[MPEG-OTSPEC] [EXTERNAL] Re: New AHG mandates and other news!

Tue May 11 18:50:24 CEST 2021

The ‘name’ table stores strings for various purposes. Some of these purposes are pre-defined in the spec; some examples:

  *   a family name such as “Arial”
  *   a subfamily name such as “Condensed Bold Italic”
  *   a foundry name
  *   a copyright string
But the format also allows for other strings for vendor-defined purposes. So, for instance, in variable fonts, the vendor can define instances (particular design variants) for some combination of variation axis values, and then they can define what would effectively be subfamily names for those specific instances.

The actual strings themselves are indexed with a two-part key that includes an ID (generally referred to as the “name ID”), which indicates the purpose (as described above), and a numeric language identifier.

So, using vendor-specific name IDs and language IDs, you could add the kind of strings you describe into a font’s name table.

But there isn’t any existing way to associate particular glyph sequences with a name ID. And that is the part that, in general, doesn’t have a clear need in the way that fonts are used.

Your original idea of localizable sentences, as I recall, involved assigning Unicode code points to particular semantic propositions, or “sentences”. Unicode has stated clearly it is not interested in pursuing that idea and banned further discussion of that idea from its email lists. I don’t think you should be trying to use this list as a back door to revisit the same idea.

Now, what you’ve described seems to have evolved from that original idea—though only slightly: now you’re talking about glyph sequences that represent “sentences”. Based on your slide presentation, it appears you want a message containing “!313125” to get associated with a string “Is there any information about the following person please?” (along with other translations), and you want to use a font table to provide a mapping from the glyphs for the character sequence “!313125” to that string (in its various translation variants). Again, there’s an unstated premise of this idea that the font will get transported with the message. If it did, the font would have some fixed sent of translations. Why not just send a message with multiple translations? Or why not create an online registry that documents the sentences and translations in many languages, and then send as a message a URL that points to the registry entry for “!313125”?

For my part, I don’t see a need to add into the OpenType / OFF spec a table that provides a mapping from glyph sequences to name IDs.

Peter

From: mpeg-otspec <mpeg-otspec-bounces at lists.aau.at> on behalf of William_J_G Overington <wjgo_10009 at btinternet.com>
Date: Tuesday, May 11, 2021 at 4:21 AM
To: 'MPEG OT Spec list' <mpeg-otspec at lists.aau.at>, Peter Constable <pgcon6 at msn.com>, Vladimir Levantovsky <vladimir.levantovsky at gmail.com>
Subject: [EXTERNAL] Re: [MPEG-OTSPEC] New AHG mandates and other news!
Thank you for replying.

I write to seek clarification please.

> You’ve described a way to organize data, but to get the functionality you described the data would be organized differently: a table that maps glyph ID sequences to string entries in the ‘name’ table.

I have found the following web page.

https://docs.microsoft.com/en-us/typography/opentype/spec/name

I am not an expert on OpenType, so as Peter mentions the 'name' table, is the implication that what I am suggesting is already implemented?

If not, can I suggest that for this discussion that we refer to my suggestion as a proposal for a 'text' table please?

I mention that use with QID emoji was just one suggested possibility and that there would be a number of other uses, even if QID emoji is never implemented. The use with QID emoji is not a central application suggestion for this proposed facility.

> It seems to me like you’re trying to propose enhancements the font format to address challenges for the QID emoji proposal.

No. My suggestion has various possible applications, many related to communication through the language barrier. QID emoji were not my idea, I have expressed my views about the idea in my responses to the Unicode Technical Committee's Public Review.

https://www.unicode.org/review/pri408/

My own research is mostly on localizable sentences and their applications, together with some research on The Mariposa System of abstract emoji for assisting communication through the language barrier when using emoji.

Although emoji are interesting, from my perspective they do not have anything like the great potential for communication through the language barrier as does the localizable sentence invention. In particular, many pictorial emoji proposals tend to be deliberately imprecise as regards meaning and implied meaning of an emoji, yet localizable sentences characters are very deliberately precise as to meaning so as to provide high provenance as to meaning in communication through the language barrier.

http://www.users.globalnet.co.uk/~ngo/

In particular, the following slide show was produced for the United Kingdom National Body to forward to the ISO/TC 37 committee.

http://www.users.globalnet.co.uk/~ngo/slide_show_about_localizable_sentences.pdf

For some recent glyph designs for The Mariposa System, please see page 5 of the following thread, starting with the fourth post on that page.

https://forum.affinity.serif.com/index.php?/topic/138654-artwork-for-greetings-cards/

Some readers might perhaps like the designs for some localizable sentence glyphs that are near the start of the thread.

William Overington

Tuesday 11 May 2021

------ Original Message ------
From: "Peter Constable" <pgcon6 at msn.com>
To: "William_J_G Overington" <wjgo_10009 at btinternet.com>; "'MPEG OT Spec list'" <mpeg-otspec at lists.aau.at>; "Vladimir Levantovsky" <vladimir.levantovsky at gmail.com>
Sent: Monday, 2021 May 10 At 22:58
Subject: RE: [MPEG-OTSPEC] New AHG mandates and other news!
William,
You’ve described a way to organize data, but to get the functionality you described the data would be organized differently: a table that maps glyph ID sequences to string entries in the ‘name’ table.
But the scenario you have in mind is to use fonts as a way to carry descriptions of Unicode character sequences, and specifically QID emoji sequences—which is an idea that has been proposed but has not been approved by Unicode. Even _if_ the QID emoji proposal were adopted by Unicode—and it’s far from clear that it will be—, I don’t think it’s a good idea to use fonts as a vehicle for transporting descriptions of glyph ID sequences.

  *   For the QID emoji sequence scenario, Unicode strings in general are sent between applications or between devices 99.99% of the time without any font data. So, it’s very unclear that it would provide much useful benefit for that scenario.
  *   If it is assumed that text containing QID emoji sequences would _need_ font data to be sent along with the text, then that raises a question of whether the QID proposal provides significant benefit over using PUA characters.
  *   The formats added to the font would not be inherently specific to QID sequences—that is, the design suggests a much more general usage: strings describing arbitrary glyph sequences. But I don’t see any real need for such a general mechanism.
It seems to me like you’re trying to propose enhancements the font format to address challenges for the QID emoji proposal. For my part, I don’t think it’s a good idea. Fonts are not the best way to solve those problems. If Unicode is going to consider the QID proposal, then proponents of the proposal need to come up with better ways to address any shortcomings in the proposal.
Peter
From: mpeg-otspec <mpeg-otspec-bounces at lists.aau.at> On Behalf Of William_J_G Overington
Sent: May 6, 2021 8:34 AM
To: 'MPEG OT Spec list' <mpeg-otspec at lists.aau.at>; Vladimir Levantovsky <vladimir.levantovsky at gmail.com>
Subject: Re: [MPEG-OTSPEC] New AHG mandates and other news!
> As part of the mandate #2, we are also encouraged to start exploration activities to discuss the next round of changes that will become the basis for the new OFF 5th edition work item – your contributions to these topics (both on this list and / or new issues on MPEGGroup/OpenFontFormat GitHub<https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FMPEGGroup%2FOpenFontFormat&data=04%7C01%7Cpconstable%40microsoft.com%7Cfe51c13919584abc23dd08d9146ec3d0%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C637563288688874579%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=Slzh2fSOikrSWx%2F0HhjBIhycZjbUfbMoSWvUkFQ22xA%3D&reserved=0>) are much appreciated.

Would it be good to have a new table which is similar in structure to a GSUB table but which can have in the part to the left of each -> either one postscript name or a sequence of postscript names and to the right of each -> a string of Unicode text characters in UTF-16 format - that is, a string of text characters as one might have in, say, a computer program written in Pascal, for the avoidance of doubt specifically not a sequence of postscript names.

I am thinking that this could have various uses, for example, for text to speech in a language of the font designer's choice, transliteration, on-screen explanation of emoji - including perhaps the potentially millions of QID emoji that may soon become encoded into Unicode, so that a font that supports just a few QID emoji could also include an explanation of them in a language of the font designer's choice. The output of the table could be used for any of screen display, tooltip display, speech output. The use of the table in a font would be optional and could be simply ignored by an application that does not support it: also an application that does support the use of the information that is in the table could have a button to switch that use on or off.

William Overington

Thursday 6 May 2021

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.aau.at/pipermail/mpeg-otspec/attachments/20210511/88e4e6e9/attachment-0001.html>