by Michael S. Kaplan, published on 2008/08/02 14:01 -04:00, original URI: http://blogs.msdn.com/b/michkap/archive/2008/08/02/8806350.aspx
We were at a meeting yesterday.
A 4:00 meeting.
Yes, a 4pm on a Friday meeting.
When the sun was out.
I probably would have left work early that day and tried to enjoy the sun, were this meeting not on the schedule.
The organizer of the meeting, self-consciously aware of this, said as much in the invite:
Sorry for the late time on Friday but we have varying OOFs coming up.
Obviously it wasn't really late. Hell, after I had dinner I probably worked for several hours. but folks in the Pacific Northwest as a tribe seem to be quite jealous of people who would deny us our sun time.
Anyway, it was an hour meeting.
Well, technically it was two 30-minute meetings, back-to-back.
I was only invited to the first one, but I suspected that unless I insisted they'd be happy if I stayed the full hour.
Turns out I was right. :-)
It was a few minutes after five when I finally aimed the scooter at the door.
Everyone but Peter and I had already left.
Essentially everyone but Peter and I were smart enough to leave....
Peter asked me a question when the scooter was literally sticking out the door like something unmentionable sticks out that you have to tuck back in.
Obnoxiously enough, it was interesting question.
Peter was confused about something.
He was wondering about all of these books like the Kano book1, they start taking about internationalization from the NLS side. It seemed to him like such an unnatural place to begin if you are trying to introduce the subject to people who had no background in the area. Wouldn't it make more sense to start from somewhere like text? Everyone has to deal with input or display, after all.
On the one hand, he has a point. It does seem like an odd place to start.
On the other hand, however, he is actually a font person and linguist, asking an NLS person with notions of linguistic aptitude why books like the quintessential tome for the area starts from what the latter is most familiar with when what the former is most familiar with seems like so much more natural of a place.
When it doesn't, for the former. :-)
I don't recall exactly what my answer was -- it focused on the viewpoint thing, and the fact that there were more applications that had to start from a simple place where their UI was a series of printf statements than people writing code that built huge word processing applications. So that it was attempting to start simple from what was perhaps the commonest (f not the leastest) of the least common denominators.
I do know that I did not think of that line at the time; what I said was less poetic. But I was honestly thinking about the last bit of sunlight for the day and not focusing on remembering the exact words this time. Plus this seems more clever, with the proper amount of self-deprecation for me to feel comfortable with a situation where i was explaining why in a world of competing viewpoints, the one closer to mine happened to be the winner this time....
The truth is perhaps a bit more complicated, though.
After all, the same content could have been organized very differently and it would still be great content. Since it is the kind of book where almost every time it is opened it is to a specific place found from the TOC or the index or memory, and since few will read it from end-to-end even if you leave out the appendices, how it is organized is not, strictly speaking, as important as my answer implied.
HOWEVER, many of the people doing the writing here came from that development background.
Not the typographic one.
And if someone is putting together a white paper or a book, one will only most comfortably write from the point of view of one's own experience; it's only natural.
How would you have done it, if it were up to you?
Would you have started with text display (the typography side) or locales (the NLS side)?
And is it for any reason beyond your background that you can identify? :-)
1 - Developing International Software for Windows 95 and Windows NT (first edition). Nadine Kano's name was on the cover but everyone (including her) knew that it was a collective effort from many people throughout Microsoft. We still call it the Kano book most of the time, or occasionally DISv.1 if we are contrasting....
This blog brought to you by ਁ (U+0a01, aka GURMUKHI SIGN ADAK BINDI)
Mihai on 4 Aug 2008 3:40 PM:
I would start with localizability, then locales.
And I think there are several reasons for that (I think):
1. Display and input are taken for granted. The OS deals with it (correctly).
Use any Unicode API and standard controls, and you get the right result.
It is only when you start messing with it (custom controls, etc.) that you risk bad results, but then you are asking for it :-)
2. You cannot carry the bad practices from school to Windows text output. You can't call printf to display text in a window. So you will use the right API, like it or not.
On the other side, it is very easy to call strcmp, strftime, and all the other crap from standard C or C++.
3. In many people's minds internationalization == localizability. If I extract the text in resources, the application is internationalized.
Then you do the first Japanese build, the first complain is about corrupted text. Or the impossibility to type. So that's the first thing you have to solve. No need for a book.
But you don't often get bugs reported about date/time/number formats and such.
They are also easier (unfortunately) to postpone for the next version, or discard them as unimportant.