90 likes | 228 Views
IDN issues. ITU-T SG16-Q7 26 July 2010. Objective. Provide a set of examples why IDN study at ITU-T could be considered Demonstrate the need of ITU recommendation for at least one issue.
E N D
IDN issues ITU-T SG16-Q7 26 July 2010
Objective • Provide a set of examples why IDN study at ITU-T could be considered • Demonstrate the need of ITU recommendation for at least one issue. • Trigger the discussion between local expert communities about other issues which still need to be addressed.
Current status • Standards issued by IETF : IDNA 2008 in their final phase, major revision of IDNA2003. • Mainly focus on character conversion and registration protocol, no focus on applications. • ICANN launched IDN ccTLD “fast track” on 2009. • The “full fledged” IDN gTLD process will be launched at start of 2011. • We believe there are still certain “holes” in the current standards which need to be covered. Mainly at the application level.
Variants and visual confusion • Old problem : Paypal visual confusion attack (The “P” is cyrillic and not Latin). • IDNA sol. : No mixing of characters between “scripts”. Notice: scripts and not “languages”. • Prob.: Scripts could be shared by many language communities. Example: Arabic Script (Arabic, farsi, Urdu, jawi, etc). • IDNA2008 relies on UNICODE, which also defines scripts and not languages. • Result: Visual confusion still possible. Even worse: visually confusing characters within the same language.
Variants and visual confusion • Problem: What would happen to a registry if he opens door for IDN registration w/o considering visual confusion issues ? • Problem currently existing with Arabic scrip, but not sharp (yet) because there are no IDN TLDs, but this will change soon (1 year, 2 years ?). • Proposed solution: Apply a master key algorithm before domain allocation to generate all variants and block them. • http://arabic-domains.org/adn_tools/mk/index.php • Algorithm needs further development: acceleration, check with other languages/script communities if the approach is applicable or not.
Variants and visual confusion • Action needed: Issue an ITU recommendation about variants in domain names and how to deal with them. In case a master key algorithm is found to be appropriate, a set of rules to designate the key should be included within the standard. • Who needs this recommendation ? • Registry operators • Developpers of registration software for registries and registrars • Developpers of applications using IDN domains • A liaison could be established with IETF to avoid redundancy.
URL structure • A web address is composed of three parts: • Scheme://Domain Name/Path • IDN adresses only the domain name part. • W3C/IETF standards would allow Path to be written in Unicode, but not Scheme. • Even worse: if we do not write scheme, then it defaults to http. • Result: We got the gTLDs in IDN, but it is difficult to use. • Action: Issue a recommendation about Full Unicode web addresses. A liaison could be established with W3C
Conclusion • Certain IDN issues need to be adressed, mainly for application support. Not adressed by current standards because out of scope. • Some of them are clearly identified (such as visual confusion), some needs more study (such as URL structure, email structure, etc.). • The outcome will definitly be needed by the community once large scale IDN registration starts (within 1-2 years). • Propose to start working on clearly identified issues, while others are being prepared within local expert communities.