Sorry if this message got sent already. (Outlook Express promised me it was just saving a draft so I could edit it later.)
William Overington <WOverington at ngo dot globalnet dot co dot uk> wrote: > Yet ConScript has now withdrawn that allocation and now uses that > code point for Ewellic. > > http://www.evertype.com/standards/csur/ewellic.html > > What is interesting is as to how Doug produced that effect. How was > it done please? I don't think the "effect" William was talking about was getting my proposal for Ewellic listed in ConScript, but rather entering the character U+E707 into an e-mail. It was simple. I started SC UniPad <http://www.unipad.org> with a new document, opened Character Map, entered E707 in the edit control and pressed Enter, then copied and pasted the character into Outlook Express. I could have entered the character into UniPad in other ways, too, like typing "\ue707", then highlighting the sequence and converting it from ASCII-UCN to Unicode. As for the problem of sending a PUA character in an e-mail, I simply made sure the encoding of my message was set to UTF-8. (More about UTF-8 below.) > Here it came out as a black rectangle in Outlook Express. That's right, it did. And it did for almost everyone else (except my neighbor, James Kass). That was my whole point of using it; unless you were using the One and Only Font to read my message, you would see a black rectangle, which is WORSE -- not better -- than if I had just used "c" and "t". > So I did > two things. Firstly I looked in the message source and found the > string =EE=9C=87 in the line of text. Secondly I did a copy and > paste of the text from Outlook Express to Word 97 and then did a > Save as HTML and then I looked at the source code of the HTML file > which was produced. This produced the number 59143 in the sequence >  so I then looked in the list at the following web page. > > http://www.Joern.De/tipsn128.htm#Ligaturen > > There, to my delight, was the number 59143 alongside my choice of > U+E707 for the ct ligature. Well, you probably could have guessed on your own what belongs between "Respe" and "fully". :-) > This is interesting, as the fact that your system was set up for > ConScript and Doug wrote using a character from what is now called > the golden ligatures collection provides a good practical example of > the need for the use of the classification codes which I suggested > some time ago. No, it doesn't. It illustrates the supreme lack of interoperability which would result from the quasi-standard use of these ligature code points. You saw a black box and so did most other users. And even you, the inventor of these things, had to jump through hoops to discover what the text meant. > If the Conscript registry is defined to be in one type tray and the > golden ligatures collection is defined to be in another type tray, > then, in future software, the two different meanings associated with > the code point U+E707 could be clearly signalled, indeed the two > meanings could both be signalled in the same document! > > I am wondering what is the coding that Doug used, namely =EE=9C=87 > in the line of text. It's called quoted-printable UTF-8. U+E707 is expressed in UTF-8 with the bytes 0xEE 0x9C 0x87. These three non-ASCII bytes are then converted to the nine-character ASCII string "=EE=9C=87" so they will pass through e-mail channels. This is basic character encoding stuff, something you really should have a handle on if you are going to propose grand new uses for Unicode. > I have also analysed the other black rectangle which appears in your > posting by the same process. It comes out as decimal 9785 which > converts to hexadecimal 2639 which, upon looking in the code charts, > gives a variation on a smiley, namely a frowning face. Now that glyph *is* in quite a few existing fonts. If you are using Windows 2000 or XP and a fairly common Microsoft-provided font, you should have seen it. > So, Doug has proved the benefit of my list existing and you have > proved the benefit of, in the future, using my suggested > classification codes. Nobody did any such thing. We proved that the use of this proposed ligature character obscures the intended text unless a custom-built font is used, and guarantees that a search of the Unicode Mailing List archive for the word "Respectfully" will never return a hit for that message. -Doug Ewell Fullerton, California
