means entirely, that is another question
entirely, regarding which I abstain from offering any opinion.

But I find it worthwhile to be mindful that different instances of the
original question may well deserve different answers :-) . Cheers, -- tlvp

[Jukka K. Korpela]

16.4.2016, 6:44, tlvp wrote:

> On Thu, 14 Apr 2016 15:58:15 +0300, Jukka K. Korpela wrote:
>
>> 14.4.2016, 1:55, Jaakov wrote:
>>
>>> Leaving aside historical considerations, ease of typesetting, space
>>> limits, and usage frequency, is there any other reason for using or not
>>> using superscript ordinals in English?
– –

> I rather suspect Jaakov may have been thinking particularly about the
> superscript ^a and ^o (used in Spanish? or Italian?) serving as feminine
> and masculine ordinal-number endings

Maybe, but the specifically wrote about superscript ordinals in English.

The feminine and masculine ordinal indicators “ª” and “º” (used in
Spanish and Catalan, but normally not in Italian) are rather different
beasts, in font design, too. If a font implements superscript glyphs for
writing “1st”, “2nd”, etc. with letters as superscripts, there is no
excuse for using anything but a normal letter in superscript style. In
contrast, “ª” and “º” may be, and often are, implemented in a different
style, with an underline under the letter, reflecting the traditions.

> (the HTML entities "feminine ordinal",
> &#170; or &ordf; and "masculine ordinal", &#186; or &ordm; ),

There is no reason (and never really was, as these as Latin 1
characters) to use numeric or named character references for these
characters, except that you might find them easier to type than to tweak
your keyboard driver to let you type them comfortably.

> and their
> analogue in French, superscript ^ême or the like.

I think you mean superscript ème and the like. They are typographically
comparable to using superscript rendering for English ordinals like “1st”.

> As to the technical means by which to produce them -- via HTML <sup> tags,
> as dedicated entities, or by other means entirely, that is another question
> entirely, regarding which I abstain from offering any opinion.

But that’s really the only font-related question here. The answer is
simple: anything you get with <sup> or with comparable tools, like
superscript style in word processors, is inferior to using superscript
glyphs designed by a typographer. It is so inferior that it should not
be used.

The use of <sup> or other coarse tools for *mathematical* superscripts
is an entirely different issue. I recently read in a local magazine a a
story that mentioned Fermat’s great theorem using the formula xn + yn =
xn (without using italic or superscripts). It was just plain wrong of
course, and any method that makes the n’s superscripts in any sense is
infinitely better.

The difference is that in “1st”, superscripting is a stylistic issue
that does not change the meaning at all, whereas in “xn”, it changes the
whole meaning (from a product to a power).

--
Yucca, http://www.cs.tut.fi/~jkorpela/

[Peter Flynn]

On 04/13/2016 11:55 PM, Jaakov wrote:
> Dear all:
>
> Leaving aside historical considerations, ease of typesetting, space
> limits, and usage frequency, is there any other reason for using or not
> using superscript ordinals in English?

Yes, see my article on this in TUGboat at
https://www.tug.org/TUGboat/tb26-3/tb84inn.pdf
and my later comments in _Formatting Information_ at
http://latex.silmaril.ie/formattinginformation/specials.html#superscript

> I.e., I'm asking for aesthetic, technical, and psychological reasons for why reading or writing
> 1^{st}, 2^{nd}, 3^{rd}, 4^{th}, ..., n^{th}, ...
> (where ^{xy} means that xy is typeset as a superscript)
> should be more or less pleasant, rewarding, or just quick than reading
> 1st, 2nd, 3rd, 4th, ..., nth, ...

It's aesthetics. The only reason people do it nowadays is because that's
what Word does. If Microsoft hadn't reintroduced it, it would have
remained a Victorian or at least post-Edwardian foible.

Ordinal superscripts have a long and honourable history deriving from
their use in mediaeval Latin. This was carried over into early printing,
and was commonplace until the dominance of the typewriter in the early
20th century, when the inability of most typewriters to produce a
superscript except by manual adjustment led to its dmise in print.

It remains the normal way of indicating ordinality in many non-English
Latin-alphabet languages, but there it is usually a single character
(the raised o or a as tlvp has shown, wrt printing and binding). I think
English is the only language where two letters are used. Mathematics is
of course a completely separate issue.

Done with care, a superscript st, nd, or th can look fine; but most of
the time nowadays (in English) it just looks hack, and can sometimes
indicate that the author is a Word user and may have no clue.

///Peter

[Jukka K. Korpela]

17.4.2016, 0:37, Peter Flynn wrote:

> On 04/13/2016 11:55 PM, Jaakov wrote:
>> Dear all:
>>
>> Leaving aside historical considerations, ease of typesetting, space
>> limits, and usage frequency, is there any other reason for using or not
>> using superscript ordinals in English?
>
> Yes,

“Yes” is a rather short and even incomplete answer to a question like this.

> see my article on this in TUGboat at
> https://www.tug.org/TUGboat/tb26-3/tb84inn.pdf
> and my later comments in _Formatting Information_ at
> http://latex.silmaril.ie/formattinginformation/specials.html#superscript

These are longer but indirect answers. It seems that your opinion is
that superscript ordinals should not be used since they are “a
historical relic of Victorian and earlier typography”. I don’t quite see
what type of argument this is. But as I wrote earlier, the question as
such looks off-topic. Yet, you continue, in the latter document: “Unless
done with skill and finesse by a typographer, they are usually ugly and
unnecessary, and are never used in modern professional typesetting.”
This may make the question on-topic, at least in part.

Your notes revolve around the behavior of Microsoft Word that I
mentioned earlier. It is not really a relic of anything. Rather, it
imitates typographic superscripts using reduced font size and elevated
position for normal glyphs.

If proper superscript glyphs are used, e.g. via OpenType feature sups,
your arguments don’t seem to apply. It might be difficult to find fonts
that have such glyphs and suit your needs in other ways, too, and it can
be difficult to make software use those glyphs. But this is not what was
asked (though it would have been on-topic here).

> Ordinal superscripts have a long and honourable history deriving from
> their use in mediaeval Latin. This was carried over into early printing,
> and was commonplace until the dominance of the typewriter in the early
> 20th century, when the inability of most typewriters to produce a
> superscript except by manual adjustment led to its dmise in print.

Similar things have happened to other glyphs and rendering features,
too. Should we also refrain from using en dashes (–) and em dashes (—)
and use Ascii hyphens (-) instead? ☺

> It remains the normal way of indicating ordinality in many non-English
> Latin-alphabet languages, but there it is usually a single character
> (the raised o or a as tlvp has shown, wrt printing and binding). I think
> English is the only language where two letters are used.

It is common in French to write e.g. 1er and 2ème, with the letters as
superscripts. (What is correct French is a different issue, and
apparently a debated one.) The use of an accented letter like “è” makes
it even more difficult to find a font with a superscript glyph.

> Done with care, a superscript st, nd, or th can look fine; but most of
> the time nowadays (in English) it just looks hack, and can sometimes
> indicate that the author is a Word user and may have no clue.

If you draw glyphs yourself, then it requires care. But in most
situations it is a matter of finding out whether there is a feasible way
to use superscript letters and what that might be, rather than designing
how to draw a superscript glyph. It’s challenging, but a matter of
knowing what to do rather than glyph design.

And it can be a difficult challenge, judging from the observations that
in the vast majority of cases where people use superscript two in an
expression like m², they don’t use a superscript glyph but a
superscripting command in Word, or something similar. (This produces a
rather poor result, especially in large font sizes: the “2” is too small
and too thin.) Yet, you can get a superscript two glyph very easily by
using the superscript two *character*.

--
Yucca, http://www.cs.tut.fi/~jkorpela/

[Peter Flynn]

On 04/17/2016 02:49 PM, Jukka K. Korpela wrote:
> 17.4.2016, 0:37, Peter Flynn wrote:
>
>> On 04/13/2016 11:55 PM, Jaakov wrote:
>>> Dear all:
>>>
>>> Leaving aside historical considerations, ease of typesetting, space
>>> limits, and usage frequency, is there any other reason for using or not
>>> using superscript ordinals in English?
>>
>> Yes,
>
> “Yes” is a rather short and even incomplete answer to a question like this.

It was intended more as an introduction than a complete answer...

>> see my article on this in TUGboat at
>> https://www.tug.org/TUGboat/tb26-3/tb84inn.pdf
>> and my later comments in _Formatting Information_ at
>> http://latex.silmaril.ie/formattinginformation/specials.html#superscript
>
> These are longer but indirect answers. It seems that your opinion is
> that superscript ordinals should not be used since they are “a
> historical relic of Victorian and earlier typography”. I don’t quite see
> what type of argument this is.

It's an argument for moving onwards. They are simply outdated, and while
it's fun to preserve them for historical purposes, they are not regarded
as useful by any designers I know.

> But as I wrote earlier, the question as
> such looks off-topic. Yet, you continue, in the latter document: “Unless
> done with skill and finesse by a typographer, they are usually ugly and
> unnecessary, and are never used in modern professional typesetting.”
> This may make the question on-topic, at least in part.

If I have misled the readers, I apologise.

> Your notes revolve around the behavior of Microsoft Word that I
> mentioned earlier. It is not really a relic of anything. Rather, it
> imitates typographic superscripts using reduced font size and elevated
> position for normal glyphs.

Microsoft's behaviour is indeed not a relic of anything; on the contrary
it is a relative novelty introduced (I think) with Word in Windows 95.

> If proper superscript glyphs are used, e.g. via OpenType feature sups,
> your arguments don’t seem to apply. It might be difficult to find fonts
> that have such glyphs and suit your needs in other ways, too, and it can
> be difficult to make software use those glyphs. But this is not what was
> asked (though it would have been on-topic here).

I'm not familiar with typefaces providing s/t/n/d/r/h as special glyphs
for superscript use -- it's an interesting idea. But my argument is
against superscript ordinals in English text, regardless of the glyphs
used to represent them.

> Similar things have happened to other glyphs and rendering features,
> too. Should we also refrain from using en dashes (–) and em dashes (—)
> and use Ascii hyphens (-) instead? ☺

No, they have remained standard practice all the time. The point you
miss about superscript ordinals is that they died out and were then
revived by Word.

> It is common in French to write e.g. 1er and 2ème, with the letters as
> superscripts. (What is correct French is a different issue, and
> apparently a debated one.) The use of an accented letter like “è” makes
> it even more difficult to find a font with a superscript glyph.

Very. The codepoints exist in Unicode, but I doubt very many
type-designers would spend time including them.

One possible trick is to use a Demibold, if the typeface provides it.
This compensates visually for the otherwise "too-slim" look of Regular
or Medium fonts at very small sizes. It's also possible, if tedious, to
thicken the stroke values manually.

> [...] It’s challenging, but a matter of

> knowing what to do rather than glyph design.

I think this is the key to so much design: knowing what is possible.

> And it can be a difficult challenge, judging from the observations that
> in the vast majority of cases where people use superscript two in an
> expression like m², they don’t use a superscript glyph but a
> superscripting command in Word, or something similar. (This produces a
> rather poor result, especially in large font sizes: the “2” is too small
> and too thin.) Yet, you can get a superscript two glyph very easily by
> using the superscript two *character*.

Exactly. Slowly, very slowly, software is now handling UTF-8 characters,
where previously people were stuck with a huge number of dead-end
encodings. It will take a long time for the abuses to die out.

///Peter

[Jaakov]

On 04/17/2016 09:26 PM, Peter Flynn wrote:
> On 04/17/2016 02:49 PM, Jukka K. Korpela wrote:
>> 17.4.2016, 0:37, Peter Flynn wrote:
>>
>>> On 04/13/2016 11:55 PM, Jaakov wrote:
>>>> Dear all:
>>>>
>>>> Leaving aside historical considerations, ease of typesetting, space
>>>> limits, and usage frequency, is there any other reason for using or not
>>>> using superscript ordinals in English?
>>>
>>> Yes,
>>
>> “Yes” is a rather short and even incomplete answer to a question like this.
>
> It was intended more as an introduction than a complete answer...
>
>>> see my article on this in TUGboat at
>>> https://www.tug.org/TUGboat/tb26-3/tb84inn.pdf
>>> and my later comments in _Formatting Information_ at
>>> http://latex.silmaril.ie/formattinginformation/specials.html#superscript
>>
>> These are longer but indirect answers. It seems that your opinion is
>> that superscript ordinals should not be used since they are “a
>> historical relic of Victorian and earlier typography”. I don’t quite see
>> what type of argument this is.
>
> It's an argument for moving onwards. They are simply outdated, and while
> it's fun to preserve them for historical purposes, they are not regarded
> as useful by any designers I know.

Well, "outdated", "old", "dead" are historical reasons, which I excluded
specifically in my question.

In fact, there is nothing wrong in

- keeping the outdated features dead if there are good reasons for this
and
- reviving the dead features if there are other good reasons for this.

MSWord made a choice toward reviving. So far I did not hear good reasons
for why MSWord did so. (You can turn this default autocorrection off if
you wish.)

But thank you both for driving my attention to special glyphs for the
superscript ordinals. If I use them at all, I now have a choice of how
to make them look beautiful.

[Jukka K. Korpela]

17.4.2016, 22:26, Peter Flynn wrote:

> They [superscript ordinals] are simply outdated, and while

> it's fun to preserve them for historical purposes, they are not regarded
> as useful by any designers I know.

My half-educated guess is that most designers of text formatting do not
simply know how to produce superscript letters in a typographically
acceptable way, even in situations where it is possible. It is natural
to regard something as not useful if you don’t know how to do it. ☺

> I'm not familiar with typefaces providing s/t/n/d/r/h as special glyphs
> for superscript use -- it's an interesting idea.

They include the “C fonts” Calibri, Cambria, Candara, Consolas,
Constantia, and Corbel as well as Palatino Linotype, as shipped with
Windows 7, and Source Sans Pro, as available from Google Fonts. These
are hardly outdated fonts; rather, it seems that superscript letter
glyphs are making a comeback, at the technical level at least. This does
not imply you should use them; just that you can, in some environments.

> But my argument is
> against superscript ordinals in English text, regardless of the glyphs
> used to represent them.

I still don’t know what that argument is, but I think that part of the
discussion belongs to other formus.

>> Similar things have happened to other glyphs and rendering features,
>> too. Should we also refrain from using en dashes (–) and em dashes (—)
>> and use Ascii hyphens (-) instead? ☺
>
> No, they have remained standard practice all the time.

They didn’t. Typewriters did not have them, early computer keyboards did
not have them, and even in the 1990s it was risky to use them on web
pages. (It’s safe now, provided that you do it right.)

>> It is common in French to write e.g. 1er and 2ème, with the letters as
>> superscripts. (What is correct French is a different issue, and
>> apparently a debated one.) The use of an accented letter like “è” makes
>> it even more difficult to find a font with a superscript glyph.
>
> Very. The codepoints exist in Unicode, but I doubt very many
> type-designers would spend time including them.

Unicode has a code point for superscript “n” (probably due to its use in
math), but not for letters in general. So you cannot write “1st” with
superscript letters as plain text; you can just use markup of some kind
to indicate that superscript glyphs be used for them.

The point, from the perspective of font design and usage, is that you
cannot have, say, superscript “s” as character, but you can have it as a
glyph, to be selected, using special methods, for rendering a normal “s”.

The practical difficulty is that if you apply, say, the OpenType “sups”
feature to a string containing “1st”, you get all the characters,
including the digit “1”, as superscripts, when applicable. Thus, you
need some low-level markup like “1^st” in HTML—but beware that
this creates, in all user agents I suppose, “st” as “fake superscripts”,
i.e. as reduced-size elevated-position glyphs “st”. To get typographic
superscripts, you need to add a style sheet that fixes this (a
nontrivial task, and starting with 1<span class=sup>st</span> actually
makes it essentially simpler).

> One possible trick is to use a Demibold, if the typeface provides it.

It may work in some contexts, but it’s really a hack.

>> And it can be a difficult challenge, judging from the observations that
>> in the vast majority of cases where people use superscript two in an
>> expression like m², they don’t use a superscript glyph but a
>> superscripting command in Word, or something similar. (This produces a
>> rather poor result, especially in large font sizes: the “2” is too small
>> and too thin.) Yet, you can get a superscript two glyph very easily by
>> using the superscript two *character*.
>
> Exactly. Slowly, very slowly, software is now handling UTF-8 characters,
> where previously people were stuck with a huge number of dead-end
> encodings. It will take a long time for the abuses to die out.

I agree, but my point was that people are playing with superscript
commands rather than superscript characters. The superscript two “²”
character is in the Latin 1 repertoire and thus rather safe even in the
pre-UTF-8 world.

--
Yucca, http://www.cs.tut.fi/~jkorpela/

[Jaakov]

>>
>> Exactly. Slowly, very slowly, software is now handling UTF-8 characters,
>> where previously people were stuck with a huge number of dead-end
>> encodings. It will take a long time for the abuses to die out.
>
> I agree, but my point was that people are playing with superscript
> commands rather than superscript characters. The superscript two “²”
> character is in the Latin 1 repertoire and thus rather safe even in the
> pre-UTF-8 world.
>

Remember that Unicode is a mess. It is worth being replaced already now.
Just two examples come to my mind:
- The double-stroke alphabet (as well as certain others) is scattered: ℕ
is not between 𝕄 and 𝕆!
- There is bold sans serif greek but no normal-weight sans-serif greek!

Of course, UTF-8 is a good implementation of Unicode.

[tlvp]

On Sun, 17 Apr 2016 16:49:52 +0300, Jukka K. Korpela wrote:

> It is common in French to write e.g. 1er and 2ème, with the letters as
> superscripts. (What is correct French is a different issue, and
> apparently a debated one.) The use of an accented letter like “è” makes
> it even more difficult to find a font with a superscript glyph.

For another usage example, coming from US publishers' practice rather than
French custom, I'll note the use of what can be described in HTML as
2^e to stand (in book-flog materials or front matter) for "Second
Edition" or "2^nd Ed'n".

That's a very low-density example, but a persistent one if you look in the
right places :-) .

Another is the use of a final ⁿ in place of the final " 'n " in
abbreviations such as Ed'n (for Edition), funct'n (for function), etc.

Again: done rarely, but systematically (in the right places).

[Peter Flynn]

On 04/18/2016 08:19 PM, Jukka K. Korpela wrote:
> 17.4.2016, 22:26, Peter Flynn wrote:
>
>> They [superscript ordinals] are simply outdated, and while
>> it's fun to preserve them for historical purposes, they are not regarded
>> as useful by any designers I know.
>
> My half-educated guess is that most designers of text formatting do not
> simply know how to produce superscript letters in a typographically
> acceptable way, even in situations where it is possible. It is natural
> to regard something as not useful if you don’t know how to do it. ☺

Very true. Even LaTeX, which usually provides good controls for this
kind of thing, just uses math superscript dimensions for *text*
superscripts.

>> I'm not familiar with typefaces providing s/t/n/d/r/h as special glyphs
>> for superscript use -- it's an interesting idea.
>
> They include the “C fonts” Calibri, Cambria, Candara, Consolas,
> Constantia, and Corbel as well as Palatino Linotype, as shipped with
> Windows 7, and Source Sans Pro, as available from Google Fonts. These
> are hardly outdated fonts; rather, it seems that superscript letter
> glyphs are making a comeback, at the technical level at least. This does
> not imply you should use them; just that you can, in some environments.

As a non-Windows users, I had missed these: thanks very much.
Do you happen to know what Unicode code-points they represent?

>> But my argument is
>> against superscript ordinals in English text, regardless of the glyphs
>> used to represent them.
>
> I still don’t know what that argument is, but I think that part of the

> discussion belongs to other forums.

You may have missed that part of the argument: they make the text look
like something out of the 1890s. Which is valid in a reconstruction of a
historical document, but rarely elsewhere.

>>> Similar things have happened to other glyphs and rendering features,
>>> too. Should we also refrain from using en dashes (–) and em dashes (—)
>>> and use Ascii hyphens (-) instead? ☺
>>
>> No, they have remained standard practice all the time.
>
> They didn’t. Typewriters did not have them, early computer keyboards did
> not have them, and even in the 1990s it was risky to use them on web
> pages. (It’s safe now, provided that you do it right.)

I'm sorry: I was talking about typesetting, not typewriters or
computers. Those characters have remained standard and in daily use.

> Unicode has a code point for superscript “n” (probably due to its use in
> math), but not for letters in general. So you cannot write “1st” with
> superscript letters as plain text; you can just use markup of some kind
> to indicate that superscript glyphs be used for them.

That is one of the problems (and the reason I asked the question above).

> The point, from the perspective of font design and usage, is that you
> cannot have, say, superscript “s” as character, but you can have it as a
> glyph, to be selected, using special methods, for rendering a normal “s”.

And that illustrates your earlier point, that it's left as an exercise
to the formatter.

> I agree, but my point was that people are playing with superscript
> commands rather than superscript characters. The superscript two “²”
> character is in the Latin 1 repertoire and thus rather safe even in the
> pre-UTF-8 world.

I don't think the existence of superscript characters would entirely
solve the problem, though, unless they were universally available (eg
given Unicode code-points *and* available in all fonts).

///Peter

[Peter Flynn]

On 04/23/2016 02:22 AM, tlvp wrote:
> On Sun, 17 Apr 2016 16:49:52 +0300, Jukka K. Korpela wrote:
>
>> It is common in French to write e.g. 1er and 2ème, with the letters as
>> superscripts. (What is correct French is a different issue, and
>> apparently a debated one.) The use of an accented letter like “è” makes
>> it even more difficult to find a font with a superscript glyph.
>
> For another usage example, coming from US publishers' practice rather than
> French custom, I'll note the use of what can be described in HTML as
> 2^e to stand (in book-flog materials or front matter) for "Second
> Edition" or "2^nd Ed'n".
>
> That's a very low-density example, but a persistent one if you look in the
> right places :-) .
>
> Another is the use of a final ⁿ in place of the final " 'n " in
> abbreviations such as Ed'n (for Edition), funct'n (for function), etc.

Thank you, excellent examples of the same kind of precious affectation
as the 2^nd practice.

I rest my case.

///Peter

[tlvp]

On Mon, 25 Apr 2016 22:12:35 +0100, Peter Flynn wrote:

> Thank you, excellent examples of the same kind of precious affectation
> as the 2^nd practice.

LOL ... precious affectation in your eyes, standard practice in others'.
What do you make of the German and Italian hotel practices of using, e.g.,

Zimmer N^<u>r</u> 315
Camera N^o 315

for "Room number 315" on check-in forms?

[Jukka K. Korpela]

26.4.2016, 0:11, Peter Flynn wrote:

> On 04/18/2016 08:19 PM, Jukka K. Korpela wrote:

– –

>>> I'm not familiar with typefaces providing s/t/n/d/r/h as special glyphs
>>> for superscript use -- it's an interesting idea.
>>
>> They include the “C fonts” Calibri, Cambria, Candara, Consolas,
>> Constantia, and Corbel as well as Palatino Linotype, as shipped with
>> Windows 7, and Source Sans Pro, as available from Google Fonts. These
>> are hardly outdated fonts; rather, it seems that superscript letter
>> glyphs are making a comeback, at the technical level at least. This does
>> not imply you should use them; just that you can, in some environments.
>
> As a non-Windows users, I had missed these: thanks very much.
> Do you happen to know what Unicode code-points they represent?

They correspond to normal Basic Latin code-points, e.g. superscript “s”
is assigned to the code point U+0073 LATIN SMALL LETTER S. The
superscript glyph can be selected using the OpenType property “sups”.
There is really no other sensible way to do it, since there is no code
point for superscript “s” and no way to specify superscripting at the
code point level.

> You may have missed that part of the argument: they

> [superscript ordinals] make the text look

> like something out of the 1890s. Which is valid in a reconstruction of a
> historical document, but rarely elsewhere.

How many people have read texts from the 1890s these days? I must admit
that my experience with texts from different decades is so limited that
I cannot tell the time range when superscript ordinals were dominant or
common.

>> I agree, but my point was that people are playing with superscript
>> commands rather than superscript characters. The superscript two “²”
>> character is in the Latin 1 repertoire and thus rather safe even in the
>> pre-UTF-8 world.
>
> I don't think the existence of superscript characters would entirely
> solve the problem, though, unless they were universally available (eg
> given Unicode code-points *and* available in all fonts).

I didn’t mean to suggest adding superscript characters to Unicode. I was
just trying to point out that even in the important special case of
superscript two (and three), people very often don’t use the available
characters and glyphs. Instead of m², they write m2, possibly so that
the digit two is formatted as a superscript the way Microsoft Word (or a
word processor in general) does it—even though it should be clear that
the result is inferior.

--
Yucca, http://www.cs.tut.fi/~jkorpela/

[Peter Flynn]

On 04/30/2016 05:57 AM, Jukka K. Korpela wrote:
[...]

> How many people have read texts from the 1890s these days?

Historians and other researchers in the Humanities, I guess.

> I must admit that my experience with texts from different decades is
> so limited that I cannot tell the time range when superscript
> ordinals were dominant or common.

In the days of hand-set type, a superior figure would just be a smaller
size of the same font, with space material of the same width below it so
that the metal occupied the whole of the body height of the adjacent
sorts (this meant choosing a superior whose width did indeed correspond
to one of the standard widths of spaces in the smaller font), and high
enough to fill the line-height. This is fiddly but is the only way to do
it for occasional requirements, unless you were in a position to buy a
font of superior figures and letters cast on the size of the text body.
Only the large houses like Clowes would have been able to afford to do
this (they typeset things like the Bible in different languages for the
BFBS, so they would have needed all kinds of odd sorts). Some popular
book faces bought as foundry type like Old Style had superior figures
cast on the body size available as sorts when you bought a whole font.
Given that they were the expected and normal way of producing ordinals,
fine printing would therefore go to the trouble of getting it done
right, one way or another; jobbing printing would resort to whatever
contrivance the compositor was able to make available with the sorts on
hand.

Once machine-setting became common (end of 1800s to late-mid-1900s), the
fiddly bit of the problem went away, because you could add the matrices
for the superior figures needed, and they would then get cast on the
right body-size along with everything else (and of course there was no
need to keep them afterwards). But a matrix-case is of limited capacity,
so you would only do this if there were very few needed and you had the
space for them. My deduction is that this corresponds with the decline
in their usage, but it would need a lot of research to find out. Maybe
someone has already done this -- the PHS and the St Bride Library would
know.

> I didn’t mean to suggest adding superscript characters to Unicode. I was
> just trying to point out that even in the important special case of
> superscript two (and three), people very often don’t use the available
> characters and glyphs.

Education, I think. Most people simply don't know they exist.

> Instead of m², they write m2, possibly so that the digit two is
> formatted as a superscript the way Microsoft Word (or a word
> processor in general) does it—even though it should be clear that the
> result is inferior.

Most people wouldn't notice. Until recently, of course, different
encodings of the ¹, ², and ³ would mean some recipients would get the
mess of multiple bytes we still come across...

///Peter

[Peter Flynn]

On 04/26/2016 05:29 AM, tlvp wrote:
> On Mon, 25 Apr 2016 22:12:35 +0100, Peter Flynn wrote:
>
>> Thank you, excellent examples of the same kind of precious affectation
>> as the 2^nd practice.
>
> LOL ... precious affectation in your eyes, standard practice in others'.
> What do you make of the German and Italian hotel practices of using, e.g.,
>
> Zimmer N^<u>r</u> 315
> Camera N^o 315
>
> for "Room number 315" on check-in forms?

Perfectly normal: the use of the superscript in Romance languages
remains standard. But as Jukka pointed out in relation to the ²
superscript, there is a perfectly good Numero symbol № for this. Alas
there doesn't seem to be one for Nr but I may have missed it.

///Peter

[tlvp]

On Fri, 13 May 2016 22:23:42 +0100, Peter Flynn wrote:

> ... there is a perfectly good Numero symbol № for this. Alas

> there doesn't seem to be one for Nr but I may have missed it.

Heh-heh ... I missed even the one for № , so you have my thanks for calling
my attention to it :-) . Cheers, -- tlvp