Odd capitalization question

[KeithD]

That doesn't mean that everyone else is the most accurate. The government is (or was at one time) very careful about hiring the best and most consistent translators. Just because it's a real popular game to "it's always someone else's fault and nobody knows as much as I do--about everything" dump on government doesn't mean it's justified to do so.

 

[Bramblethorn]

I would imagine that you're right about addressing documents to individuals. I only had involvement in how the government indexed them in their research databases (and the indexes didn't include personally preferred spelling variations. Seems they probably should have)--and that was well before Google was being used by anyone.

There were some older tools for that kind of problem - not as reliable as the approach you describe, but still quite useful for situations where getting a human to enforce consistent spellings isn't an option.

I understand law-enforcement agencies still make use of Soundex, which was invented way back in 1918, and I used it myself for name-matching a few years ago. It's surprisingly effective for how simple it is. For example, "Mohammed", "Mehmet", "Muhamad" and almost all other variants get coded as M530, representing an M-m-d sound.

 

[KeithD]

The Intelligence agencies haven't been much on off-the-shelf systems. They want full control from bottom up. My office did take the message-handling system of the deceased Washington Evening Star once, but only as a temporary fix.

 

[Touch_type]

How did you handle Western names with multiple spellings, like Graeme/Grahame/Graham, Ian/Iain, Brian/Bryan/Bryen/Brien/Briain, Chris/Kris/Krys, Robin/Robyn, Sean/Shane, and so on? If you were standardising "Muhammad" but not those, I can see that giving offence.

Not really. Anglophone writers spelled their own names; interviewers always checked the spelling of US, European, Turkish (basically people who used the Latin alphabet) interviewees. We used to contact the Japanese and Chinese embassies if we needed to check spellings of their nationals' names. There seems to be an internationally accepted system for transliteration of names from the Indian sub-continent. Discussions about names mainly came from students. They often wrote essays in English and had an opinion on how they transliterated their names, although, when I taught occasional classes, I noticed that individuals would write their names differently on different days.

Ironically, the most confusing transliteration was the other way around. 'John' and 'Joan' are transliterated the same way from English to Arabic (and Hebrew). People wanting to contact my colleagues with those names found it very hard to get their heads round why a man would have what they knew to be a woman's name and vice versa.

 

[ArdentHemingway]

Perhaps you should ask your Juliet.

 

[Bramblethorn]

Ironically, the most confusing transliteration was the other way around. 'John' and 'Joan' are transliterated the same way from English to Arabic (and Hebrew). People wanting to contact my colleagues with those names found it very hard to get their heads round why a man would have what they knew to be a woman's name and vice versa.

That surprises me - I'm no linguist, but I know both Hebrew and Arabic have quite a few unisex names e.g. "Gal", "Qamar".

 

[Touch_type]

That surprises me - I'm no linguist, but I know both Hebrew and Arabic have quite a few unisex names e.g. "Gal", "Qamar".

Don't think so. There are similar sounding names, but they'd have gender indicators - usually suffixes. A bit like 'George" and 'Georgette' or "Georgina' in English; Sayeed and Sayeeda is an example which springs to mind. At least in my day, Hebrew names tended to be of Biblical origin. And that version of God tends to be pretty strict on gender divisions. Names like Uri and Ora could have the same roots but be gender specific. The absence of vowels in the written versions of both languages may also explain some similarities.