Re: Re[2]: The meeting with Boshiamy result
On Fri, 21 Jul 2000, hashao wrote:
> »¡¨ì¦rÅé¡A²MµØÁÙ¬O¥_¤jªº¤@Ó¤°»ò¤½¥q(¤è¥¿¡H¡^¡A¦n¹³§â¤@Ó¨t¦Cªº
> °ê¼Ð¦rÅé©ñ¦bºô¤W§K¶O¤U¸ü¡A¤@®M¦@¡@¢·¢·¡@Ó¡@¢â¢â¢Ô¡@¦rÅé°Ú¡C
> ˬO¤S§Æ±æÅý¥Ļâ³oӦۥѤF¡A¤Ï¥¿¬O©ñ¦bºô¤W§K¶O¤U¸ü¡C
> okey, ¬d¤F¤@¤U¡A¥_¤j¤è¥¿ªºÄõ«F¦r®w¡A§Ú¥H«e´£¹L¨S¦³¡H¦L¶H¤¤¦n¹³´£¹L¡C¡@
> or deja vu? http://www1.founderpku.com/fontweb/ ¥þ¬O°ê¼Ð½Xªº¡A
> Ác²¦r«¬³£¦³¡C
> ¤£¹L»¡¹ê¦bªº¡A¦r§Î³oºØ¨Æ¨à¡A´£¨Ñ¤@¨â®M¨t²Î¦r®w´N¦æ¤F¡A¦pªG¥Î¤á
> ¦³»Ýn¡A¦Û¤v¥h¶RÁÙ¬O¦X²zªº¡C¨ì©³³oªF¦è¬O°ø¨×«~¡Aªá¿ú¤]¬OÀ³¸Óªº¡C
> ¤£¹L¬Ý¨ì 88 ®M¦rÅé§K¶O¤U¸ü¡AÁÙ¬O¤j¬y¤f¤ô¡A¥þ¤U¸ü¤F¡C 300 ¦h¥ü¡A
> ¥Î¤£µÛ¤]©ñ¦b¨º¨à¡A¤ß¸ÌµÎªA¡Cn¤£¹ï¤£°_¦Û¤v¡A¤]¹ï¤£°_¤è¥¿¡C
I downloaded those 88 fonts a few months ago. The "jian" ones (e.g.,
fzssjw.ttf) do not have all the characters in GBK, but just the GB2312 (as
far as I can tell) ones--no problem here. However, the "fan" ones (e.g.,
fzssfw.ttf) are very odd--while the glyphs are "fan" (traditional) ones,
they are allocated in the positions for their simplified
equivalents--something is wrong with the CMAP table in the fonts. e.g.,
simplified guo2 'country' is in U+56FD in the "jian" version fzssjw.ttf,
which is okay, but traditional guo2 'country' is in *U+56FD in the "fan"
version fzssfw.ttf, when it should really be at U+570B in that file. It
seems it is intended for someone who is typing a document using a "jian"
font, and thus, simplified characters, to instantly switch to a
"traditional characters" version by switching to a "fan" font, without
proper conversion of the codepoints. (In any case, the "fan" ones only
have slightly more characters than the "jian" ones, but not enough to
cover all the characters in Big5, even if someone were to fix the CMAP's.)
In their current state, I think the "fan" (*fw.ttf) ones should be avoided
because of the latent threat they pose to data integrity.
Thomas Chan
tc31@cornell.edu
Reply to: