Hi, Apologies for not attending the call (if there is an archived mp3 recording, I’d appreciate a link), so forgive me if my comments are a bit out of context. Whether or not this group should mandate automated methods for language detection is not something I have a ready thought on, but I imagine that whether or not the contact data language detection is automated, it needs to be accurate in its detection. Is that something that is possible? There were studies done a couple of years ago on accuracy of automated T/T, and those did not prove fruitful. Too many identified languages scripts with large error margins (more on this can be found on the PDP WG wiki page here: https://community.icann.org/display/tatcipdp/7+Studies+and+Background+Docume...). Language/script detection may be a different matter, but are there any indications out there that this can indeed be done reliably? That might be something worth considering before contemplating mandatory automated detection as an implementation measure. Also on this from Roger:
On Jan 5, 2017, at 10:17 PM, Roger D Carney <rcarney@godaddy.com> wrote:
[snip]
In regards to the last point, and as I have stated before in different ways, my expectation coming out of the T/T PDP was that transformation (in its entirety) was not mandatory. And to me this meant none of the recommendations were mandatory but that if you did translate and/or transliterate then these recommendations would apply. My expectation for the T/T implementation group was to come up with an implementation for those that choose to translate and/or transliterate.
Yes…, this sounds right to me. If a “gTLD Provider”, which I have always understood to be a registry and/or registrar chooses to provide registration services in local languages and scripts with optional transformation, the rest of the consensus policy kicks in. So if a contracted party chooses to transform, it will be required to fill in language tags. It is however my understanding the the whois replacement system needs to be able to provide language tags as a feature, so that in the event that the contracted party chooses to transform data, the ability to tag certain fields is in place. If I recall correctly, the PDP WG also recommended that it be indicated who actually transformed the authoritative data in its local language/script. You may also find an exchange that took place between the ICANN Board and GNSO Council on some of this to be insightful. Letter from the ICANN Board to the GNSO Council: https://www.icann.org/en/system/files/correspondence/crocker-to-bladel-11may... GNSO Council response: https://gnso.icann.org/en/correspondence/gnso-council-to-icann-board-21dec16... Thanks. Amr