Wishlist

[dhinostroza]

Hi,
I can't seem to find these in Junicode/JuniusX. If not included, who should I send the request to?
Best regards,
Daniel

[psb1558]

On the top line it looks like you have a variant of a ligature of q and "et" (U+A76B). The recommended way to produce this in Junicode 2 is to enter the two characters and apply the OpenType feature hlig (Historical Ligatures) to them. To add the acute, type this sequence:

q + U+0301 (combining acute accent) + U+A76B (et sign)

Remember that you have to apply hlig to all three characters. This won't work in old Junicode 1, and it is not yet available in the italic of Junicode 2 (but should be soon). Also, in the roman the anchors that position the acute accent correctly are messed up, but I have corrected them and will post the rebuilt font later today. MUFI encodes this ligature at U+E8BF, but I don't think you'll be able to position an acute on that.)

How important is that loop connecting the q and the et sign? Remembering that Junicode can't reproduce the stylistic features of the many early modern typefaces out there. Also, the loop makes it look a little g-like. It doesn't fill me with enthusiasm.

The stroke above various vowels is a stylistic variation of the common mark of abbreviation, usually represented with a macron (U+0304). I could include the variant in the font. I would ask: how important is this?

The most interesting thing here is the q with tail on torq(?)eatur. I don't see anything to match it precisely in Cappelli. Assuming that this is an abbreviation, how do you expand it, and how common is it? Can you supply a higher-resolution version of it, or a link to one?

BTW, what book is this?

[dhinostroza]

Dear Peter,
Thank you very much for considering solutions to these characters. It's the Cautio Criminalis, 2nd Ed. I'm using JuniusX, with just a tad of Adobe Caslon Swash for some uppercase letters in italics and Hoefler italic for the abbreviated "etc".

---

..."On the top line it looks like you have a variant of a ligature of q and "et" (U+A76B)"
Yes, the closest I have is q + U+A76B. I'll update my font to Junicode Two Beta and try q + U+0301 + U+A76B. Thank you!

---

..."How important is that loop connecting the q and the et sign? Remembering that Junicode can't reproduce the stylistic features of the many early modern typefaces out there. Also, the loop makes it look a little g-like. It doesn't fill me with enthusiasm."
The loop really scratches me the wrong way, hehe. It's definitely not a g (your JuniusX "g" is perfect for the book), but I wonder if it's rather a q + U+0301 + the German eszett in subscript. I don't know what the abbreviation means. If I find it, would you be interested?

---

..."The stroke above various vowels is a stylistic variation of the common mark of abbreviation, usually represented with a macron (U+0304). I could include the variant in the font. I would ask: how important is this?"
Well, there are vowels with macrons in the text, yet these half-macrons ending with a half-breve seem to be predecessors to á é ó ú. I haven't found this pattern on the i, yet. So, yes. If it's not too much of a burden, please include this variant in the font.

---

..."The most interesting thing here is the q with tail on torq(?)eatur. I don't see anything to match it precisely in Cappelli. Assuming that this is an abbreviation, how do you expand it, and how common is it? Can you supply a higher-resolution version of it, or a link to one?"
Yes, this is a quite unusual character. It only appears once in the book (page 168, last line). I also looked into Cappelli and it's not documented there.
Cautio_Criminalis_ocr_la-0186.tif.pdf
(https://download.digitale-sammlungen.de/pdf/16298249948888bsb10393951.pdf)
If you compare both sources attached, the second scanned image does not show the q + tail. But then again, there seems to be a "si" with a Latin small letter long s with diagonal stroke (U+1E9C) at the end of that line that is different from the "si" in the first image. Maybe it's just a smudge, maybe not.
All the very best,
Daniel

[psb1558]

The italic ligature looks like this right now:

Not a terrific match for the roman: suggestions welcome.

[dhinostroza]

You discovered something I had overlooked. This ligature only appears in italic. Perhaps a subscript eszett or et might be closer?
All the very best,
Daniel

[dhinostroza]

I looked at the first edition in hopes the q with tail on torq(?)eatur would repeat itself. Doesn't seem so.

All the very best,
Daniel

[psb1558]

"Thus should the accused be tortured." What in heaven's name is this book?

Oh well. I think the thing that seems to hang down from the q in one of the images is pretty clearly a smudge. There are no letters to be supplied there, "torqueatur" being perfectly complete.

You don't need to look anything up. The q + et sign is a common abbreviation for -que. I don't know why there's an accent, but that doesn't matter. The forms are perfectly intelligible. What Junicode needs, I think, because this is a very common variant, is a subscripted et sign that ligatures with q. Vaguely like this:

I already have a feature for alternates of the et sign, so it'll be no problem to make this work.

This is a busy time, so probably nothing's going to get done till the weekend.

[dhinostroza]

...""Thus should the accused be tortured." What in heaven's name is this book?"
It's the book that questioned, and ended, the Inquisition. A truly interesting read. The English translation is said to have been wrongfully altered from the latin source. I'm comparing the translations and the little nuances in the font type, page by page. It seems to have been written by multiple Inquisitors and printed in multiple locations.
Again, thank you very much for your offering to include these characters in Junicode Two!
All the very best,
Daniel

[psb1558]

Here is a first try at alternate et and a ligature (examples are italic, since that is the more interesting problem).

No. 1 is a followed by et U+A76B: this is the standard abbreviation for -que 'and'. No. 2 si the same, but with alternate et (somewhat subscripted). No. 3 is a ligature of q and alternate et. I wasn't going to have a go at the loop, but without it, it looked a lot like No. 2—and then what would be the point?

Suggestions?

[dhinostroza]

It's a step in the right direction and, with the acute accent, would be useful immediately. If you could reduce the horizontal space within the oval of q, and flatten the ligature between the loop and the et character, your end result would be much closer to the source. Thank you and please continue fleshing it out!
All the very best,
Daniel

[psb1558]

Here's the state of things. Various OpenType features applied to q + U+A76B (et), which together constitute the common abbreviation for -que "and.":

Feature cv83[1] transforms the "et" to the same character with a shape like a semicolon. (This is in the last roman sample above, line 3 from bottom). The feature has been in Junicode for a while.

cv83[2] transforms the "et" to a smaller, somewhat subscripted form, common in MSS and printed books.

The last two lines show what happens when hlig "Historic Ligatures" is applied to q + et without and with cv83[2] applied.

A couple of notes about this. There will be no gyph with acute added, but you can get that by adding the combining acute U+0301 after the q and applying hlig to the whole three-character sequence:

And, the q + et ligature can only go so far towards imitating an original, since it's got to be true to the design of Junicode. The q is the shape of q in Junicode, and the et is the Junicode et. The only new thing in the italic ligature is the loop that attaches the q to the et, and that can't be quite the same shape or join the two elements in quite the same way as in the italic example above because (1) the q in the original descends way farther below the baseline than the Junicode q and (2) as I stated in an earlier post, I don't love the original ligature (because it looks a little like a g + et sign.

This is an extremely common abbreviation in medieval manuscripts and early printed books (here's one from a Gutenberg Bible)

And so it is right that Junicode should have it in several forms, with the ability to apply different marks to it. But it would make no sense to imitate any one font.

I may do some work on improving consistency between the roman and the italic, though I don't think the roman ligatures will get a loop.

New version on the website later tonight.

[dhinostroza]

Hi,
Latest Junicode Two Beta (regular and italic) is not validating on MacOS Big Sur nor Windows 11. Thought it had something to do with Junicode 1, so I removed all Junicode 1 and retried installing Junicode Two: Font not recognized (fail).
Installation attempts did not affect JuniusX. Tried downloading/installing the individual fonts and repeated with the entire zip package. Same message.
All the very best,
Daniel

[dhinostroza]

Update for MacOS: I manually placed the fonts (downloaded as a zip file) in Users/Me/Library/Fonts and restarted. All is well now.

[dhinostroza]

Here's the state of things. Various OpenType features applied to q + U+A76B (et), which together constitute the common abbreviation for -que "and.": Feature cv83[1] transforms the "et" to the same character with a shape like a semicolon. (This is in the last roman sample above, line 3 from bottom). The feature has been in Junicode for a while.

cv83[2] transforms the "et" to a smaller, somewhat subscripted form, common in MSS and printed books.

The last two lines show what happens when hlig "Historic Ligatures" is applied to q + et without and with cv83[2] applied.

A couple of notes about this. There will be no gyph with acute added, but you can get that by adding the combining acute U+0301 after the q and applying hlig to the whole three-character sequence: And, the q + et ligature can only go so far towards imitating an original, since it's got to be true to the design of Junicode. The q is the shape of q in Junicode, and the et is the Junicode et. The only new thing in the italic ligature is the loop that attaches the q to the et, and that can't be quite the same shape or join the two elements in quite the same way as in the italic example above because (1) the q in the original descends way farther below the baseline than the Junicode q and (2) as I stated in an earlier post, I don't love the original ligature (because it looks a little like a g + et sign.

This is an extremely common abbreviation in medieval manuscripts and early printed books (here's one from a Gutenberg Bible) And so it is right that Junicode should have it in several forms, with the ability to apply different marks to it. But it would make no sense to imitate any one font.

I may do some work on improving consistency between the roman and the italic, though I don't think the roman ligatures will get a loop.

New version on the website later tonight.

---

Hi, Peter:
So, it took me this long to finally comprehend (somewhat) your solution. I had to learn how to render CSS within a database field and how to apply character variants and historical ligatures. Yet, I haven't been able to pass Stage two (see screenshot below).

You will notice that multiple variations of cv83 2 and hlig have been applied (with corresponding code on the right), and I've even swapped the order but I haven't been able to render the loop that attaches the q to the et.
The combining Overline, Macron, Cyrillic and Grapheme joiners are there to prove that character-variant is active and rendering correctly. Where am I failing with the combination?
Best regards,
Daniel

[AugustusMarcus]

Hello, @dhinostroza. I know it's a little late, but I managed to fix your issue after coming across it on my own. How to fix it:
Note: I'm using the "JunicodeVF-Italic.woff2" file.

Do not use different html tags for the 'cv83[2]' and the 'hlig' features like you did. For some reason that makes it not work. Put both of these features inside a single tag or class. Example:

Check the final result here: https://reviewsheet.xyz/quiz%20epid/pr2/qu3st1.html; inspect the html if you're having any trouble.

Good luck!

[psb1558]

I'm not getting this error on Big Sur. I'll try various validators and see if I can figure out what's going on.

[psb1558]

A first go at the common mark of suspension (most often for a missing n or m, and represented by a combining macron:

I don't know how much I like this. I've seen this shape of the mark before, though: I may take a prowl through some early printed books to get a good sample of them.

[dhinostroza]

This looks amazing!

[jsbien]

represented by a combining macron

Is this a common practise? I use U+0305 COMBINING OVERLINE for abbreviations (despite the fact it doesn't look nicely; MUFI F00A-F00B also are not positioned correctly), as in Polish texts the macron can be used over some letters as an equivalent of the contemporary dot. No example at the moment, but I will look for one.
BTW, perhaps F00A-F00B qualify as cv of, say, macron or overline?

[psb1558]

This btw seems related to a more extravagant variant, which is in the MUFI recommendation and has therefore been in Junicode for a number of years:

[jsbien]

Which MUFI character do you mean? F1CC COMBINING CURLY MARK ABOVE?

[psb1558]

Yes. I've never been completely clear about what this is supposed to do (I think I may have asked once on the list without getting a reply), but it looks like a pretty common variant of the standard macron-like abbreviation for missing n/m and a few other thing.

[psb1558]

This is an improvement, I think, in both shape and positioning.

It's a bit anachronistic, being from an Aldine Press book, but then all the Latin abbreviation marks are, really: the Oxford Press doesn't seem to have used a lot of abbreviations in their books ca. 1700.

[psb1558]

Slimmed down a bit for italic.

[dhinostroza]

Hi,
This looks very nice in CSS. How do you make it work using InDesign?

[psb1558]

I think that InDesign is edging towards making OpenType features available directly, so that eventually you should be able to apply character variant features to selections and whole texts. But for right now I don't see a method other than going through the text and tediously replacing each macron with its variant using the glyphs palette. In theory you should be able to do a search-and-replace using the "glyphs" tab of the search dialog, but I can't make it work for this purpose.

Maybe some visitor will be able to offer pointers.

[jsbien]

Some examples from a Polish 16th century book.

[psb1558]

I'm fascinated by the variability of these marks, which I also saw in the Aldine prints I was browsing at the Beinecke Library site (pretty good collection of digit material with a good search engine). They come in a variety of shapes, but what they have in common is an avoidance of the simple straight line. In early typefaces they're emulating pen strokes, which are rarely straight or perfectly consistent.

[dhinostroza]

Can't seem to capture the characters on my keyboard. The Apple font book displays them correctly, but when I copy and paste them they become garbled. Pasting them in Word + Advanced typography (Junicode Two Beta Italics + all ligature combinations), latest BBEdit, LaTeXiT didn't work either. Are they in a specific PUA location?

[psb1558]

Amazed that characters can be copies and pasted from Font Book. I didn't know that. It looks as if characters that don't have encodings in the font can't be copied from Font Book and pasted into Word. I don't know about BBEdit or LaTeXiT. BBEdit is proprietary, and of course I can't collect proprietary software for the purpose of testing a free font. LaTeXiT looks like a LaTeX editor for the Mac, and I don't run LaTeX on the Mac.

The model that Junicode employs for getting characters that lack Unicode encodings (I mean encodings assigned by Unicode, not encodings in the PUA) is that you enter a Unicode character and apply one or more OpenType features to it (details in Feature_Reference.pdf). You can see how to do this for LaTeX (XeLaTeX or LuaTeX flavors) in the Feature_Reference.tex file in the docs folder of this repository. You can see how to do it on a webpage in my little essay about Searchability: view the page source for a commented example of how to apply OpenType features to HTML text. Unfortunately, I have no idea whether BBEdit or XeLaTeXiT provide access to the features you would need.

Word is a special problem because it provides access to only a few OpenType features, and Junicode uses many. In fact, Word is the most primitive of the major word processing apps in support for advanced typography (by contrast, LibreOffice gives you access to essentially all of a font's features, and even humble TextEdit and Pages offer a lot of them). If you're stuck with Word, the only way around its limitations is to include PUA-encoded versions of the glyphs you want. I can do that for the q_et ligatures. Note that they will not be searchable if you use the PUA encodings, but if you're generating printed documents, that won't matter.

PUA encoding for combining marks is a different problem. In most apps, including Word, a combining mark with a PUA encoding will never be positioned correctly over its base character (the reasons are technical and boring). The only way around this is to provide a bunch of precomposed (base + mark) characters in the PUA for each variant mark. I won't do that, because every PUA encoded character creates a little accessibility nightmare, and I'm willing to go only so far to work around MS Word's stupidities. If you have to use Word and you want access to that variant suspension mark (treated in Junicode as a variant of U+0304 combining macron), I'm afraid you're in trouble.

[psb1558]

I don't know FileMaker, and I don't know how much of the OpenType spec it supports. Since you can do a ct ligature, it seems to support dlig "Discretionary Ligatures," but I'd be surprised if it gives access to the more exotic ones like hlig "Historic Ligatures" or the cvNN "Character Variant" features.

I can think of a couple of ways to get around the program's limitations. One is to use some kind of markup to signal formatting you'll have to do later, when you massage the text into a reader-friendly form. You might, for example, use entity reference-like notation for your q_et ligature and other things, e.g. &q_et;, &oslantingmacron;. Or you could tag Unicode-standard characters with XML-like tags to indicate the features that will have to be applied to them: <cv90>ō</cv90> (btw, you should use the macron for the mark of abbreviation in co(n)cludens, as that is the standard way of transcribing the mark that signals a missing n or m.).

I've never used LibreOffice's database management program, and I can't promise that it supports OpenType features as fully as the other LO programs do. But I'd guess that it does. I don't suppose it would be a particularly difficult matter to export your database from FileMaker to a format that LO can work with, but I can imagine complications that might make it better to simply stick with FileMaker—especially if you've already done of a lot of your transcription.

The main difference between JuniusX and Junicode Two Beta, which produces the problems you are observing, is that Junicode Two Beta has an italic face, which is not as complete as the roman face. As JuniusX didn't have an italic, FileMaker simply slanted the roman when you asked for italics. On most days I add a dozen or more characters to the italic, and I enable OpenType features whenever possible, but it will be months before it's as finished as the roman (there are 900+ characters still to go).

So for now, I would suggest deleting the italic fonts from your system and letting the program slant the roman as before. When the italic is complete I'll announce it on this list and elsewhere.

[dhinostroza]

Thank you. I'll follow your advice!

[jsbien]

you should use the macron for the mark of abbreviation in co(n)cludens, as that is the standard way of transcribing the mark that signals a missing n or m

In what sense it is standard? Why not e.g. COMBINING OVERLINE?

[psb1558]

Actually, I can't claim that it's standard, but only that it's my preference (sheepish grin emoji here). I see that MENOTA uses U+0305 combining overline for this abbreviation, and since Unicode doesn't provide anything for this specific purpose, I can't say it's wrong.

But to me, U+0305 looks too wide. The purpose of that character, at least if you can trust Wikipedia, is for certain mathematical expressions, for indicating numerals in Ancient Greek, and for × 1000 in Roman numbers. It is supposed to go the full width of a character and is especially appropriate when it has to join up with another overline on left or right. In explaining the difference between a macron and an overline, J.C. Wells, in his "Orthographic diacritics and multilingual computing", says "Macrons on successive letters are separate; overlines link up," and gives this illustration:

By this way of thinking, the abbreviation that stands for a following m or n is more like U+0304 combining macron than like U+0305 combining overline, since it generally sits over a single character. (If it goes over two characters, we have U+035E for that.) Further, it doesn't always go the whole width of a character (and U+0305 extends over both the width of a character and its sidebearings—the white space on either side).

So my recommendation is to use U+0304 or one of the macron combining forms for the common suspension and U+0305 for Roman numbers.

[jsbien]

There is also U+FE26 COMBINING CONJOINING MACRON which can be used in a sequence or between U+FE2E COMBINING CYRILLIC TITLO LEFT HALF and U+FE2F COMBINING CYRILLIC TITLO RIGHT HALF. As I mentioned earlier, I want to distinguish abbreviation mark from a diacritic over a letter. This is probably not strictly necessary, but convenient. I seriously consider using for Latin and Polish abbreviations, despite its name, U+0483 COMBINING CYRILLIC TITLO.

[psb1558]

I would also like to have a combining mark that's strictly for the abbreviation. Someone should propose it to Unicode, which has lots of character pairs that look alike but are semantically distinct. In the meantime, we have the macron, which was originally a length marker but now has many uses, so that you can use it without people coming down on you for abusing a Unicode code point.

You can also distinguish the abbreviation mark from the length mark with markup. You can do it in TEI with the <abbr> and/or <expan> tag. You can do it in HTML/CSS by enclosing all abbreviations in <span> tags with an appropriate class, like abbr. You can even do it in LaTeX by creating an \abbrev macro. The XML-based solutions, at least, would permit search programs to find all the abbreviations in a text.

[jsbien]

OK, but the link to abusing Unicode doesn't work for me.

[psb1558]

Sorry—I've corrected it.

[marrus-sh]

U+034F COMBINING GRAPHEME JOINER (ignore the name) is the typical option for introducing semantic difference into character sequences which Unicode would consider equivalent. For example, CH versus C<CGJ>H might distinguish the digraph from the two letters in sequence for purposes of sorting or searching (but not display). Similarly, E<combining macron> (for the letter plus diacritic) might be distinguished from E<CGJ><combining macron> (for the abbreviation).

(See What is the function of U+034F COMBINING GRAPHEME JOINER? on the Unicode FAQ, and the following question regarding distinguishing tréma from umlaut, which is similar in principle.)

Is this advice practically useful? Uh, it probably depends on your exact technical situation / what you need to achieve. But I figured I would mention it as one approach.

[psb1558]

(Replying here also to one from above.)

Strange and annoying thing about OpenType features in CSS: they don't get inherited. So if you use font-feature-settings, exactly those features invoked there get set, and everything else gets reset to the default. So if you do (here I'm doing it as CSS/HTML would do it, since that's all I know):

.hlig { font-feature-settings: "hlig"; }
.altet { font-feature-settings: "cv8" 2; }

and then, in the HTML

<span class="hlig"><span class="altet">q&#xA76B;</span></span>

you'll only get the alternate et, not the hlig. So what you want is:

.que { font-feature-settings: "cv82" 2, "hlig"; }

then apply the "que" class to the same sequence. It ought to work, and an inserted ̄ shouldn't make a difference. Try something like this and report back whether it works.

I'm not sure what's going on with the alternate macron, except that in the current version it's cv84[2], not cv83[2]. Are you using the current version?

[dhinostroza]

Managed to reduce the CSS code but got more or less the same results:

When font-variant-ligatures: character-variant; is added to the CSS, the ligatures become active. Otherwise, the que and macron font-feature-settings: "cv82" 2, "hlig"; are not enough. In the screenshot you will see that cv82 2, cv83 2, and cv84 2 combinations were all tested, just in case I had the second digit wrong.
Indeed, character-variant cv84 2 does something slightly different to the macron + u letter, so I have to be missing something.

[psb1558]

This is not Junicode's italic q:

which looks like this when standing alone:

Note the quite different shape of the bowl. It looks like it might be just the roman q + et combo, slanted. Are you sure you have the italic font installed?

In fact there's quite a bit of non-Junicode mixed into your illustration, so it's hard to figure out what exactly you're pointing at. But I've been trying different things with the alt macron thing in LibreOffice, LaTeX and various browsers, to see if I can break it, but it holds up very sturdily, if you've got the characters in the right order: a͏̄. The only feature you should need to make that work is cv84[2]. Definitely that. (An earlier version of the font had macron variants on cv90, so you've got to be up to date (current version is 1.032).

If this doesn't start working in the next couple of days I'll fix up a kludge for you that will work.

[dhinostroza]

That has to be it! I had uninstalled italics, as per your recommendation a while back. Thought the behaviour would be the same with the regular font. I'll test it on another machine and let you know.
Thank you!

[psb1558]

The behavior is the same, and you got the ligature, I think, but it was just the roman ligature, slanted. Doesn't have the pzazz of the italic, which is more or less usable now, though things are still missing.

Now I'm more worried about why the macron isn't working. Make sure you're version is up to date and let me know.

[psb1558]

Some time ago, responding to an issue on my Elstob font, I fixed a variety of problems with its math symbols. Discovered that math has its own particular set of requirements stemming from the way the symbols get used in expressions. In my to-do list is the task of correcting all the math symbols in Junicode the way I did for Elstob. But of course that is just a very basic set of symbols. The prospect of adding a whole lot more is, as you say, daunting.

No. 1 task is completing the italic face and applying the various tweaks and fixes needed to get the font to version 2.0. Then I can take another look at math.

[pgratrex]

Thanks for the answer! I suspected that math symbols would be lower down the priority list. I look forward to seeing 2.0 :)

[carbeck]

I just wanted to throw in this type of de ligature again that sometimes appears in German-language manuscripts, lest it be forgotten. It's not in MUFI (yet?), though it might be made available by means of a stylistic alternative to U+F096 LATIN SMALL LETTER D ROTUNDA WITH LATIN SMALL LETTER E ABOVE.

[psb1558]

Thanks for reminding me of this, and sorry it slipped my mind before. It looks as if it should go with the Historical Ligatures (hlig). Is that deum in this image and deus in the one as Sourceforge?

[carbeck]

Nice, thanks :) No, that is the German dative singular masculine/neuter definite article dem, and the completely visible part in the image on Sourceforge reads deſ. So this ligature is really just that, a graphic contraction, not an abbreviation.

[psb1558]

Yes, I visited the MS after writing that query and saw it was German. Very nice MS.

[psb1558]

Does this capture it?

[carbeck]

Yes, that looks like a reasonable adaptation to a Roman typeface to me.

[carbeck]

On closer inspection, it seems a little high, no? I mean, it tops even letters with ascenders. As I understand it, it's basically the delta-like d with the top curl modified into an e-shape. I'm no expert on type design, though, and I know the design of this gylph is tentative.

[psb1558]

I see that. Trouble is, there's too much going on vertically. Got to have the bowl of the d and enough of the ascender for identification, and then an e big enough to be identifiable as not-a-diacritic. Too much squashing might be fatal. Perhaps I should shrink the bowl a bit, as that seems part of the strategy in the MS.

[psb1558]

I think this will probably work:

Each of the vertical elements is squashed a bit. I think the shorter bowl of the d works because a number of fonts adopt this strategy when a letter has a crowded upper storey (e.g. ð). In making the e beaked (as in the MS) I was mainly after that diagonal stroke, which enlarges the upper counter (in heavier weights it had gotten so small it was threatening to close up completely) and uses up some of that empty middle space.

The height of the ligature now matches that of f and ſ.

[psb1558]

The italic already looks like this. It just needs to be scaled vertically and the weight adjusted.

[psb1558]

Five masters for d_e.hlig.

Regular, bold, light, condensed, expanded.

[psb1558]

And a first run at an italic. Not great, Curves are award. But maybe the fundamentals are okay.

[psb1558]

The italic, a bit improved (though I still don't love it).

It's in the (approximately) weekly build, just uploaded. Access via hlig.

[psb1558]

Actually, for the italic I think it works better to keep the beak of the e. This will be transparent enough to medievalists.

[dhinostroza]

Hi, Peter:
I have a few more requests for the wishlist, please.

1. Final e at end of word
   

2. Swash Uppercase italics
   

3. Contextual ligature (clig, italic) for final s after a, e, i, and u
   

4. et italic (not &)
   
   

They are most certainly not at the top of your priority list, though I'd love to see them soon. Maybe they're already in Junicode 1?
Best regards,
Daniel

[dhinostroza]

Here's a swash e for you. The tail closely matches the one in the printed book, while the "e" part on the left stays as close to Junicode as it can manage. There's a small collection of swashes in the italic, and it may grow:

That looks nice! I'll add it over the weekend.

Here's a swash e for you. The tail closely matches the one in the printed book, while the "e" part on the left stays as close to Junicode as it can manage. There's a small collection of swashes in the italic, and it may grow:
Great, I can use them right away! I'd only be missing the Q, N and T swashes for the Cautio Criminalis.

[dhinostroza]

Made a mistake: I put the vowel + s ligatures in Elstob, not Junicode, but they're a standard feature of the Fell fonts: So they'll be in Junicode 2.

I tested Junicode 1.032 italic on another machine and both the ampersand and the accented q + et are looking beautiful!
Thank you so much!

[psb1558]

I have a Q in progress, bu,t I don't know how soon I'll have an N or a T. I'll post a new build with this stuff in it probably tomorrow afternoon or evening.

[psb1558]

These ligatures will be in version version 1.033 (in dlig "Discretionary Ligatures").

[dhinostroza]

Hi, Peter,
Perhaps the M, N, R, and T swashes can still be considered for Junicode?

[dhinostroza]

Hi, Peter:
Again, the unsufferable way InDesign handles Opentype alternates.

Your modified macron will place itself over the consonant, or place itself behind or underneath the macron. You can also place multiple modified macrons over the vowel but there is no way of adding just the modified macron to the vowel. If the excess macrons are removed, the default flat macron reappears. Guess I'm stuck with applying CSS in Indesign, somehow.
Best regards,
Daniel

[psb1558]

It is an amazing script. It's working for me:

Make sure you're applying cv83[2] (not cv84[2]) to get the et variant you need, and don't forget that you have to apply both cv83[2] and hlig to the whole sequence q + U+0301 + U+A76B. Best, in fact, if you apply both cv83[2] and cv84[2] to the whole text (if you're not using any of the other cv84 variants of macron) and hlig where needed. (Edit: Oh yes, and for your paragraph style, under "Justification" choose for the Composer "Adobe World-Ready Paragraph Composer.")

[dhinostroza]

Oops! Changing to Adobe World-Ready Paragraph Composer makes some really nasty changes. I'll have to look at it further before applying. I made q+accent+et work in the meantime by placing the accent before the q and applying -340 tracking to the accent.


Best regards,
Daniel

[psb1558]

What breaks with the world-ready composer?

[dhinostroza]

It changes the word spacing for the entire style and reflows the text, which messes up the many manual changes done to the spacing in order for the text to look as closely as possible to the source. It might become useful in another task for the text, but for now I can't allow it to reflow without close supervision.

[psb1558]

That's too bad. I was hoping it was something I could help with.

[dhinostroza]

Hi, Peter:
The next character variants for the wishlist are for uppercase Q:

1. A Qv ligature, where Q has a tail with a flat swash and v is smallcaps+expanded. Here are the screencaptures:
   
   

2. An Uppercase Q swash, similar to the Uppercase Q available in Adobe Caslon Swash:
   

Thank you very much for your effort on Junicode Two Beta!
Daniel

[TheKnightWho]

Hi - could I please request that the extended tail form also be added for small caps?

[eclecticfluff]

In version 1.049 VF, CV33[3] is now the only way to get a Qu, or the ligature moved to some different feature. Is that change intended? Rereading, I see the change is intended. It might be relevant to mention it in the documentation in section 4.14, though so other people don't need to go through the confusion I did.

[psb1558]

I'm sorry to have confused you. Ligatures are disappearing fast in Junicode, most of them invisibly replaced by contextual alternates. But Qu really did disappear--I think I picked up somehow that not everyone liked the long tail making an appearance whenever a u followed--and of course there was no way to make all possible ligatures for the Q with longer tail.

I'd rather not document the disappearance of a ligature in a section where such negative announcements otherwise have no place. Instead I'll try to work up the discipline to keep a change log, which I'll keep up to date at the back of the manual.

[eclecticfluff]

That makes sense. It is not completely standard, and if truly needed it could allow avoid collisions when followed by some lowercase with a descender. Speaking of ligatures, would it be appropriate to open a discussion thread for ligatures in general? I have a few more questions about the philosophy and organization of them.

[dhinostroza]

I'm wondering if the behavior of U+0051 Latin Capital Letter Q, i.e. stretching its tail to embrace the following letter (alt, alt3, q_u, q_pcap), can be applied to Q swash. It seems the slant of the tail would need to be less pronounced. Screenshot below:

[psb1558]

Here's another go at a swash Q, interpolated in various ways (Regular at top left)

[dhinostroza]

Dear Peter,
As you are quickly approaching the release of Junicode 2.00, I was hoping you could give one last look.
Source examples are .
I am aware of your need to maintain consistency with the entire Junicode font, but even a little tweak slanting the ligature would be very significant for the Cautio Criminalis.
Very kind regards,
Daniel

[psb1558]

I'm not sure what you're asking here. The ligature has the same slant as the rest of the italic face.

[dhinostroza]

Some time ago, you commented that the q in the source looks more like a g. Starting the ligature from the straight tail of q makes the slant of the ligature look almost reversed in Junicode. I was hoping the ligature could lean a bit more.
The weight seems to be different when compared to other characters in Junicode italics:
Best,
Daniel

[psb1558]

I think you're right that the weight of q_et looks different: the ligature is too crowded, and that makes it look darker than the surrounding text. But changing the slant won't solve the problem. It seems to me that most of the problems with the ligature arise from my trying to reproduce that loop while trying to avoid the flaws of the original (the q too much like a g, the et too tiny). There isn't room for the loop if the et is larger.

My solution would be to get rid of the loop, like so:

That would make the ligature more legible and comprehensible. This would make it less like your original visually, but here I would recommend section 7.1 of the Junicode Manual. My little sermon there is aimed at the folks transcribing old records, but it applies here too—especially to your use of v with hook (U+028B) instead of u or v in utilissimus, which now cannot be searched as either u or v.

(BTW, the right sidebearing of U+028B was too narrow: I have changed that now.)

[dhinostroza]

A few more swashes for the wishlist, please:

These two additions (G and T) would eliminate Adobe Caslon Swash from the file.
The uppercase i is actually a 1 using Junicode, but placing a number there is not the best solution. It's very odd to have that character in such an important page (page 1), but again, the book was printed quite hastily.
Thank you very much beforehand,
Daniel

[psb1558]

I'll think about more swashes. Right now I've frozen new features while searching out bugs and design flaws, but I may think about it later.

BTW, I don't think that's an I in MAGISTRATUS: I think it actually is a one. I've never seen a swash I remotely like it (a swash I generally has a top like a swash J). What I have seen a good bit of, though, over the years that I've been mucking around with early printed books, is old typesetters taking desperate measures when they can't find a particular bit of type in the font they're using. Here the typesetter didn't have an I in the correct size, and in desperation he used a 1 instead. See how the spacing is off? The 1 is not designed for this context.

Are you sure you want to reproduce this "feature" of the old book? You're reproducing the look, but the word can't be searched as MAGISTRATVS.

BTW, the 1 is not the only substitution in MAGISTRATVS: the first S is also from another font, being shorter than the other letters and in a different style.

[dhinostroza]

Hi, Peter,
A ligature request for U+0133 LATIN SMALL LIGATURE IJ:

At first I thought the character was a lowercase y + diaeresis and forgot about it, but when I saw the ij ligature, it opened my eyes.