We are in the final stages of completing the Ubuntu font suite. Unhinted monospace fonts were revealed at UDS in Budapest and seem to have gone down quite well from what I can gather. We currently furiously hinting the fonts to make sure that their appearance on screen is meeting expectations.
We are now putting the two complex scripts Hebrew and Arabic together. Although the basic design of the Arabic was finalised some months ago we have been doing a lot of background work investigating language support, and thus defining a glyph set. This work has led us to have aprox. another 1,000 glyphs in the font supporting languages such as Kashmiri. Our designer Jonathan Pierini currently has the task (some would say unfortunate) of compiling the various glyphs and their diacritics. Whilst doing this we have come across a few situations which we hope someone from the Ubuntu community can help us with.
In Kashmiri the glyph ‘waw with ring’ is used. In the Naskh, Nasdaliq and similar scripts the tail of the ‘waw’ often stretches out downwards which gives the designer the opportunity to attach the ring to the tail on the inside of the character. However, Kufic style fonts like Ubuntu tend to have a more closed ‘waw’ tail shape making it difficult to place the ring on the inside of the character. We have now created a versions that has the ring on the outside of the glyph but are uncertain if that creates a legibility issue. I am of course aware that experience of Latin design cannot be directly transferred to non-Latin design but looking at the Danish a-ring for example, I know that there is some room for interpretation. Normally, the ring is placed above the lowercase glyph, as a separate element to the ‘a’. But where space is tight the ring can be part of the ‘a’ as long as a ring shape is visible and the character can be correctly identified.
A Naskh style ‘wawring’ on the left, Ubuntu ‘wawring’ on the right
Kashmiri (and possibly other languages) also use a wavy hamza. Again, we’re uncertain if this is a distinctly different character to the Arabic hamza or is it a stylistic variation that is script dependent meaning that a standard hamza is acceptable.
We would love to hear from you on this as these are very detailed questions yet there is little information to be found. We want to make sure we can provide as comprehensive and high quality a font suite as we possibly can within the limits of our current remit.
Bruno Maag
Hey,
I thought the post about Arabic but it seems all about Kashmiri .. I hope the title will be changed to something clear to make lot of people participate ..
BTW, this is not Hamza, if you write this in Arabic it will be understood as “ع” ..
Good work ..
Thank you
Hello,
I did check out the font some time ago. It is one of the most illegible fonts I’ve ever come across. This is no joke, but an informed statement. Arabs in UK, even if they are professors, are not a good side to ask about legible Arabic fonts. Go ahead and use it by default, I bet most Arab users would change it to be able to read comfortably.
We appreciate the effort, but going cheap on some languages does more harm than good.
(Sample for those who want to confirm what I said: Ubuntu-Arabic.pdf)
I agree with both comments. Asking Arabs about the Kashmiri language won’t help you as it’s not really Arabic, but an Arabic derivative.
How much user testing have you done of the Arabic version with actual Arabs? I strongly suggest you contact the Tunisian Ubuntu group. Also, remember that the Arab world is huge with a lot of variety from Morocco to Sudan to Lebanon and so on. And that’s not counting all of the Arabic derivatives like in Afghanistan or Iran.
I kinda assumed that the Ubuntu font would be simple, a bit like Arial except with a bit of artistry. I don’t think most Arabs are used to reading big stretches of a Kufic style.
Better link: http://design.canonical.com/wp-content/uploads/2011/01/Ubuntu-Arabic.pdf
As an Arab, I find the font very pleasing to the eye. Very nice work. I have a couple of points though:
- The Haa letter (first letter in the sentence) is too wide.
- The Seen letter (first letter of last word) may be hard to distiguish from certain letters that come before or after it. Like in the case of the last word.
With regards to Jeremy’s comment, formal Arabic writing is identical in every way in all Arab countries. And other languages like Persian and Kurdish use the same alphabet but with a few additional letters. Much the same way French has similar alphabet as English but with a few additional letters.
http://www.brown.edu/Departments/CLS/people/facultypage.php?id=1143055226
He might be a good person to contact about the Kashmiri script.
Saa: Thank you for the first point; I’ve updated the title of Bruno’s post to include the word “Kashmiri”. Just as Latin script is used for writing more than just Latin (we’re using Latin here to write English), Perso-Arabic script is also used for many languages: Kashmiri and Pashto being two languages with letters that exhibit the ring issue Bruno/Jonathan are asking about here (U+06C4 ‘ۄ’ and U+06BC ‘ڼ’).
Kashmiri is also a particularly good test of the Arabic script support because full vowels/diacriticals are used by default, compared to written Hebrew or Arabic where vowels are dropped by default. Of course, Kashmiri is also written in several other scripts, and hopefully in the future the Ubuntu Font Family will be able to support written Kashmiri using Devanagari too.
Thanks for your input about Hamza/Alin (ع/ﺀ), do you perceive confusion coming from the similarity of the top parts of the letters? Is there any that you can suggest that might help reduce the ambiguity? (You are welcome to draw some ideas if it would be easier to convey).
Saa, I can see that the title of the post was somewhat misleading. As a typedesigner I usually don’t refer to language but to the script. As Paul has already detailed, the Arabic script supports many different languages, amongst them Kashmiri (also mentioned by Zaid).
Re Hamza – apologies for not being clear. I referred to the Hamza mark, above and below ‘alif’, not the actual letter Hamza.
Anonymus/Jeremy: the fonts were designed not only with the help of Rayan Abdullah but also with people from the Ubuntu community. It is also important to understand that the fonts not only have to work in body copy but are supposed to become Ubuntu’s localised menu fonts. The decision to design the fonts in a Kufic tradition, with some softer elements is the right one in these circumstances.
The fonts do read pleasantly in larger amounts of copy but no-one suggests you should set a book with them, as I would never suggest that the Latin is for that purpose. The aim of the Ubuntu font suite is to create a contemporary design where the various scripts are sympathetic to one another and still maintain some of their own traditions.
Arabic, Farsi and Urdu readers are indeed used to reading Naskh style (as well as Nasdaliq, Thuluth and Diwani) fonts in everyday life. These are designed for body copy but would be unsuitable within an UI or headline environment. Again, it comes down to the area of usage that determines the look and feel of a typeface.
Bruno
Hi,
I’ll be giving my opinion as a native Arab. First, the Kashmiri ‘hamza’ does’t exist in Arabic so using it instead of a real ‘hamza’ isn’t the correct move BUT there is something similar in Arabic, it’s called ‘Hamzat-o-lwasl’ or’همزة الوصل’ (U+0671 ‘ٱ’ ), it is mainly used in the Quran(Othmani).
I’m using Google’s “Droid Arabic Naskh”, it is a highly readable font and it confirms what I’m saying, it has both the Arabic ‘hamza’ and the Kashmiri ‘hamza’.
Second, I want to give my opinion about the Ubuntu Arabic font. As a designer, I use Arabic fonts from every category and I confirm what ‘anonymous’ said, it is unreadable(in some areas).
The font itself is good…for titles. but at 8~9pt it becomes a pain to read. I see three problems, the diagonal points are not a good choice, the hamza in the first word from the last line is so small and finally, there are some spaces that shouldn’t be that wide (2l 2w, 3l 3w).
“anonymous” / “
$email“: (nb. fake/mistyped email address) Thank you taking your time to contribute. Regarding the perceived readability, it’s much easier to respond to specific points—as Zaid has given below regarding Haa and Seen—than vague hand wavy statements. I would like to follow-up in detail, but without a name or contact details it’s much harder…Regarding professors, the professor behind the Arabic is Professor of Typography Rayan Abdullah based in Leipzig, Germany and who grew up in Mosul, Iraq. You can read about his visit to the old Dalton Maag offices in London on an earlier post.
I have not had the privilege of meeting Rayan Abdullah in person yet, but here in Berlin (where I am at the moment) his influence can be seen at every tram, bus, U-Bahn, S-Bahn and ferry stop of the Berlin Transport Authority (BVG). …and on the back of many of the fast German cars stuck in the traffic at rush-hour! (As well as the German federal authorities who try to stop them when it’s not rush hour). There more third-party write-ups of his influence at work: in English, auf Deutsch, باللغة العربية.
I would be more than happy to follow-up with you in detail about your concerns (the point about being an open font project is that we can and will try to fix things where there is a genuine issue), but it would be a lot easier if you could share genuine contact details!
Unless it’s been fixed very recently, there’s still a bug in the lowercase “w” in Windows at low point sizes. The right-most stroke of the w gets squeezed into the rest of the letter. Example here: http://goo.gl/jN6e9 (14pt shown).
I know this isn’t the proper place to post a bug, but I figured you might want to see it before the font is completed.
“Mail (will not be published) (required)”
Cool staff we have here.
Ryan: It’s only possible to fix bugs that are known about and documented in the bug tracker (much as we’d all like it, bugs can’t magically fix themselves). Thank you for spotting the hinting issue with the ‘w’; please could you file it in the bugtracker at +filebug?field.title=Hinting: lowercase w on MS Windows along with the exact version of Microsoft Windows that you’re using.
Please could you also copy-and-paste the two magic characters “” before taking the screenshot. This will give the version of the font that you’re testing with and also the state of the hinting engine. What matters is the PPEM (size in pixels) rather than the point size—the relationship between the two varies based on the monitor resolution (DPI) and the pixel size is the one that is important for debugging.
Ahh, the font has had some revisions since this last issue. The points are now horizontal after we did some testing at small sizes and some more fine tuning has been done, in collaboration with a number of native Arabic designers.
Re the Kashmiri ‘wavyhamza’ this is in Uni 0672 and 0673 in connection with alif.
I can assure you all that it is a good font to use. :-)
Bruno
SMTP error from remote mail server after RCPT TO:<$email>:host
$domain.com [$ip]:550-5.1.1 The email account that you tried to reach does not exist.
I’m more than happy to forward a copy of the personal email written to you earlier this morning that generated the above error message. Either the email address provided was a typo, in which case pasting it clearly here will hopefully help you identifying the mistake (since you won’t be able to see it yourself after submitting)—the real working email address can be kept secret. …or alternatively, an intentionally fake piece of data was added in the first place.
I would really like to have a detailed follow-up conversation (I try to follow on all of the bugs reported until we can get to the bottom of them). If you don’t wish to have a public conversation you may contact me privately using any of the email or IRC details listed on launchpad.net/~sladen.
Although, I far prefer having conversations in the open where possible since it helps to track things (everyone can see the same discussion). It is however very hard to have a discussion with multiple “anonymous” contributors in the same thread. Part of the personal email was enquiring if you had a preferred way of disambiguating the multiple names:
Updated: after reply below to (1) include email blockquote giving background for having an consistent identifiable name when holding a conversation; (2) to redact email address now that the poster has had a chance to check if it was correct or not; (2a) same for post two-up; (3) to give alternative contact details because the thread is nested too deep to reply in-line;
If the email was faked as you claim, does this give you the right to post it eventhough you promised not to do so? if you can’t discuss things in the open, then don’t do it. Asking people “questions” and replying to them in their emails beats the point of this blog.
Lack of transparency does not help anyone, especially you.
Thanks for the help. Here’s the report: https://bugs.launchpad.net/ubuntu-font-family/+bug/801716
Happy hinting!
استمروا انا مبسوط من الخط وعندي احساس انه حيكون شيء رائع واضافة لمجتمع البرمجيات الحرة وللفونتوغرافيا العربية :)
شكراً كانونيكال
I am very pleased with the font and i know isA that it’ll be incredible and will be an addition to the Free software community and Arabic Fontographia
Thanks Canonical :)
Mohamed: Thanks for the positive comment (and for its translation). Also spotted the logo on your RavingGeek website using the Ubuntu Mono Beta as its typeface. How are you finding the monospace?
Would you be willing to help do some more detailed testing of the Ubuntu Arabic in a user-interface environment?
He is using Beta Phase PPA .
I send to you an email regarding testing the font a few weeks ago :)
Also I suggested to you some people to help with the Arabic font like Abdoulla Aref .
http://design.canonical.com/2011/01/hebrew-and-arabic-on-track/comment-page-1/#comment-15317
I do not thing he saw your replay in that article :)
Mohammed is a good choice :)
this is his launchpad page
https://launchpad.net/~alam-hor
The Monospaced font is really one of the most comforting fonts I’ve ever used. even it beats Monaco,Dina and Consolas very clear msA.
I’ll be glad to help anyway :)”there tons of fans over LinuxAC :) can’t wait to test it :))
Hello,
I am a native Kashmiri speaker and normally use devanagari script for reading and writing Kashmiri. I can read and write nastaliq Kashmiri, but cannot call myself an expert. Therefore my input below must be confirmed by experts:
1. Please use the document at http://tdil.mit.gov.in/KashmiriDesignGuideOct02.pdf for reference. See alphabet chart on right column of page 1.
2. Waw with ring
In the “waw with ring” that Kashmiris use, the lower ring appears at the end of the tail. Please see the leftmost letter in group 6 of above pdf (page 1). Waw is never followed by a connected letter (on the left), so that should not be an issue. Both the shapes shown in the above post seem to present legibility problems, although the one on the left is better. Just make the lower ring at the end of the tail.
3. Wavy hamza
Kashmiri language uses a regular hamza and an additional wavy hamza. The wavy hamza is similar to the one shown in the above post.
Please also see the document linked below:
http://std.dkuug.dk/jtc1/sc2/wg2/docs/n3673.pdf
This document includes references to some experts at the end.
Thanks Bruno and Zaid for the followup. I have more trouble reading other Arabic script styles than what I’m used to, so it’s good to hear that it’s still comfortably legible for native Arab readers.
I have a tiny suggestion that I feel is really necessary. The three upstrokes of the ‘seen’ should gradually decrease in height. That way, we can tell the difference between a ‘seen’ and another normal character like ba or ta…
I loved the Ubuntu font. But when you will introduce Bengali??
I am a font developer from Bangladesh. It would be a pleasure for me if I could do anything for you.
I’m a Persian. Personally I don’t like the fact that dots in the letters with two dots are slanted. I mean, you can see a lot of dot pairs which are like .` but originally they are written like .. and I think changing this norm is not a good idea, at least for Persian (which has almost the same letters as Arabic).
Anyway, many thanks for designing a new font :)
Masoud: Thanks for the feedback, as Bruno Maag notes above, the dot angle has been changed after testing with some Arabic designers to a flatter one. I’ll try and get a snapshot to demonstrate the difference.
Would you be interested in helping to test the font itself on your desktop?
Siyam: The answer is mainly when somebody comes forward with the expertise! You might well be that person; there are two important things are needed for a particular script or language system to happen for the Ubuntu Font Family:
The font designer will need to work with Dalton Maag to ensure that a consistent style is maintained across the typeface. The testers need to also be reliable and be prepared to give good-quality on-going feedback (positive and negative).
Hi,
I am a native Persian speaker and a Linux / Windows programmer, and would like to help you with Persian Alphabet. It is very nice to have a good quality Persian font in ubuntu!
I think Siyam would the right person for this job. He got some extensive experience on Bengali font developing on both Free & Open Source and Professional purposes.
Siyam: Let me know if I can lend you a helping hand in anyway. :)
Yes, I typeset Persian texts almost everyday and I would like to test this new font to send feedback. I downloaded the fonts from launchpad, but apparently there is no Persian/Arabic letters in it. Am I doing something wrong, or I should get the fonts from you?
I’m a native Persian and very interested in Ubuntu Perso-Arabic font for whether using or testing, but I can’t find it anywhere for download. Can you please help me?
Oh, it’s very late, but an example of a good simple legible font for Persian/Arabic is Koodak:
http://cooltext.com/Download-Font-%DA%A9%D9%88%D8%AF%DA%A9+Koodak