September 23rd, 2010 byThe Ubuntu project is going full speed. We have started intial work on the Hebrew and would like to share a few thoughts with you. This PDF illustrates how we arrived at some of the basic proportions.
Pages 1 to 5 deal with the proportional relationship of the Hebrew height against the Latin x- and cap heights. From the start it was clear to us that a Hebrew to full cap height would be too big so we set it at 90% of cap height (T1). The other extreme is set just above x-height (T2). We then proceeded to create two variations in between to allow us to make an informed decision in regards to the correct proportion.
Examining the proofs in pages 1 to 5 our design team agreed that T2 and instance 400 are too small to stand up against the Latin. Instance 600 was closer and was considered but in the end we all felt that T1 provided the right proportional relationship, and also feels the right size within itself. Page 6 shows some sample setting of T1 and I feel it sits very comfortably on the page. Page 7 displays a glyph overview of the characters designed so far.
We are aware that the design of the glyphs still needs a lot of work but their design allows us to establish the basic grid, the proportions and how some design details are to be treated. At present we would like to hear your opinion in regards to its proportional relationship with the Latin, also bearing in mind that the Hebrew will be used within a GUI environment.
Bruno Maag
The toolkit
21 Responseshide comments
I always find these articles exciting to read. The creation of a new font, and the look behind the scenes is fascinating. You were so enthusiastic during your UDS talks that I’m really enjoying learning about the details that go into font design that I’ve thought of and excited to hear about that I’ve never even considered. Making a unified design that honors the typographic traditions of various character sets is even more interesting still.
Thank you for this series of articles. I look forward to seeing what the Hebrew community thinks about the height proportions.
Oh, I forgot to mention that I will be at UDS in Orlando. I am planning to give another talk, this time more a where we are at and what’s cooking now kind of thing. I certainly am still as excited as at the beginning.
During that talk I would like to get into some details, explain our reasoning for doing stuff the way we do. I am also aware that I had some critisism in that we’re using proprietary tools to make these fonts. I’d like to answer that too, with a call to arms. And if possible, I want to be available to speak face-to-face with people who can give us constructuve critisism.
See you there, then.
This is fantastic. I hope the next RTL language is going to be Arabic
I find these articles too. And sometime Jewish Community helped me to find some articles.
Yes, we’re already working on the Arabic, first trials. I hope to be able to put up some nice images for that, soon.
In respect to the Arabic, we’re also working on writing an enconding table which supports as many languages as possible. So, besides the basic Arabic, Farsi und Urdu support you’ll also get some of the Central Asian languages included.
The Arabic is going to be a Kufic/Naskh hybrid design, so it’s going to have a fairly strong horizontality. However, we are aware that a degree of ‘calligraphic flourish’ is needed to make it body copy friendly. This design decision means, however, that we’ll only include a basic ligature set, besides the standard ‘lam/alif’.
Great to see!
Really good initial work.
I really think that T2 and 400 are the right size, eventually most of the Latin letters you see on the screen are lower case letters and the hebrew text should align with them.
But as I said earlier, you must compare yourself to other common Hebrew fonts and not just to the relations inside this font.
For the glyphs, it seems like you struggle with the same problems I did when I tried to design it:
The “א” (Aleph) is very unbalanced, the handles are too thick comparing to the center line and the right handle doesn’t look in place. This letter is a tough one, many fonts fail to design it correctly.
The “ג” (Gimel) looks ‘depressed’… like a person standing with hands in pockets and the head down. I really think that my solution (ver 0.9 of my design) captured the right balance of this letter.
The “ה” (He) – in some of the smaller sizes the free leg looks a bit outside the glyph, maybe it should be a couple of points inside the curve, or maybe it should be shorter.
The top curves of “י,ו,ן” (Yod, Vav, Nun Sofit) feels too long, though your take is definitely better than the straight lines I did.
“ז” (Zain) – The top feels a bit too long. I think it should be 90%-85% of its current width
The “מ” (Mem) looks out of context. If you look at other Hebrew fonts, many fail in designing this letter and design a traditional glyph that usually doesn’t really look as a part of the overall font design. I think it should be somewhere between the traditional and the minimalist design I initially tried to do.
The “ק” (Kof)… Something there is not right. I think it is too square and kind of dull. I struggled with this one a lot, and I believe you are right with giving it a curve (Contrary to my tries), but it still needs some kind of a twist.
Last but not least the “ש” (Shin), if you look closely you’ll see that it looks good in larger sizes, but looses balance (mainly the width of the middle leg) in smaller sizes (look at T2 12/14). It looks good, but should be balanced a bit more.
Hope my points are helpful. Can’t wait to start and test this font
Hi Tom,
thanks for your feedback. It’s indeed very helpful. You are definitely right about the ‘gimel’ – that was my first impression too, when I saw it. We decided to get back to it once we are all agreed on the height proportion.
It’s interesting to year that you prefer the Hebrew aligned with x-height and your reason is perfectly valid. Personally, I still feel that it would be too squat but we’ll certainly rethink our position. One of my concerns is also how thhe font will behave within an UI environment and I feel that if it’s set at T2 or 400 it may be too small to be useful. It’s probably worth testing.
I hope to get more feedback from you as we develop the fonts.
Thanks for what is developing into a tremendous font!
Just wanted to raise another concern in support of T1 (or thereabouts). I find myself often employing Latin, Greek, Hebrew, etc. in academic writing and having large enough letters is valued for non-native readership (journals will even raise the Hebrew font size so that it is not T2/x-height).
Personally I find huge Latin letters sticking out of T2 sized Hebrew text (when in a Hebrew article) quite distracting as well.
Just some thoughts.
I like this a lot.
The only real problem, IMO, is the Mem (מ) – it looks way too low in some cases (see page 5).
A personal preference of mine is a Lamed (ל) with a higher top than what’s currently in the font but I don’t know if this would be doable.
Overall, very very good first release. Looking forward to the the next one.
it looks as the typographer don’t know any hebrew.
looking at good hebrew fonts will help!!!
http://odedezer.com/ezerfamily/
is a good place to start
An Israeli newspaper just published a lengthy article about Hebrew fonts:
http://www.haaretz.com/weekend/magazine/fonts-of-wisdom-1.315474
Fonts of Wisdom
Is it time to update the Hebrew alphabet for the Internet age? Do letters designed for the computer screen make reading easier? Or are these issues as irrelevant as yesterday’s newspaper? Linguists and graphic designers have their say
Nice
have you seen the video of Shai Golan’s graduation project comparing Arial to Frank Ruhel? Enlightening! Too bad it’s only in Hebrew.
http://www.200404.net/index.php?/projects/from-frank-ruehl-to-arial/
Hi,
I also think that the Hebrew characters should align with the lower-case latin characters. I believe that this is also how the Arial font is designed. Otherwise, English phrases embedded in Hebrew might seem a little small.
I also agree with Oron about the Mem (מ) being too low.
The Lamed (ל) also stands out as more straight than the other letters.
Overall it looks like a good font, and I’m looking forward to see its development.
Hi all,
we’re actually reviewing the Hebrew as we speak, conferring with my colleague Oded Ezer (mentioned in an earlier post). Oded, too, prefers a lower height of the Hewbrew and growing closer to the design we, too, feel that the full height T1 is maybe too large. We’re now looking at 400 again, and maybe even create a version that sits at 500.
One of my major concern is how the Hebrew is going to behave within an UI environment. A low height will compromise legibility and its usability at small sizes. I am also concerned with the texture of the setting when it’s placed against the Latin. Indeed, the majority of Latin text is lowercases but Hebrew uses the Latin/Arabic numerals. But I guess these can be designed to match the Hebrew, accessible via OT language specific feature.
We’ll post more visuals as soon as we’re ready.
W9MAZC HHIS I should have tghohut of that!
Bruno, I’d glad to hear that you’re reviewing the height issue. To me, mixed text where the Hebrew top line doesn’t align with the Latin one produces dissonance. When I looked at the T2 examples, on the other hand, it actually looked right, even if the x-heights weren’t identical, since there was the illusion of alignment.
Have you seen Simon Prais’s dissertation (which he mentioned in the previous Hebrew post[*])? He had some interesting images of how printers resolved the x-height issue, and the best-looking ones had a Hebrew x-height that was slightly greater than the Latin x-height (such as on page 4).
I think there are a couple of other reasons to keep the Hebrew x-height smaller. The first is to have an aesthetically-pleasing ל. That extra space can be used to restore Hebrew’s sole ascender to its proper glory, rather than cutting it off like an unwanted typographic complication. The other reason is the numbers (which Simon also mentions in his dissertation). If the x-heights align (more or less), the font’s old-style numerals work quite well with both Hebrew and Latin. Even the modern-styled numerals look good this way. With the taller Hebrew characters, however, both styles of numerals look out of place with Hebrew.
[*] http://design.canonical.com/2010/09/which-way-to-slant-the-hebrew/#comment-3398
One last point: having shorter Hebrew characters doesn’t necessarily mean having text that is harder to read. A good design can make those characters more legible at smaller sizes. I think you’ve already made a good start, especially with the way the tops of the letters curve like their Latin counterparts. I’m, therefore, confident that you can solve the legibility issue.
David, thanks for your confidence in our abilities. I hope to be able to post an update of the Hebrew soon but there is still soooooo much other stuff to be done for the Ubuntu font family.
I tried to access the dissertation but had no joy. I registered but was never sent the file, unfortunately.
Thanks. I can send the file to you, if you’re interested. Hopefully, Simon won’t mind. I don’t know where to send it, but you should have my e-mail address through this blog.
This is truly fascinating. They really look good, the Hebrew letters. I can’t wait to see it used in my desktop!
Good job, everyone.
The Ubuntu Hebrew font, last I’ve seen, looks like Hebrew script that is only pretending to be Hebrew print. Hebrew script is for handwriting only while Hebrew print is for almost everything else, including computer interfaces. Instead of taking the Ubuntu Latin font and modifying it to resemble Hebrew print, Hebrew print should be used as the base and then modified to resemble Ubuntu Latin.
I looked at the PDF. Even though I can’t put my finger on what’s wrong, the font just doesn’t feel right.
Happy New Year and I hope you have a great year in 2011!