Re: OpenType update for Unicode 5.2/6.0?

From: Saqqara (
Date: Fri Oct 15 2010 - 17:59:08 CDT

  • Next message: Markus Scherer: "Re: Derived age regexp"

    Thanks Peter for your comments.

    It would be interesting to hear from whoever is responsible for the OT registry but good to know they have the tags.

    I agree there has been good progress, much though in a perfect world everyone here would like to see a faster pace. Indeed the history of Windows going back to the NT 3.1 introduction of Unicode APIs has been of great value in promoting the use of Unicode in applications. More recently, the .Net, WPF and Silverlight development stack with Visual Studio has encouraged Unicode thinking and multiscript applications. I've been pleased to see developments in mathematical layout. Microsoft has been getting some things right.

    Occasionally we have pleasant surprises, such as when I recently observed the 'experimental' web browser on Kindle has font support for the egyptological aleph and ain transliteration characters introduced in Unicode 5.1.

    Of course you are right to highlight the fact there is really no such thing as 'Unicode n' support, lots of components have and applications natural lifecycles and priorities need to be decided on. There remains much to be done, some aspects of which need the long term perspective.

    However I singled out the OpenType topic as a relatively straightforward side to all this. Same goes for basic character output etc. in most cases.

    Some specific points on Microsoft Windows products and my perception that the pace has changed:

    1. All the browsers have taken Web Open Font Format (WOFF) on board (unlike the situation with EOT) so it is now possible to construct web pages using specialist fonts and reach a small proportion of internet connected devices, a proportion that will increase rapidly over the next year or two. Opentype is therefore very much an important dimension of HTML5 whether on not everyone realizes it yet. We can anticipate a change in user expectations from the current situation where people just accept 'unknown character' and substandard typography as a fact of life on the web. A game changer that IE must address to compete.

    2. Surely the Unicode 6.0 symbols will appear in the wild rather rapidly, especially in Japan. Android, iOS and WP7 will ecounter commercial pressures to provide support and compete with one another. A very different dynamic to the more leisurely approach on the desktop found in the past with OSX and Windows releases. WP7 sure doesn't have the luxury option of waiting a couple of years.

    3. The corporate and public sectors that have been slow to upgrade from XP are unlikely to be impressed by a multi-language strategy that requires a Win7->Win8 upgrade. I can't imagine there can be much need for discussion of focus groups in Microsoft as to whether this is an SP candidate feature.

    These and a whole bunch of other factors underly my comments about there being evidence of a change in pace to Unicode deployment, whether or not your managers and execs realize it yet.

    You alluded to there being still a limited understanding in this broad area. Too true. I continue being surprised at how many software developers still think Unicode is a 16 bit character set. I'm not sure what we do about it other than raise the points in forums like this and hope the message gets around sufficiently to stimulate some competition.

     Bob Richmond

      ----- Original Message -----
      From: Peter Constable
      To: 'Saqqara' ; Unicode ML
      Sent: Thursday, October 14, 2010 5:16 PM
      Subject: RE: OpenType update for Unicode 5.2/6.0?

      The tags will likely match ISO 15924 IDs. I don’t maintain the OT registry or work on the OFF standard myself, so I’m not sure what the practice is. I’ve just delivered the proposed tags.


      Win7 supported Unicode 5.1. I’m pretty sure Vista pre-dated 5.0, but I don’t recall now.


      When I say “supported Unicode 5.1”, keep in mind that that doesn’t mean that every component in the product supports every language / script / character supported by Unicode 5.1. Given the size of Unicode, it’s a simple reality that some things take longer and cost more to implement than others. For example, updating default collation is much simpler and less costly than supporting text display for a new script, and that’s simpler and less costly than supporting speech or handwriting recognition. When we ship the next version of Windows, default collation will be updated at least to Unicode 6.0, and you can bet we’ll support the Indian rupee symbol, but there will still be many parts of Unicode 6.0 that are not supported in a number of components.


      Note that an industry emphasis on standards does not necessarily imply that all standards are equally important to all people. For instance, I was speaking recently to someone whose product was subject to certain compliance requirements for a certain major government, and apparently they were required to document all functionality that related to some standards, but not others that were clearly also very significant for that product. In the Web world, there’s a lot of talk about standards, but the focus there is on things like HTML5, CSS3 and SVG, not Unicode. Regrettably, as much as Unicode has developed and been embraced, there’s still a lot of ways in which understanding is limited. There have been some important wins with certain governments committing to Unicode rather than endorsing non-Unicode encodings, but they are not necessarily as advanced in understanding all the other things needed to support the languages of their country. A great example here is the complete lack of clear conventions for keyboard layouts in Africa. Things like that make it harder for implementers to know how to design components and mean it’s far less likely that managers and execs will give priority to and fund work needed to support those languages.


      So, I haven’t seen indications that the increased emphasis on standards necessarily translates into quicker response in supporting all of Unicode. Windows has for several years been further along than other OSs in the number of scripts that are supported. I can’t speak to what the priorities are for the next version of Windows, but I can tell you that Mandaic isn’t even close. I’d speculate that for Apple there are some priorities are around getting adoption of iPhone and iPad in additional markets, but that Chinese will be given more priority than Hindi, Hindi more priority than Amharic, and Amharic way more priority than Mandaic. And I’d speculate that Google is in a similar mode.


      In other words, we’re still grinding along.


      Let’s keep this in perspective: consider just how much progress there has been in the last ten years. Windows 2000 supported 13 scripts to some degree; Windows 7 supports a lot more characters for those scripts, and also supports 28 additional scripts as well as a large number of symbols, including support for rich layout of math formulas. Products like Windows, Office and IE are getting localized into around 100 languages. We have IDNA and countries are getting TLDs in native scripts. CLDR has data for something like 186 languages and 159 regions and is growing.


      Still, I’m with all of you that wish things progressed more quickly.







      From: Saqqara []
      Sent: Thursday, October 14, 2010 6:27 AM
      To: Peter Constable; Unicode ML
      Subject: Re: OpenType update for Unicode 5.2/6.0?


      Pleased to hear the tag list is being updated. Will the 4 character tags follow ISO 15924 (as I'm assuming)? If so at least we all write code and mark up fonts accordingly then sit back and wait for system software support, browsers etc. to kick in.


      My reading of the OFF standard is the tag set can be extended by the registrar (i.e. Microsoft) without the need for a new rev. of the full ISO standard. Is this correct?


      Traditionally, support for new versions of Unicode has been slow in coming, in part for product lifecycle reasons. For Microsoft, Unicode 5.0 for Windows was associated with Vista release if I'm not mistaken (which I may well be!). Anyhow, seems to me now that with the internet browser standards focus of IE9 etc. and the flood of mobile devices - Windows Phone 7 (and for all I know Windows Slate 7) v Android v iOS etc. - this tradition has come to an end and a more agile approach is needed for commercial relevance. I mention this on Can you say anything on this topic from the Microsoft perspective? Love to hear any views from Google and Apple too.



      Bob Richmond


        ----- Original Message -----

        From: Peter Constable

        To: John H. Jenkins ; Unicode ML

        Sent: Tuesday, October 12, 2010 9:39 PM

        Subject: RE: OpenType update for Unicode 5.2/6.0?


        We are in the process of updating the tags to sync with Unicode 6.0. This has to be coordinated with the ISO Open Font Format standard, so may take a little time.





        From: [] On Behalf Of John H. Jenkins
        Sent: Monday, October 11, 2010 9:32 AM
        To: Unicode ML
        Subject: Re: OpenType update for Unicode 5.2/6.0?


        You might start with at


        On Oct 11, 2010, at 5:11 AM, Saqqara wrote:


        Given that OpenType is the de-facto standard for fonts, it is disappointing to see the 'Script tag' list for OpenType has not been updated in almost three years. I'm a patient person but the lack of inclusion of new scripts in Unicode 5.2 a year after the fact seems like carelessness. I've elaborated a little further on my jtotobsc blog, see


        My particular interest being 𓌃𓂧𓏏𓏯𓀁𓏪𓆎𓅓𓊖 (mdt-kmt, the Egyptian language in hieroglyphs).


        Any ideas who needs to be prodded to make an update happen? It would also be very useful if HTML5/WOFF could spec Unicode 6.0 or later as a step towards a multiscript web.


        Bob Richmond






        Siôn ap-Rhisiart
        John H. Jenkins


    This archive was generated by hypermail 2.1.5 : Fri Oct 15 2010 - 18:02:11 CDT