3 Thursday Plenary

Wednesday Plenary

Current Meeting Report

´╗┐Technical Plenary - 22. March 2007

Agenda:
Welcome and introduction
IAB update
IRTF Report (Aaron Falk + NMRG RG Chair(s)) Technical Discussion:  Internationalization Technical Discussion
	Technical presentations
	Moderated open discussion
IAB open Mic

1. Aaron Falck: IRTF update (see slides)

anti-spam:

Scott: it says on the slides that an RG will publish a BCP?!

Aaron: if they come up with a document, it will certainly go through the IETF last call and the IESG.

2. Aiko Pras, University of Twente (see slides)

3. Technical Discussion:  Internationalisation - Technical Discussion
	
- Leslie gives the high-level introduction (see slides)

- Ted Hardie goes more into detail about the important of contexts
   (see slides)

- John Klensin talks about current work around internationalisation at
   the IETF (see slides)

Open Mic:

Carasco ??: Unicode is only standardizes code-points. It does not specify how to actually draw them. Also fonts are left out.



John Klensin: This is certainly out for our purposes. He expresses sympathy for those people who developed unicode. True what you said, but there are also compatibility characters. For example there is one for font variation, there are several pages for characters that are described as mathematical, but you cannot distinguish them from others except that it is defined in the description as mathematical. So, in that case it would not a font difference, but a mathematical character.  That is a problem, because in our environment it causes complications.

Carasco: There is no definiton of what is a character. Everything to do with characters, is solved to a big extent. But what we would call 'the higher level internationalisation' is not solved. Most of the people in this room will not know: If you go to character modeling in unicode, they will get a headache.

Henning Schulzrinne: Question about concept of IRIs? The deployment community is not clear if you can put non-ascii characters in it and what is happening? Did we do all the right things when it comes to terminology?

Pete Resnick:  you already used the wrong terminology: you used the word character while you should have used non-ascii encoding.

Patrick Faltstrom: Maybe you now understand why we had headaches for the last three years.

Harald Alvestrand: 'devil's advocate question': the IETF does not have a particular expertise in character sets. The Unicode Consortium claims that. Why should we not just wait until they give the right answer in unicode 6.0.

Patrik: We do not want to wait. There is a clear boundary between the Unicode Consortium and the IETF: what we are doing is relying on the data the Unicode Consortium is providing. We are referencing the expertise of the classification of each of the unicode characters.

Dave Crocker: I heard a lengthy description of this complex issue.
Believes everything he was told. Lot of details on decisions, not unreasonable. But there was a lot of detail. And there was no concise statement of the goal of all those details. We did not get a probelm statement.

Ted Hardie: one set of human Internet users cannot use the Internet the way we do it. Would like to change that and therefore trying to make Unicode-UT8 the new ascii

Patrik: We need to have an algorithm for Unicode future versions.
One of John's slildes described why need IDNAbis.

John: a character set is not enough, it comes with a lot of supporting infrastructure and how to use it. It is not all about coding.

Leslie: meta point: there was no clear problem statement, because there is no clear problem, but all sorts of problems that crept up in different environments and that might become bigger problems.

Dave: phishing is one part of the problem. It is about human perception.

Pete: This will not fix the phishing problem. There are too many things that must go into the DNS that will continue to suport the phishing problems.

Dave: guidance is needed on why restrictions are made.

Patrik: The reasons why we do not talk about phishing is that we cannot solve it. But we can try hard to include as few as possible - together with the Unicode Consortium - confusing characters.

Iljitsch van Beijnum: There are ideas to allow people that use different scripts than Latin. Doesn't this lead to balcanisation? What if I get an email written in a different script, I will not be able to read it. Aren't we breaking worldwide operation of the Internet?

Pete: Yes, we are worried about that problem. There is an attempt in the eai WG to come up with a new form of email addresses. Then you will at least be able to reply and get back to the person with the email address written in a different script.

John: You need to keep this in perspective: we are already partially balcanised. The difference is if that person is on the network at all. We are trying to allow more people to participate in the Internet by allowing them to use their own language.

Margaret Wasserman: I remember when Patrik gave a talk on how to compare strings.  Learned a lot, but today got abit lost in the details.  Confused who 'we' is when you talk about who is working on it.

Patrik: it is a mailing list specified on John's slides. Not a WG yet. WE is a group of individuals for now.

Margaret: is this a heads-up on what may be coming into the IETF as new work?

Leslie: yes

Andrew Sullivan: It seems that some of the work is pushed to user space or client space that is currently done with protocols. Do we have to start worrying about the user and applications space. Will different applications not deal with this in different ways which can create new problems again?

JohnK: This problem is not new and we cannot solve it. If I look at an Arabic browser, I cannot read it, unless someone provides a translation into english. The new problem is that someone decides to pass a URL to someone that does not use that script or language. That person better passes real URLs and not just what they happen to be typing (or what they think is the URL).

Pete: This is real user interface work. We might have taken it as granted that the IETF is not doing that.

Andrew: exactly what he is worrying about. The scope for confusion will be bigger if people start communicating among language and script borders. Worried about talking about user interface, and we might not have the expertise here.

Ted: Most cases are localisation issues. Now we focus so much on a multilingual solution, that they will not use it in their local context anymore. We need to make some advances. We are going to have to look at the multilingual context as well, but we have the mechanisms we have now (restriction), we can do that. but the aim is:
to allow the enormous number of people that cannot participate today, because we fail to allow them to localise properly.

someone from Japan (?): if we just extend email, don't forget to look into HAN-unification.

John: no help with that right now, other than what we do here at the IETF provides enough handles that we can do this right at the user interface.

Japanese guy: There is some effort going on in other communities. We need to work together so we all understand each other.

Xiaodong Lee: but on the other hand, also the people in Asia have to understand

Avri Doria: as this new IDNAbis becomes new standard, it will look like we still use the old prefix. Will the old IDNA registrations still work for people?  If we talk about what names should be used, what lable is allowed etc. Where is the boundary what should be used vs. a technical decision what works or doesn't work in the new lable

Patrik: Regarding background compatibility: what happens to names that are registered now and with IDNAbis might not be valid anymore? Yes, this might happen. Most of these cases are corner cases and only those people that wanted to do phishing etc. used them. Working with ICANN on the actual legal issues.

Avri: what about scripts that have not been added yet. The new version is only an extension, only 14 rules have been changed.

John: the vast number of lables that will not be valid in the new system, were not valid under the old system either, the old system was just more permissable. We are already in the territory you are worried about. And we are trying to work to make this less possible and standardise it more.

Patrik: Regarding Avri's second question, we are not explicitely looking code point by code point, because the Unicode Consortium is looking at that. We are looking at the meta data around those code points and that is where we get the rules from right now. Policies about the actual code points are done by the Unicode Consortium. We are looking at the technical implications of allowing certain code points or blocks of code points.




4. Leslie: IAB Update

Leslie handed the IAB chairmanship over to Olaf Kolkman. Olaf thanked Leslie for her outstanding work and handed a present to Leslie. She received a long round of standing ovations from the plenary attendees!

Old and new IAB come up on the stage.

No people at the microphones!


A person at the microphone (?): Do people care more about micromanging the meeting than about the architecture of the Internet? Maybe then we should all go home?

Another person at the microphone: Advise about using SOAP over HTTP for 'everything'.

Brian: this came up in netconf, where peole asked to use web services to build xml? You really have to look at sizes of devices.

Olaf closes the meeting.



Slides

Introduction & IAB Update
IRTF Update
NMRG
Internationalization & Internet Engineering -- Overview
Internationalization in IETF Contexts
Current work in Internationalization around the IETF