Stupidity of the day
I’m leaving for Texas, USA soon. Me being a German living in - guess - Germany, causes the need to apply for the US “Visa Waiver Program” using https://esta.cbp.dhs.gov (thanks again to Otto for reminding me *g*) these days. While clicking through and filling out the forms of the electronic variant of that “green sheet of paper” (the one you used to have to fill out on the plane prior to landing on US soil) I was greeted by this notice…
It’s almost 2010! Have the people, who implemented that web-interface, ever heard of Unicode or do they expect international travelers to not use anything but ASCII to supply their (usually non-english) names, which carry the high probability with them to not use ASCII characters only? For me it’s just the ü in my surname. I wonder what people with funkier names do, when they have to diverge from the correct name-spelling to something this ESTA-system accepts. Once they succeed there, I bet they have a hard time trying to convince the staff at customs, that they are really themselves, because the spelling of their name on the passport doesn’t even remotely match the spelling in the visa-waiver-form.
I once almost wasn’t let aboard a plane in Germany, because the travel-agency booked my flight on “Mueller”, but my passport says “Müller”. Is all that the legacy-fault of Cobol?
November 5th, 2009 at 8:05 pm
Unfortunately I encounter this somewhat frequently even in South Africa where the local population often have names with diacritics. I wrote about my experience with a local airline here: http://translate.org.za/blogs/friedel/content/vrot-mango
November 5th, 2009 at 8:28 pm
Yeah, same here, even in my own country there’s still a lot of sites (even government ones) that don’t accept anything but ASCII.
Still, I think the trend is to be more and more compatible with unicode and everything.
By the way, for the “Mueller” thing, it looks like the transliteration was done quite right:
$ echo “Müller” | LC_CTYPE=de_DE.UTF-8 iconv -f utf-8 -t ASCII//translit
Mueller
(Still, unicode is hundreds of times better than transliteration!)
November 5th, 2009 at 8:33 pm
@ F Wolff: Holy… your encounters in this domain are even worse than mine.
November 5th, 2009 at 8:59 pm
You’ll still get to fill in that green sheet of paper. With the same set of questions as esta.
November 5th, 2009 at 9:01 pm
Unfortunately, you will still have to fill out the green form on the plane as well.
At least they still made us do it back in September when I visited the states for vacation
November 5th, 2009 at 9:03 pm
I have experienced the same when I registered on that site. My credit card does not feature the beloved ‘ö’ either, my first name is transliterated as Goetz. My collection of conference badges have many funny symbols in that place as well.
November 5th, 2009 at 11:19 pm
This is great. You are visiting me this time! Texas looks forward to you.
November 6th, 2009 at 12:11 am
Anybody with a passport issued by a reasonably modern country in the last twentysomething years shouldn’t have any problem figuring out the required transcription, if any, of his name. See http://en.wikipedia.org/wiki/Machine-readable_passport . “The data of the machine readable zone consists of two rows of 44 characters each. The only characters used are A-Z, 0-9 and the filler character <.” So just look in your passport and you will see what to input in the web form.
November 6th, 2009 at 12:31 am
Btw, I found this interesting document: http://www.icao.int/icao/en/atb/meetings/2008/TAGMRTD18/TagMrtd18_wp02.pdf about the issues related to transcribing Arabic names into the A-Z of the machine-readable lines in passports. (Note that the document for some reason uses “Arabic font” when it means “Arabic script”.)
It’s not like Unicode would be an unknown concept to people who actually work with these things. The document above even discusses the possibility of using Unicode numeric code values on the machine-readable lines, but rejects the idea because it would take too much space. (And be utterly non-readable to humans, of course.) Just having arbitrary Unicode on the machine-readable lines is not feasible either. It’s surely much harder to do OCR of arbitrary Unicode text than of just A-Z. (Allowing just some accented Latin letters but not for instance Greek or Hebrew would be silly.) And note that the font on the machine-readable lines of passports is one specifically designed to be easily OCRable, OCR-B.
November 6th, 2009 at 5:33 am
It’d be nice if the ETSA form actually saved some time, but when I travelled through the US earlier in the year I still had to fill in the old green form. I’m not even sure if everyone on the flight had filled out the ETSA form …
As for representing non-ASCII names on passports, I doubt you’ll see them change the machine readable zone. If they are going to store a unicode name anywhere these days, it will be on the chip. No need for clever OCR tricks if the name is already available in binary form.