[XeTeX] Ligatures and searching in PDFs

Andy Lin kiryen at gmail.com
Wed Jun 9 22:43:22 CEST 2010


> My understanding is that the only way to make PDFs 100% searchable is to use
> fonts that avoid the PUA altogether.  In other words, the 'Th' ligature
> would be accessible only via an OT or AAT feature such as discretionary
> ligatures; the user could not put it into a document by entering a PUA value
> because the 'Th' has none in the font.  Many font makers have included both
> PUA values and OT features in their fonts because of the lack of widespread
> support for advanced typographic features, particularly on Windows and
> Linux.  Some are now abandoning the PUA, partly because of the problem with
> seaching PDFs, as support on Windows gradually gets better.   You have to
> check the particular font you want to use.

David: It's funny because I remember reading an article about the
version of Palatino Linotype that MS bundled with Windows XP and later
(and used for the Silverlight OT demos); people were complaining about
how there was no way to access its many ligature glyphs in the font
through a character map/palette because they weren't present in the
PUA.

Anyhow, with your suggestion in mind, I tried out Palatino Linotype
(and several other fonts). Palatino Linotype's f-ligatures are
searchable, but not its other ligatures, such as 'Th'. Calibri is
problematic because it enables 'ti' ligatures by default, and these
are not searchable (its f-ligatures are, however). Strangely enough,
ff-ligatures are searchable in the Adobe fonts that have them, but not
the f-ligatures. Out of the Windows fonts I tested (Cleartype
collection), only Cambria did not have searchable f-ligatures.

I've attached a PDF of the fonts I've tested so far. This is without
any TECkit mapping, and with Common and Discretionary ligatures
enabled. I hope someone will find this list useful.

Ross: Thank you very much for the clarification. Your post was very
informative and helpful, as always.

-Andy
-------------- next part --------------
A non-text attachment was scrubbed...
Name: fitest.pdf
Type: application/pdf
Size: 184976 bytes
Desc: not available
URL: <http://tug.org/pipermail/xetex/attachments/20100609/cd5e4723/attachment-0001.pdf>


More information about the XeTeX mailing list