Call for design suggestions: Anglicana W

John Hudson

In terms of contrast with W/w, that seems to me very much a style-specific, and is the sort of issue we already face with some extended Latin and IPA characters that, in an italic font, might be easily confusable with conventional cursive forms of common letters.

Of your recent suggested forms I think a hybrid of forms 4 and 6 is possible, keeping the rounded bottom on the right, but straightening the middle diagonal. I made some pen experiments with this idea, in a couple of different size/modulation variations:



Although we're talking about typographic forms as distinct from the attested historical manuscript forms, I still find it helpful to see if a proposed form can be reasonably written, and what instinctive proportional relationships emerge in terms of counter sizes. This shape is interesting because — as in these examples — with a little forethought the letter can be written without lifting the pen; alternatively, both diagonals can be written beginning from the top with a broken ductus.

In a seriffed type, I would either have only a left serif on the middle stroke, as in some of your examples, or even leave it off completely.

Andreas Stötzner

Michael, could you post a small set of MS samples, showing the character in its context?
I need to watch this …

Andreas Stötzner

… what exactly is the carolingian w?

Andreas Stötzner

> I still find it helpful to see if a proposed form can be reasonably written, and what instinctive proportional relationships emerge in terms of counter sizes …

Very good point, John. (we seem to work rather simultaneously right now…

Michael Everson

John, the way you have written it in one stroke is exactly the way I am doing it in my notes on the printout of the manuscript I am preparing for publication. The issue typographically with a strong middle seems to be the interplay between it and the hairline that leads from lower left to upper right, as well as the angle the corresponding stroke might have for the Carolingian w. 

Andreas, I use the term Carolingian here to distinguish our (v+v) w from the (ỽ+ỽ) Anglicana w. The best way to see samples is to study the many examples in the proposal document http://www.unicode.org/L2/L2017/17238-n4838-anglicana-w.pdf. I have indicated Carolingian v and w as well as Anglicana w, and for some of the latter in those examples you can also compare the letter b.

Michael Everson

Michael Everson

With serifs, it's interesting. Unless the width of the character is to be made rather wider than the standard w, it's hard to see how a curve could work well, given the usual space between serifs in this font and the distance of the dark strokes in the w. 

Michael Everson

№ 3 and № 4 are a reasonable hybrids, though № 5 retains more of the Middle-Welsh w flavour. I think I am happy with the italic.

Michael Everson

Andreas Stötzner said:

Michael, could you post a small set of MS samples, showing the character in its context?
I need to watch this …

The MS samples are in the original proposal. http://www.unicode.org/L2/L2017/17238-n4838-anglicana-w.pdf

I will do up some prose samples (running the verse lines together) in due course, but I am travelling tomorrow and need to do some packing.

Andreas Stötzner

ah, thanks for pointing at the Pdf again. I just missed it …

Right now my favourite would be w_a6. If ascenders shall be, probably w_a9, in which I referred to the detailing of the j. A double full-strenght ascender with serif terminals (w_a7) feels rather unusual, but may do as well.

John Hudson

With serifs, it's interesting. Unless the width of the character is to be made rather wider than the standard w, it's hard to see how a curve could work well, given the usual space between serifs in this font and the distance of the dark strokes in the w. 

Given the complexity of the double bowl on the right, I would expect this letter to be somewhat wider than the w. But I would also be willing to forego the serif on the middle diagonal. If the ductus is understood to not involve lifting the pen, terminating in a rising diagonal stroke, rather than a descending one, then the middle serif can logically be dispensed with, which helps manage the width and the counter space.

Steve Gardner

This isn't a discussion I would usually dream of getting involved in, since my knowledge in this area is zero, but I think I'm seeing the construction of the anglicana w differently.

I've shown a single stroke, script-like version below, which shows the bowl/s being formed from below, not above; followed by two more formal versions (upper & lower?), which I've paired with Arno Pro for visualisation.  

Apologies for the crude forms: they were created in a hurry using Inkscape.








EDIT TO ADD: A more Arno-like version

John Hudson

As I suspected would happen eventually, Brill have asked me to add these characters to their typefaces. Still no easier to conceive in common roman and italic type styles than it was two and a half years ago. I'm leaning towards different constructions for the roman and italic letters, with the former being more formal and modelled on the treatment of the Welsh v.

Andreas Stötzner

thanks for showing, John. I find the two pointed terminals at the top tricky (especially the close gap), and the emphasized roundness does not ease it for me to get a ‘w’ feeling for it. On the other hand, your anglican w and v bear some obvious kinship.

The anglican w is also in the latest version of Andron Mega. I’m not unhappy with it, although, the anglican v/w kinship is hardly obvious here. – How important is it that these two letters bear some visible likeness? (a question for Micheal Everson, I suppose)

John Hudson

Michael hates my suggested form.

There are lots of different shapes for this letter in manuscripts, and I think Michael would point to most of them being less related to the shape of the Welsh v. That's fair, and I'm happy to consider other options. I was working from the basis that the letter was originally derived from the Welsh v and there are some manuscripts in which a clear ductus correspondence exists between the two.

Brill works as a typeface because >5,500 glyphs use a consistent set of methods to construct shapes. I try very hard not to introduce additional construction methods or shapes that don't already have some precedence in the existing set. It's a design that favours round shapes. I'm also trying hard to avoid shapes that significantly differ from the typical spatial frequency, which is why I really don't like some of the options for this letter: they have too much going on in too small an area.

Michael Everson

There's far more manuscripts than John knows about, and no, they don't all look like Middle Welsh V.

The forms from the Ordinalia do not harmonize well with standard Latin. The Anglicana W has many forms. While it derived from Middle-Welsh V (ỽ) it developed in a variety of ways including some that harmonize better with standard W. The Unicode charts have an interim version. I will be providing better glyphs. 

I just published a volume in which the Cornish Passion Poem has a palaeographic transcription, and I would recommend the shapes I settled on. I read Middle Cornish and I find the example shown above NOT to be legible.

Why were these letters encoded? For palaeographic use. Most editions of Middle English and Middle Cornish texts replace Anglicana w with w, just as most editions of Old English texts replace ƿ with w. Who will use the Anglicana w? People interested in reading palaeographic ME and MC texts.

While the Anglicana w appears to be in origin a ligature of ỽỽ as w derives from vv, it's still the case that ỽ on its own had a more limited range of use (Middle Welsh and Middle Cornish). The @-style Anglicana w is extremely common, perhaps even predominant, but one doesn't suggest using that.

What happened is that the two stems and the two bowls of the ỽ separated and became two halves of the new letter. Hers's an attempt to show something of the ductus. The @-style fifth arose from this.

At the same time the v and w were forming. There was a tendency to extend the left arm of both. And there was a tendency for the Anglicana w to borrow this same shape, keeping only its double bowl to indicate its history.

The letters w and Anglicana w don't distinguish phonemes. The distinction is of palaeographic interest where the two appear together in a manuscript. That's why they're encoded. And in that context they often appear side-by-side, and the later harmonized form is more legible.

Middle Welsh v acquired a typographic form long ago, and is suitable as it is, since it has been used also orthographically in non-palaeographic texts. Here I show the index of The Charter Fragment (ordinary w does not occur in that text).

That's all Baskeverville Bold. [sic] Here is an example, with 6 and 3 and b for reference, of plain, italic, and bold. Note that the ductus of the lowercase Anglicana w is similar to one a modern writer might use. The shapes are familiar enough not to make a reader stumble. It works!

One can build the lowercase Anglicana w out of v and ɛ and the capital out of the \\'s of the W. Note the verticals in the italic, and that their angles may not be identical; I follow the W and add a serif to the second \.

Michael Everson

Index screenshots above taken from:The Charter Fragment and Pascon agan Arluth (Corpus Textuum Cornicorum; 1). Dundee: Evertype. ISBN 978-1-78201-182-8 
http://www.evertype.com/books/ctc1-passion.html

John Hudson

I find the indeterminate sticky-up bit in the middle visually distracting in the context of a typical typographic featural grid. The lowercase t is the only letter that conventionally has this neither x-height nor ascender element.

Michael Everson

Fancy that! That's exactly what I chose. Anglicana w is an unconventional letter. But it conforms to the grid as partially defined by the t. Embrace it! The grid is there to help, not to rule. Anyway the t is as much a part of the grid as anything.

I've been reading it for a good while now, and it reads more easily this way than for instance the Andron one. Though the Andron one is more like my capital. 

Michael Everson

I find the "worty" example above perfectly legible. It's not distracting. And it does give a nod to the Middle Welsh V. (If you make it just as high as the Middle Welsh V then it does look distracting.)

John Hudson

Brill has the option of serifless terminals, which I use where I want a cleaner shape with less going on. How does this read?

Dreading the heavier weight.

Michael Everson

I find your r rotunda to be very unsatisfactory. While the half-R shape is found in some Fraktur fonts, in Latin the 2 shape is much easier to read. As to your Anglicana w, I can live with the stump (though I don't like it) but I really think that the hairline must join the top of the double-bowl. See your own drawing from 2018. 

Michael Everson

With regard to the r rotunda as used by modern medievalists, see the MUFI specificiation:

John Hudson

Round version. I suspect Michael won't like this either, but I'm trying to come up with a shape that breathes and fits into the spatial frequency of Brill text: that isn't a jumble of criss-crossed strokes of varying angles with a sticky-up bit that breaks the x-height.

John Hudson

BTW, Andreas, I like that you made the right side lean outward to mirror the diagonal angle of the left side, making the whole more w-like. In an renaissance style type like Andron that seems to make sense, as does the pointed bottom of the Welsh v.

Andreas Stötzner

Hello Michael! I’d wish you had shown us these samples earlier … 

I’m thinking about which features are mandatory, seen from the MS sources point of view:

middle-welsh v:

• the bottom being rounded (like in o or c) – yes/no
• the left ascender arm reaching full ascender height – y/n

anglicana w:

• the two arms reaching full or half ascender height – y/n
• the thin stroke crossing at the centre
  

And yet another question, Michael: the centre part of the angl. w, does it write top-›down or rather bottom-›up?

I would be glad to get these points answered to as precise as possible.

Andreas Stötzner

Andreas Stötzner

I made the welsh v more slender.

Of the two angl. w’s here, I lean towards the bottom one. It is like the one shown above, yet I tweaked the right part, raised the central stem and gave the whole form more space in order to get more pleasing counter shapes.

I think a more ‘synchron’ shape of the two stems is desirable. With the middle part bended (as in Michael’s Baskerville) the whole thing becoms very busy.

John Hudson

[The Brill r rotunda has been that way for 10 years, so is unlikely to be changed now, unless the client were to request it. The roman is a formal letter constructed following the principles of the rest of the typeface; the italic is a cursive form. See also:
https://typedrawers.com/discussion/comment/14942/#Comment_14942]

John Hudson

Ignoring the incomplete stroke on the right, I think this example provides a good model of how one can handle the short ascender without curving the whole stroke. This is the approach I would take in a renaissance style type like Andron with a translation stroke modulation.