I think I am getting old

by Michael S. Kaplan, published on 2005/07/05 02:30 -04:00, original URI: http://blogs.msdn.com/b/michkap/archive/2005/07/04/435522.aspx

I had dinner with Andrea last night (yes, the same Andrea who I was talking with about the Korean Unicode sort) in the before time, in the long long ago.

Just dinner folks. She has a boyfriend now, and he is not me. Man, if my life were only as interesting as people who email me seem to think it is! :-)

Anyway, she is not attending TechEd, but she figured it would be nice to have dinner and catch up. She did smile when she saw that old post, and was amazed at how much detail I remembered. She even asked if I had recorded the conversation (no, I didn't!) or if she was that interesting (ordinarily I'd probably say yes to that but she was fishing so I just said maybe and smiled), or am I just that much of a geek (yes, I definitely am!).

But after dinner we parted ways and I went back to my hotel room. I shot up with the appropriate sub/q meds I am supposed to, and took all of the pills I am supposed to take (now I have them in one of those caddies with separate little bins for AM and PM of each day, easier to be sure I am taking all I am supposed to). And as I looked at the pill container, I realized that I am getting old.

It's not just the M.S. thing, though of course that makes it a lot more apparent, a lot more often.

I was having a conversation with someone at Maya's wedding, a friend of the family who practically wanted to see ID when I told her I was almost 35 (though when she looked closer and saw the grey she grudingly admitted that she could maybe see it). She did go on for a while about how well I seemed to be dealing with it all, though of course her only sampling was seeing me in a Tux at a wedding reception, so it is probably hard to judge.

I didn't dance with anyone at the reception. But I did watch a lot of people dance. And got to talk to a bunch of people, some of whom I have not seen for years.

That reminds me, I will get up a tux picture as soon as I can, the adapter to take them off the camera is at home so it will not be right away unless someone forwards a picture to me sooner!

Of course I do seem to regress a little at events like TechEd, but in the end not all that much. It is astounding how often 'mature wisdom' resembles being just too damn tired. :-)

But I went back to my room and worked for a while, some cool engine/tool type stuff. Can't say what it is yet, but one day.

I am getting old, though. I guess we all are. The problem with M.S. is that it makes denial a lot harder to do convincingly....

Continuing the discussion you and Norman had on that old Korean Unicode thread: the point he missed and I don't think you made successfully is that the Japanese and Korean standard fonts are broken - they actually map U+005C to a glyph of the Yen or Won character as appropriate. You can see this in Batang, BatangChe, Dotum, DotumChe, Gulim, GulimChe, Gungsuh, GungsuhChe, MS Gothic, MS Mincho, MS PGothic, MS PMincho, and MS UI Gothic, to name the fonts that have this issue on my system (XP SP2). From a purist's point of view, yes, U+005C should always be mapped to a glyph that is recognisable as a REVERSE SOLIDUS. But that's up to the font author.

There must be reasons why U+005C has ended up mapped incorrectly in those fonts - and it looks like at least one reason is that the first 128 code points are identical to ISO-646-US (US-ASCII) in code page 932 and 949, *except* for 0x5C (== YEN SIGN, WON SIGN respectively). The WON SIGN U+2089, at least, isn't even in the Windows Glyph List 4.0 set, although (presumably through being in ISO-8859-1 and hence Windows-1252 at 0xA5) the YEN SIGN U+00A5 is.

I was trying to investigate this further but I'll be honest: I got very lost in the OpenType documentation. The character design standards on the Typography website only cover Latin and Cyrillic glyphs. Perhaps this issue originated in TrueType before version 1.66, which is the earliest (and also latest) version that is currently available.

I have trouble considering it a bug, since it actually is what customers in Japan and Korea have wanted for their backslash for as long as they have had a version of DOS to see it on. The font is matching the user expectation.... any other behavior would break code that relied on the path separator's value across different language settings....