I am not a fan of i18n. We should all speak one language (and have one culture, and use symbology the same way). Okay, I'm lying. I like i18n and wish to embrace it in my applications. But there is so much to do!
Not only do I have to worry about using other languages and character sets, but also whether it's an RTL language (which means I'll probably have to make a really different interface for them since they're *usually* right-side anchored instead of our left). And I have to worry about people with names that don't follow our standard First Last format (John (given/first) Smith (family/last) vs Mao (family) Ze (generational) Dong (given) vs Björk (given) Guðmundsdóttir (paternym) (and worse, not even sorted by the "family/last name", but by the given name)).
Colors mean different things everywhere. For example, red is considered "stop" or "warning" in the US, but in China, if a name is in red, it means that the person is deceased. I don't want to accidentally claim that some guy is dead just because I used red to signify something about him. Specifically for my situation, I could understandably use red to signify that John Smith's computer is unavailable or that there is a pending error status for his computer (assuming that his computer happened to be named after him). I don't mean that he's dead, just that there is a problem. What does this mean to me? In the US it is acceptable. In China, it probably isn't. This means that I just can't use red that way. I'll need to use something else to indicate this state.
Symbols in the US mean different things to us than they might mean to someone in India, China, Japan, or even England. For example, using a mailbox with a raised red flag in the real world (away from our computers) means that there is outgoing mail to be picked up by the mailman. But online, it usually means that you have *incoming* mail in your mailbox (yeah, in the US we can't even use this symbol correctly!). And in other countries, they might not even know that it has anything to with mail or messages at all and will not likely understand that the red flag means something.
I mentioned names earlier. I also have to worry about transliterated names. For example, if I were to come across a contact with the name 宮崎駿, how the heck am I supposed to know who that is? If my program is being used in an environment where everyone understands how to read that name, it's okay. But what about mixed or international environments? I want to be able to have that person be addressed using their name in kanji by those that can converse in Japanese (by email/letter/etc), but I would want to be able to *say* their name when I end up talking to them on the phone and I would want to be able to identify the person by their contact record without being completely lost. That means that I need to be able to collect his name transliterated into rōmaji (romanji). Only then would *I* be able to tell that the contact record is actually for MIYAZAKI Hayao, my favorite director.
Did you know that Vaqif Səmədoğlu and Вагиф Сәмәдоғлу are the same person? I didn't either. Storing a transliterated name (Vagif Samadoghlu) would be oh-so-helpful here. But now we have another problem: should we store both the Latin version as well as the Cyrillic version of his name in addition to the transliterated version? It could very well not be worthwhile to worry about it. But how should I know?
And how the heck would you deal with محسن مخملباف if you saw it in a database somewhere? I wouldn't have a clue if that's really a person's name. But since you see it in a contact record, you would probably assume it was a person (or is it a company name?). Since I already know that it's a person's name (I copied it from a webpage that claims that it is the name of an Iranian director), we can move on to another problem: what do I call this person? There is no way I'd get "Mohsen Makhmalbaf" out of that mess. And even then knowing that, should I call the person "Mohsen"? Or "Mr. Mohsen"? Or "Mr. Makhmalbaf"? Or is it really a female? Is "Mr" even appropriate here? Oh, so many possibilities.
I care about these things. Maybe I won't even need to talk to Miyazaki or Makhmalbaf. But my customers probably will. Someday, probably soon, one of the users of my code will need to deal with names like these and they won't want to be left wondering what to do. Sure, we could force the rest of the world into the first-name/last-name business, but why? If we can figure out how to do this right, maybe they'll feel that we really understand their situation or their culture. Maybe they'll feel more comfortable using our software. Maybe they won't feel that all Americans are self-centered, naïve, or stupid (just some of us, eh?).
I believe that opening our software and our eyes to the other cultures of the world will bring us closer together. It might be a lot of work, but it may be worth it in the end.
We use Amazon S3 for data backup for all of our servers. Each individual server gets its own bucket in S3. Actually, it gets two (-backups and -chunks). Apparently there is a limit to the number of buckets you're allowed to have in S3.
So now the last server I added can't be backed up. Greaaat.
I'm in the process of merging a few junk server backups into a single pair of buckets to see how it works. If it works well, I'm just going to merge *all* of the servers into a single pair of buckets. I don't like it, but I'm not sure I have much of a choice. I don't want to stop using S3 for backup (super cheap, easy, and reliable).
