Small Caps `t` and `h` with dot under them

[khalid-hussain]

I'm not a professional typographer but if I'm not mistaken the small letters t and h with a dot under them in small caps do not render properly.

In the picture above, the first set is from Linux Libertine and the second from EB Garamond.

[georgd]

@khalid-hussain I know, it’s a bit late. But if you use ss20 (Stylistic Set 20) with smcp you should get the correct small caps.

[khalid-hussain]

I tried it on Microsoft Word. Yes, you're right.

Thank you very much for the clarification. I don't have much knowledge about font creation but wouldn't ss be used for variations for a glyph and not to actually store a glyph?

[georgd]

Well, I know that, what ss20 does is not deduceable from the feature’s name. There is in fact a feature ccmp (something like Character composition and decomposition). But that is often activated automatically and I don’t want every precomposed glyph decomposed everywhere because users might get confused. Therefore, I chose ss20 which I didn’t use for anything else and which is available in many applications.

What exactly a feature does or has to do or even how it has to do something isn’t standardized so there’s no problem, technically, with my use of ss20.

Am 10.08.2017 um 16:38 schrieb Mohammad Khalid Hussain:

Thank you very much for the clarification. I don't have much knowledge about font creation but wouldn't ss be used for variations for a glyph and not to actually find a glyph?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/georgd/EB-Garamond/issues/86#issuecomment-321571037, or mute the thread https://github.com/notifications/unsubscribe-auth/AAND8gqlmo3h_zGqzPew2seQmTc6SiBoks5sWxXKgaJpZM4LkKD-.

-- EB Garamond: http://www.georgduffner.at/ebgaramond

[khalid-hussain]

What exactly a feature does or has to do or even how it has to do something isn’t standardized so there’s no problem, technically, with my use of ss20.

If I understood this correctly, then shouldn't this be applied to the other letters too? S and Z don't have this problem even though, they too have a dot under them. I understand the ss category has no hard-encoded rules but isn't consistency important?

Please forgive my ignorance.

[georgd]

That’s an optical illusion. Lower case s and z share shapes with the upper case glyphs, therefore it looks as though there were smallcaps (which use uc shapes). If you look closely you’ll notice that they are a bit smaller than small caps s and z.

Am 12.08.2017 um 05:37 schrieb Mohammad Khalid Hussain:

What exactly a feature does or has to do or even how it has to do
something isn’t standardized so there’s no problem, technically,
with my use of ss20.

If I understood this correctly, then shouldn't this be applied to the other letters too? S and Z don't have this problem even though, they /too/ have a dot under them.

Please forgive my ignorance.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/georgd/EB-Garamond/issues/86#issuecomment-321954946, or mute the thread https://github.com/notifications/unsubscribe-auth/AAND8ndHADYgypvdHYmvwqGc5DksSbtcks5sXR4BgaJpZM4LkKD-.

-- EB Garamond: http://www.georgduffner.at/ebgaramond

[khalid-hussain]

I didn't know that. I created a sample file using in LuaLaTeX as follows,

\documentclass[ebook]{memoir}

\usepackage{fontspec}
\setmainfont{EB Garamond}
\setlength{\parindent}{0pt}

\begin{document}
{\addfontfeatures{StylisticSet=20}\textsc{ḥadīth S Ṣ s ṣ}}\\
\textsc{ḥadīth S Ṣ s ṣ}
\end{document}

I got the following,

There is a clear difference for the SC small s when not using ss 20. Is there any other difference I should watch out for?

[georgd]

The first one is fine. Beyond the bigger western European languages you should always use ss20 with small caps. Also when transliterating from Arabic like in your case.

Am 18.08.2017 10:39 schrieb "Mohammad Khalid Hussain" < [email protected]>:

I didn't know that. I created a sample file using in LuaLaTeX as follows,

\documentclass[ebook]{memoir}

\usepackage{fontspec} \setmainfont{EB Garamond} \setlength{\parindent}{0pt}

\begin{document} {\addfontfeatures{StylisticSet=20}\textsc{ḥadīth S Ṣ s ṣ}}\ \textsc{ḥadīth S Ṣ s ṣ} \end{document}

I got the following,

[image: ebgaramond] https://user-images.githubusercontent.com/3102714/29450945-513558da-8433-11e7-898f-af94f4553fb8.png

There is a clear difference for the SC small s when not using ss 20. Is there any other difference I should watch out for?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/georgd/EB-Garamond/issues/86#issuecomment-323295720, or mute the thread https://github.com/notifications/unsubscribe-auth/AAND8q9oDaSIDVVqHpdOK57MH1IQzeveks5sZU3LgaJpZM4LkKD- .

[skurzinz]

For reference: Using ss20 in XeLaTeX also avoids a strange error that seems to use a lowercase f in all \textsc{} instead of the smcp. Not filing a separate issue…

[skurzinz]

For reference: Using ss20 in XeLaTeX also avoids a strange error that seems to use a lowercase f in all \textsc{} instead of the smcp. Not filing a separate issue…

The issue seems to have been a strange interaction with polyglossia:

\setmainlanguage[babelshorthands=true]{german} was triggering this, even if ss20 was used (but strangely enough, this was not consistently happening).

[skurzinz]

The issue seems to have been a strange interaction with polyglossia:

\setmainlanguage[babelshorthands=true]{german} was triggering this, even if ss20 was used (but strangely enough, this was not consistently happening).

Fixed now by falling back to babel…