[gtld-tech] Specification 5 - Country names... again..
Alexander Mayrhofer
alexander.mayrhofer at nic.at
Thu Dec 5 16:26:06 UTC 2013
> Alexander Mayrhofer wrote:
> >> Therefore, could you please clarify whether the file that's now on the
> Github was produced by translation engines, or taken from said document? I
> think that's a significant impact on whether or not that file is usable to cover
> ICANN's requirement.
>
> Thanks for your comments @Alex. The file submitted by Verisign to satisfy
> Specification 5.4.3 was created using automatic translation of the English
> Language names of the 193 UN Member states.
[Alexander Mayrhofer]
Thanks for the clarification. As i said above, it would be great to hear from ICANN what they intended with the 5.4.3 requirement - more specifically, which components of the date in the document cited in the requirement is to be used.
My point is - the requirements are vague, and hence there is no clear way to quantify the quality of a solution. Which, in turn, is bound to create interopability issues.
> @Alex, if you have a less-error prone method of arriving at a solution,
> Verisign would invest some time in reworking the data. But cutting and
> pasting from the PDF seems error prone. We would need a programmatic
> solution to consuming those records. And records meant to be consumed
> programmatically are not generally encoded in PDF. But again, we're open to
> ideas.
[Alexander Mayrhofer]
I totally agree that cut & past from the PDF is extremely error prone, even if there would be someone understanding all the scripts used in the document.
Alex
More information about the gtld-tech
mailing list