Greetings again. Some of you have asked offline for test cases for AMC-Z. The following URL is an 8.5 MB file that expands to 34 MB. It contains all 641883 unique IDN names from a recent snapshot of the VGRS multilingual test bed. The contents of the file looks like: . . . U+D68C U+D654|dz--9x8b1d U+D68C U+D654 U+0069|dz--i-9m3grf U+D68C U+D68C|dz--vz8ba U+D68C U+0031|dz--1-hq3g . . . All entries in the file were prepared first by converting the VGRS zone files from RACE to U+ format, then by converting the U+ into AMC-Z. The latter conversion was done twice, first with Adam's C code from the draft, then using my Perl code which is based on his pseudocode. The two outputs were identical, which is a good (but not perfect) sign. It would be great if other folks who have written their own AMC-Z implementations would test against the file. (There is obviously no need to do so if you are just using Adam's code.) Further, if other folks have collections of example host names, please post them in a similar format so we can all test against them. The file is at <http://www.imc.org/nameprep/mltbd-amcz.gz>. Again, you only want to download this if you can actually test against it (bandwidth isn't free, y'know...). Please send any results to the mailing list. --Paul Hoffman, Director --Internet Mail Consortium
