Indeed this is quite a difficult design problem. I would recommend reading my article "Latinization: Prevention and Cure"* published in issue #4 of the Austrian magazine Spatium (as well as the 5th issue of the Greek journal Hyphen) and studying my "multilateral" solution for Armenian and Latin, a typesystem called Nour&Patria**. You could say it took me two decades to arrive at what I did; I think it works, and I hope it can inform the development of almost any multi-script type family.

* A preview here: http://themicrofoundry.com/ss_rome1.html

** From the Creative Review awards:
http://themicrofoundry.com/ad/CR.pdf

The first thing I would decide: Is the Latin meant for parallel setting, or subordinate setting? My view is that a single font can't do both - a more complex system is needed.

hhp

Thanks Hrant, Your website has given me a great deal to think about!

My first DECISION is that in Mike Hebrew font, the Latin should be subordinate to the Hebrew. Actually I thought this all along... but maybe not in words.

How should this effect the design of the Latin characters? Again it would be helpful to have explicitly formulated ideas on this. Much of it up to now has been intuitive.

Mike

Michael Cunliffe Thompson
Seascapes and Landscapes of New England and...
Hebrew Calligraphy at http://cunliffethompson.com/font

This is hard, but I would look to models of more informal designs, with some of the variation in baseline that you have in Mike Hebrew. Examples are Peanut Sans, Tempus sans and Burbank.

> How should this effect the design of the Latin characters?

The way I see it a subordinate font can sacrifice quite a bit of cultural authenticity - but only with the purpose of better serving the master design. So for example Patria's Armenian subordinate has full serifs (which the Armenian script really doesn't like). Another example is the Latin "g" below from a grot Cyrillic-dominant typeface I was commissioned to design last year. It's made to echo the form of the Cyrillic "б".

For serving a traditional Hebrew type I might recommend a subordinate Latin where the stroke contrast is flipped (thicker horizontals than verticals*) but Mike Hebrew is monoline (which you could say makes things easier). Another central parameter is vertical proportions: Latin likes to have an x-height within a certain range, but to avoid a subordinate that looks too large/small in comparison to the master script you can move quite a distance away from that range; in the case of Hebrew that probably means a Latin with quite a large x-height (which you seem to be shooting for already).

* See the work of Excoffon and Bloemsma.

--

William, depreciating the baseline is quite a fascinating idea.

hhp

>depreciating the baseline

Ada Yardeni in Hebrew Script points out that traditionally Hebrew is "hung" from the top, with the baseline not so strong. The baseline was made stronger in Ashkenazi scripts, perhaps under the influence of Latin scripts.

Mike Hebrew takes advantage of the fact you can play with the baseline in Hebrew without penalty--except for making nikud more difficult to place.

Playful designs like Peanut (that link should work, the above does not) are able to get away with an irregular baseline in latin. I don't understand how exactly they are able to pull it off, but you might ask typophiler Michael Clark, who designed Peanut, about the challenge.

Because of the informal mono-line, I think potentially the latin can be very strong also, and still harmonize. This would be a real achievement, I think--to have a latin and hebrew that really work together seamlessly. If it were me, I'd start by mirroring your kaf and using that as a c, and building the look on that. As it is, the very round sans you have doesn't work at all, IMHO.

Wow, thank you all. I now have a lot of ideas to work on,
Mike

Michael Cunliffe Thompson
Seascapes and Landscapes of New England and...
Hebrew Calligraphy at http://cunliffethompson.com/font

I have a lot of test documents in Hebrew with some English text.
My first goal was to balance the size and boldness of the English so that it
was unobtrusive - usually acheived by having the Latin characters a little
larger and a lighter. The overall effect should be for the Hebrew to stand
out on the page even when an English translation with more words will
occupy more space

In no way did I want the Latin characters to look like Hebrew ones. I avoid
giving the shape of a Hebrew character to a Latin character. My approach
was to make a small change in the Latin and then review all my test documents
to see if there was an improvent. For instance, the first change was to tilt the
center bar of the lower case "e". The change was carried out across the
six members of the font family. I was amazed that an improvement could be
seen in the appearance of whole pages after such a small change!

The table below shows which Latin characters were changed.
Further changes suggest themselves but will they improve appearance and not
diminish readability?
Mike

> In no way did I want the Latin characters to look like Hebrew ones.

1) I actually think there is functionality to be gained by doing so.
2) How then do you rationalize the changes you made?

hhp

If you show the proposed match mixed in the same line, that will tell us a lot.

Your roman characters seem awfully narrow compared to the Hebrew. I don't see how you are going to get the same "color" meaning density of gray, which is a minimum.

I agree with Hrant that making the roman more angular, like the Hebrew, is more likely to work. I also don't see why the roman should be lighter. But I'll be happy to be proven wrong.

Hrant,
There may be valuable possiblities in the approach you suggest. Certainly in brief bilingual presentation such as product name or title page, a closely related visual
style would be attractive. However I'm thinking of my Latin as being a practical
workhorse to get the Latin out with minimal baggage. The user doesn't have to LIKE the
design - ideally it should not even be noticed.

William,
The very narrow roman characters are of value in displaying translated text where the
English is so much more wordy and the Hebrew so very brief.
If you look again at my first post, I think you might agree that the English text in
the example blends in well with the Hebrew... or maybe it doesn't. What do you think?

Mike

"Blends in well" is a question of: compared to what? Generally there's a huge gap between latin and hebrew scripts. So better than many, but still far from being really harmonious.

Stylistically, they are very different, even though the monoline for both helps immensely. There are a lot of very successful squared off sans--my favorite is Klavika--and my feeling it that something like that would have a shot at more harmony.

I don't know about the squashing. I think it would be better to get them really harmonious. If squashing helps, fine, if not not.

Just slightly off-topic - you might like to view my Fontstruct creation HebRoni:

http://fontstruct.fontshop.com/fontstructions/show/273980

If I say "slightly off-topic" it is because I did exactly what Mike is trying to avoid: "In no way did I want the Latin characters to look like Hebrew ones". But maybe it could give some people some ideas...

I tried to upload an image comparing similar letterforms, but failed. Anyway, they are:

פמטaeg בt אx ץy כc סםo רr שmw

Have fun - and tell me how to get the pesky niqqudim in place! :-)

Ronald Kyrmse
http://kyrmse.googlepages.com

Mike,

You have here a display font (the Hebrew font); do the same thing with the Latin;

Ronald Kyrmse,
Thanks for your input on structures. Here are the characters you suggested.
Mike

Michael Cunliffe Thompson
Seascapes and Landscapes of New England and...
Hebrew Calligraphy at http://cunliffethompson.com/font

BTW, since Hebrew doesn't mind rightward slanted forms, but Latin very much does, I might make the Italics slant the other way (especially if you are in effect wanting to make the two scripts equal in importance).

hhp

David Hamuel,

So if I have a display font and...
we accept that:
Display fonts developed as advertising types, and are intended for limited use
at larger sizes, to catch readers' attention. Their readability is not their prime
function, and should not be used for the longer passages of text.

Should I design a matching display font for my Latin characters?
I is a big undertaking and as software engineers like to say it is 'scheduled
for a future release'.

After thinking about the posts and looking at the other fonts suggested, my current
Latin set looks less and less attractive. However assuming I will release
Mike Hebrew v33 quite soon, I wonder what changes should I make to the Latin
characters I have right now.
Mike

Michael Cunliffe Thompson
Seascapes and Landscapes of New England and...
Hebrew Calligraphy at http://cunliffethompson.com/font

@Mike:

In fact, I was trying to show how my font treats similar shapes, e. g. [teth mem pe a e g], but couldn't upload my image file. Anyway, you get the idea, and if you go to the trouble of downloading HebRoni you will see how (in this case) the same basic shape does duty for six glyphs.

... and I am curious to see the next version of Mike Hebrew!

Best -

Ronald Kyrmse
http://kyrmse.googlepages.com

Good readability however comes from contrast between shapes, not
a conscious -and essentially aesthetic- adoration of uniformity.

hhp

Mike, I think David, who is a native speaker and designer of excellent Hebrew types, is right on the mark. Your Mike Hebrew already is a display font, as it won't work for extended text. So a matching latin that is much more strongly 'flavored' is a more appropriate match.