Re: [Slim] Issue 43: How to know the modality of a language indication?

[Bernard Aboba <[e-mail]>]

Thanks for suggesting a way forward, Gunnar. I too would like to get Issue
43 resolved so we can move forward in the process.
Please send your thoughts to the mailing list (preferrably before the
October 30 submission deadline so we can spin a new draft version).

In thinking about the issue, a question Paul asked has stuck in my mind:
What difference does it/should it make?

Let us presume that the user agents are configured to signal the language
preferences.

In a pure peer-to-peer case (me calling you, no intermediaries), I
configure my UA to indicate a preference for Mandarin on the text modality,
French sign language on the audio modality and Swahili on the video
modality.

You (knowing my background and weird sense of humour) after being notified
of my preference, realize I am kidding and agree to accept the call,
knowing that whatever language preference you indicate, we will most likely
communicate in English since I do not speak Mandarin or Swahili and am
barely conversant in French, let along French sign language.

Would this scenario have worked out better with rules that mandated that my
odd choices for audio and video languages be labelled "undefined"? I think
not.

In a scenario where the call is between me and a call center (not a PSAP)
my flippant UA configuration might result in the call being rejected due to
a lack of Mandarin, Swahili or French sign language resources within the
call center.  But it's not clear that labelling my odd choices as
"undefined" should play a role in that decision.  For example, if the call
center did have someone who spoke Swahili, and connected me to them
(perhaps under the theory that my declared preference might indicate an
ability to lip-read Swahili), this might have improved the chance of
communication had my UA configuration been based on genuine expertise
rather than a warped sense of humour.

In other words,it is not clear to me how Section 5.4's discussion of scope
improves or clarifies the situation in any way - and there is some
possibility that it could cause problems.

On Mon, Oct 23, 2017 at 2:17 PM, Gunnar Hellström <
[e-mail]> wrote:

> Issue #43 is the only issue we have left now. I do not want to see the
> discussion stop again until we have a solution on it that seems acceptable.
>
> Section 5.4 seems to be a good place to handle Issue #43.
>
> Currently, section 5.4 is more aimed at limiting what kind of coding for
> languages and modalities are acceptable.
>
> Some viewpoints said that such limitations are not needed and that 5.4 can
> be deleted.
>
> I think we can do something in between these extremes. We can introduce
> explanations for what is required from an acceptable coding of a
> combination of media, languages, directions, and other parameters and
> explain basic ways to assess what is the resulting modality, and also
> explain that the more common media and language combinations that are used,
> the higher chance there is for a match. Thus unusual combinations are
> discouraged but not forbidden as long as the modality can be assessed from
> them. They can be used in specific applications.
>
> I might continue tomorrow with wording proposal for the reasoning above,
> hoping that we can close issue #43 and the discussions around 5.4 soon.
>
> /Gunnar
>
> Den 2017-10-17 kl. 11:02, skrev Gunnar Hellström:
>
> An even more general way to express what section 5.4 tries to say is:
>
> ------------------------------------------------------------
> ------------------------------------------------------------
> ------------------
>
> 5.4 Combinations of Language tags and Media descriptions
>
> The combination of Language tags and other information in the media
> descriptions should be made so that the intended modality can be concluded
> by the negotiating parties.
>
>
> ------------------------------------------------------------
> ----------------------------
>
> That makes us not need to investigate what is possible today, and what
> further attributes or coding rules may be added in the future.
>
> We have a risk that implementers start using some insufficient coding,
> that can cause interop issues. But instead we do not need to limit valid
> use that we just have not thought about by saying that specific
> combinations are out of scope or not defined. It is up to implementers to
> check that the combinations they use result in unambiguous modality.
>
> And it opens for use of possible new attributes e.g.  a=modality:spoken
> or a=modality:written  etc, to complement the undefined case when a
> non-signed language tag without script subtag is used in video media, and
> also for explaining any use of m=application or m=message media in
> interactive communication.
>
> It does not really answer Issue #43 by explaining HOW to assess the
> modality easily, but it requires the implementers to make sure that it is
> possible.
>
> And deducting the intended modality is the key to successful neotiation
> and communication.
>
> Do you think this would be clear enough, or do we need to go into what
> clear cases we have?
>
> Gunnar
>
>
>
> Den 2017-10-17 kl. 00:21, skrev Gunnar Hellström:
>
> Den 2017-10-16 kl. 01:21, skrev Bernard Aboba:
>
> Paul said:
>
> ""- can the UA use this information to change how to render the media?"
>
> [BA]  If the video is used for signing, an application might infer an
> encoder preference for frame rate over resolution (e.g. in WebRTC,
> RTCRtpParameters.degradationPreference = "maintain-framerate" )
>
> <GH>Right, that is a valid example of how real "knowledge" of the modality
> can be used by the application.
>
>
> And, as a response on issue #43,
>
> A simple way is to say
>
> Video media descriptions shall only contain sign language tags
> Audio media descriptions shall only contain language tags for spoken
> language
> Text media descriptions shall only contain language tags for written
> language
> Use of other media descriptions such as message and application with
> language indications require other specifications on how to assess the
> modality for non-signed languages.
>
> The current 5.4 does not mention our main problem with the language tags,
> that there is no difference on them if we mean use for spoken language or
> written language. We should have made better efforts to solve that problem
> long ago, but we have not.
>
> 5.4 can be modified to specify the simple limited case and the problems
> that block us from specifying other cases:
>
> 5.4 <https://tools.ietf.org/html/draft-ietf-slim-negotiating-human-language-17#section-5.4>.  Media and modality Combination problems
>
>
>    The problem of indicating a language tag for the view of a speaking person in a video stream is out of scope for this document.
>
>    The problem of indicating a language tag for use of written language coded as a component in a video stream is out of scope for this document.
>
>    The use of language tags for negotiation of languages in other media than audio, video and text is not defined in this document.
>
>    The problem of knowing which language tags are signed and which are not can be deducted
>    from the IANA language tag registry. How this is done is out of scope of this document.
>
>
> --------------------------------------------------------
>
>
> But if we want to allow more cases, we need to consider the following
> complications:
>
>
> 1. to assess if a language represents a Sign Language, the application can
> look for the word "sign" in the description in the  IANA language registry
> or a copy thereof as Randall already indicated.
>
> 2. For written languages used as a text component in a video stream, it is
> possible to code this for languages requiring a script subtag, but not for
> languages with suppressed script subtags
>
> 3. We have also discussed proposals for how to code written language in
> video stream for languages not requiring a script subtag, but not got
> acceptance for our proposals. So we need to say that that is currently
> undefined.
>
> 4. We also discussed how to code a view of a speaking person in video and
> said that that could be done by using the "definitively not written" script
> subtag on a non-signed language tag in video. But that was not appreciated
> by the language experts. Another option was to not allow written language
> overlayed on video, and that is the lately used option. ( up to version -16
> or so)
>
> 5. For talking and hearing audio media, we only have that case for
> language-tags in Audio. So that is easy to code and assess.
>
> 6. For written language in text media, a check can be made about if "sign"
> is part of the language tag description, and if not, it is a written
> language.
>
> 7. For signed language in text media, a check can be made about if "sign"
> is part of the language tag description, and if it is, it is a signed
> language in text notation. (extremely unusual)
>
> 8. For use with language tags in other media than audio, video and text,
> there is a need for a description on how to assess the modality, especially
> for non-signed languages before it is used.
>
>
> We can construct a section 5.4 to describe this situation, but I doubt
> that it is worth the effort.
>
>
>
> See:  https://rawgit.com/w3c/webrtc-pc/master/webrtc.html#
> dom-rtcrtpparameters-degradationpreference
>
> On Sun, Oct 15, 2017 at 2:22 PM, Gunnar Hellström <
> [e-mail]> wrote:
>
>> Den 2017-10-15 kl. 21:27, skrev Paul Kyzivat:
>>
>>> On 10/15/17 1:49 PM, Bernard Aboba wrote:
>>>
>>>> Paul said:
>>>>
>>>> "For the software to know must mean that it will behave differently for
>>>> a tag that represents a sign language than for one that represents a spoken
>>>> or written language. What is it that it will do differently?"
>>>>
>>>> [BA] In terms of behavior based on the signed/non-signed distinction,
>>>> in -17 the only reference appears to be in Section 5.4, stating that
>>>> certain combinations are not defined in the document (but that definition
>>>> of those combinations was out of scope):
>>>>
>>>
>>> I'm asking whether this is a distinction without a difference. I'm not
>>> asking whether this makes a difference in the *protocol*, but whether in
>>> the end it benefits the participants in the call in any way.
>>>
>> <GH>Good point, I was on my way to make a similar comment earlier today.
>> The difference it makes for applications to "know" what modality a language
>> tag represents in its used position seems to be only for imagined functions
>> that are out of scope for the protocol specification.
>>
>>> For instance:
>>>
>>> - does it help the UA to decide how to alert the callee, so that the
>>>   callee can better decide whether to accept the call or instruct the
>>>   UA about how to handle the call?
>>>
>> <GH>Yes, for a regular human user -to-user call, the result of the
>> negotiation must be presented to the participants, so that they can start
>> the call with a language and modality that is agreed.
>> That presentation could be exactly the description from the language tag
>> registry, and then no "knowledge" is needed from the application. But it is
>> more likely that the application has its own string for presentation of the
>> negotiated language and modality. So that will be presented. But it is
>> still found by a table lookup between language tag and string for a
>> language name, so no real knowledge is needed.
>> We have said many times that the way the application tells the user the
>> result of the negotiation is out of scope for the draft, but it is good to
>> discuss and know that it can be done.
>> A similar mechanism is also needed for configuration of the user's
>> language preference profile further discussed below.
>>
>>>
>>> - does it allow the UA to make a decision whether to accept the media?
>>>
>> <GH>No, the media should be accepted regardless of the result of the
>> language negotiation.
>>
>>>
>>> - can the UA use this information to change how to render the media?
>>>
>> <GH>Yes, for the specialized text notation of sign language we have
>> discussed but currently placed out of scope, a very special rendering
>> application is needed. The modality would be recognized by a script subtag
>> to a sign language tag used in text media. However, I think that would be
>> best to also use it with a specific text subtype, so that the rendering can
>> be controlled by invocation of a "codec" for that rendering.
>>
>>>
>>> And if there is something like this, will the UA be able to do this
>>> generically based on whether the media is sign language or not, or will the
>>> UA need to already understand *specific* sign language tags?
>>>
>> <GH>Applications will need to have localized versions of the names for
>> the different sign languages and also for spoken languages and written
>> languages, to be used in setting of preferences and announcing the results
>> of the negotiation. It might be overkill to have such localized names for
>> all languages in the IANA language registry, so it will need to be able to
>> handle localized names of a subset och the registry. With good design
>> however, this is just an automatic translation between a language tag and a
>> corresponding name, so it does in fact not require any "knowledge" of what
>> modality is used with each language tag.
>> The application can ask for the configuration:
>> "Which languages do you want to offer to send in video"
>> "Which languages do you want to offer to send in text"
>> "Which languages do you want to offer to send in audio"
>> "Which languages do you want to be prepared to receive in video"
>> "Which languages do you want to be prepared to receive in text"
>> "Which languages do you want to be prepared to receive in audio"
>>
>> And for each question provide a list of language names to select from.
>> When the selection is made, the corresponding language tag is placed in the
>> profile for negotiation.
>>
>> If the application provides the whole IANA language registry to the user
>> for each question, then there is a possibility that the user by mistake
>> selects a language that requires another modality than the question was
>> about. If the application shall limit the lists provided for each question,
>> then it will need a kind of knowledge about which language tags suit each
>> modality (and media)
>>
>>
>>
>>> E.g., A UA serving a deaf person might automatically introduce a sign
>>> language interpreter into an incoming audio-only call. If the incoming call
>>> has both audio and video then the video *might* be for conveying sign
>>> language, or not. If not then the UA will still want to bring in a sign
>>> language interpreter. But is knowing the call generically contains sign
>>> language sufficient to decide against bringing in an interpreter? Or must
>>> that depend on it being a sign language that the user can use? If the UA is
>>> configured for all the specific sign languages that the user can deal with
>>> then there is no need to recognize other sign languages generically.
>>>
>> <GH>We are talking about specific language tags here and knowing what
>> modality they are used for. The user needs to specify which sign languages
>> they prefer to use. The callee application can be made to look for gaps
>> between what the caller offers and what the callee can accept, and from
>> that deduct which type and languages for a conversion that is needed, and
>> invoke that as a relay service. That invocation can be made completely
>> table driven and have corresponding translation profiles for available
>> relay services. But it is more likely that it is done by having some
>> knowledge about which languages are sign languages and which are spoken
>> languages and sending the call to the relay service to try to sort out if
>> they can handle the translation.
>>
>>>
>>>
>>> So, the answer is - no, the application does not really have any
>> knowledge about which modality a language tag represents in its used
>> position. If the user selects to indicate very rare language tag
>> indications for a media, then a match will just become very unlikely.
>>
>> Where does this discussion take us? Should we modify section 5.4 again?
>>
>> Thanks
>> Gunnar
>>
>>     Thanks,
>>>     Paul
>>>
>>>       5.4
>>>> <https://tools.ietf.org/html/draft-ietf-slim-negotiating-hum
>>>> an-language-17#section-5.4>.
>>>>       Undefined Combinations
>>>>
>>>>
>>>>
>>>>     The behavior when specifying a non-signed language tag for a video
>>>>     media stream, or a signed language tag for an audio or text media
>>>>     stream, is not defined in this document.
>>>>
>>>>     The problem of knowing which language tags are signed and which are
>>>>     not is out of scope of this document.
>>>>
>>>>
>>>>
>>>> On Sun, Oct 15, 2017 at 10:13 AM, Paul Kyzivat <[e-mail]
>>>> <mailto:[e-mail]>> wrote:
>>>>
>>>>     On 10/15/17 2:24 AM, Gunnar Hellström wrote:
>>>>
>>>>         Paul,
>>>>         Den 2017-10-15 kl. 01:19, skrev Paul Kyzivat:
>>>>
>>>>             On 10/14/17 2:03 PM, Bernard Aboba wrote:
>>>>
>>>>                 Gunnar said:
>>>>
>>>>                 "Applications not implementing such specific notations
>>>>                 may use the following simple deductions.
>>>>
>>>>                 - A language tag in audio media is supposed to indicate
>>>>                 spoken modality.
>>>>
>>>>                 [BA] Even a tag with "Sign Language" in the
>>>> description??
>>>>
>>>>                 - A language tag in text media is supposed to indicate
>>>>                 written modality.
>>>>
>>>>                 [BA] If the tag has "Sign Language" in the description,
>>>>                 can this document really say that?
>>>>
>>>>                 - A language tag in video media is supposed to indicate
>>>>                 visual sign language modality except for the case when
>>>>                 it is supposed to indicate a view of a speaking person
>>>>                 mentioned in section 5.2 characterized by the exact same
>>>>                 language tag also appearing in an audio media
>>>> specification.
>>>>
>>>>                 [BA] It seems like an over-reach to say that a spoken
>>>>                 language tag in video media should instead be
>>>>                 interpreted as a request for Sign Language.  If this
>>>>                 were done, would it always be clear which Sign Language
>>>>                 was intended?  And could we really assume that both
>>>>                 sides, if negotiating a spoken language tag in video
>>>>                 media, were really indicating the desire to sign?  It
>>>>                 seems like this could easily result interoperability
>>>>                 failure.
>>>>
>>>>
>>>>             IMO the right way to indicate that two (or more) media
>>>>             streams are conveying alternative representations of the
>>>>             same language content is by grouping them with a new
>>>>             grouping attribute. That can tie together an audio with a
>>>>             video and/or text. A language tag for sign language on the
>>>>             video stream then clarifies to the recipient that it is sign
>>>>             language. The grouping attribute by itself can indicate that
>>>>             these streams are conveying language.
>>>>
>>>>         <GH>Yes, and that is proposed in
>>>>         draft-hellstrom-slim-modality-grouping    with two kinds of
>>>>         grouping: One kind of grouping to tell that two or more
>>>>         languages in different streams are alternatives with the same
>>>>         content and a priority order is assigned to them to guide the
>>>>         selection of which one to use during the call. The other kind of
>>>>         grouping telling that two or more languages in different streams
>>>>         are desired together with the same language content but
>>>>         different modalities ( such as the use for captioned telephony
>>>>         with the same content provided in both speech and text, or sign
>>>>         language interpretation where you see the interpreter, or
>>>>         possibly spoken language interpretation with the languages
>>>>         provided in different audio streams ). I hope that that draft
>>>>         can be progressed. I see it as a needed complement to the pure
>>>>         language indications per media.
>>>>
>>>>
>>>>     Oh, sorry. I did read that draft but forgot about it.
>>>>
>>>>         The discussion in this thread is more about how an application
>>>>         would easily know that e.g. "ase" is a sign language and "en" is
>>>>         a spoken (or written) language, and also a discussion about what
>>>>         kinds of languages are allowed and indicated by default in each
>>>>         media type. It was not at all about falsely using language tags
>>>>         in the wrong media type as Bernard understood my wording. It was
>>>>         rather a limitation to what modalities are used in each media
>>>>         type and how to know the modality with cases that are not
>>>>         evident, e.g. "application" and "message" media types.
>>>>
>>>>
>>>>     What do you mean by "know"? Is it for the *UA* software to know, or
>>>>     for the human user of the UA to know? Presumably a human user that
>>>>     cares will understand this if presented with the information in some
>>>>     way. But typically this isn't presented to the user.
>>>>
>>>>     For the software to know must mean that it will behave differently
>>>>     for a tag that represents a sign language than for one that
>>>>     represents a spoken or written language. What is it that it will do
>>>>     differently?
>>>>
>>>>              Thanks,
>>>>              Paul
>>>>
>>>>
>>>>         Right now we have returned to a very simple rule: we define only
>>>>         use of spoken language in audio media, written language in text
>>>>         media and sign language in video media.
>>>>         We have discussed other use, such as a view of a speaking person
>>>>         in video, text overlay on video, a sign language notation in
>>>>         text media, written language in message media, written language
>>>>         in WebRTC data channels, sign written and spoken in bucket media
>>>>         maybe declared as application media. We do not define these
>>>>         cases. They are just not defined, not forbidden. They may be
>>>>         defined in the future.
>>>>
>>>>         My proposed wording in section 5.4 got too many
>>>>         misunderstandings so I gave up with it. I think we can live with
>>>>         5.4 as it is in version -16.
>>>>
>>>>         Thanks,
>>>>         Gunnar
>>>>
>>>>
>>>>
>>>>             (IIRC I suggested something along these lines a long time
>>>> ago.)
>>>>
>>>>                  Thanks,
>>>>                  Paul
>>>>
>>>>             _______________________________________________
>>>>             SLIM mailing list
>>>>             [e-mail] <mailto:[e-mail]>
>>>>             https://www.ietf.org/mailman/listinfo/slim
>>>>             <https://www.ietf.org/mailman/listinfo/slim>
>>>>
>>>>
>>>>
>>>>     _______________________________________________
>>>>     SLIM mailing list
>>>>     [e-mail] <mailto:[e-mail]>
>>>>     https://www.ietf.org/mailman/listinfo/slim
>>>>     <https://www.ietf.org/mailman/listinfo/slim>
>>>>
>>>>
>>>>
>>>
>> --
>> -----------------------------------------
>> Gunnar Hellström
>> Omnitor
>> [e-mail]
>> [phone]
>>
>>
>
> --
> -----------------------------------------
> Gunnar Hellström
> [e-mail]
> [phone]
>
>
> --
> -----------------------------------------
> Gunnar Hellström
> [e-mail]
> [phone]
>
>
>
> _______________________________________________
> SLIM mailing [e-mail]://www.ietf.org/mailman/listinfo/slim
>
>
> --
> -----------------------------------------
> Gunnar Hellström
> [e-mail]
> [phone]
>
>