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

[William_J_G Overington]

To "'MPEG OT Spec list'" <mpeg-otspec at lists.aau.at>

Hello Peter
Thank you for taking the time to explain. I appreciate your help.

> But it isn’t really practical from a linguistic perspective. It 
> assumes that semantic propositions can be defined that can always be 
> used as translation pivots between two arbitrary languages, and that 
> those can be formulated independent  of any larger discourse context.

I was unfamiliar with the term "translation pivots" as such.

https://en.wikipedia.org/wiki/Pivot_language

However, upon investigation I found that I a> m familiar with the 
concept, just not under that name.

> It’s also not scoped: Adapting examples from your slides, I might want 
> to send a message to the hotel saying, “If my dog has hair rather than 
> fur, and doesn’t shed, will I be able to keep my dog in your hotel 
> rooms?” There  are an unbounded number of semantic propositions 
> someone might want to communicate in your hotel scenario. There’s no 
> clear way to scope.

Yes, exactly.

I have long recognised that. I have always been suggesting that 
localizable sentences can be used for communication through the language 
barrier in some particular circumstances. I have not advocated the 
invention for every possible use.

The slide presentation is about finding information through the language 
barrier after a disaster. I opine that the invention would be good for 
that application.

This limitation of not encoding every possible sentence is applied upon 
in Chapter 10 of my first novel.

http://www.users.globalnet.co.uk/~ngo/localizable_sentences_the_novel_chapter_010.pdf

However, emoji cannot depict every possible image that can exist, yet 
emoji are being encoded for some pictures.

Also, clip art cannot depict every possible image that can exist, yet 
clip art collections can be useful.

I have used clip art. For example in the slide presentation that we have 
been discussing.

  >> I am just told that it is 'out of scope'…

> That’s because Unicode is an encoding for _characters / text 
> elements_, not semantic propositions.

Thank you for explaining.

Yet what about U+1F91F?

https://www.unicode.org/charts/PDF/Unicode-13.0/U130-1F900.pdf

There are glyphs for the localizable sentences. There seems to me to be 
no reason why sequences to use those glyphs interoperably with those 
sequences including tag characters could not be reasonably formally 
accepted into regular Unicode, thereby providing robustness and formal 
encoding in a non-proprietary manner such that they could be used 
without problem or uncertainty as to provenance or intellectual property 
rights by any individual or organization that chooses to use them.

It would seem that adding a note in the code chart explaining the 
meaning and its context would not be unreasonable.

Now maybe I am wrong about that, and there is some reason of which I am 
not aware.

However, I opine that the way to settle the matter is to have that 
formally and fairly assessed by the Unicode Technical Committee, with 
reasons given if the idea is refused.

That would seem to be the fair way and the professional way to do 
things.

For example, the sentence "It is spring." would have the context of "A 
season of the year".

I remember at one stage, though not now, an online translation facility 
translated into French "It is spring." to "C'est source." as in a spring 
of water from a hillside rather than "C'est le printemps." as desired. 
So the context is important as a guide to the human translator producing 
a sentence.dat file.

> If the idea is that the display strings for a the set of localizable 
> sentences is resident in the receiving device, then that’s an instance 
> of something that is frequently done:  it’s very common for software 
> running on devices to have access to lots of static strings to display 
> messages to a user. However, fonts are not the storage mechanism used 
> for such strings, and software vendors definitely would not want to 
> start using fonts  as a new way to store static strings for 
> user-interface messages.

So maybe the way to do localizable sentences is to use the following 
format that I devised. The format is such that it uses a text file that 
can be produced using the Microsoft WordPad program. WordPad is bundled 
with the Windows operating system so is widely available without needing 
to use additional software to produce or study the file.

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

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

William Overington

Thursday 13 May 2021


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.aau.at/pipermail/mpeg-otspec/attachments/20210513/0b8c1de5/attachment.html>