>>Suppose that one wishes to produce a chess diagram in a Unicode compliant >>manner in a document produced using Word 97 running on either a Windows 95 >>platform or a Windows 98 platform, with a view also to save the document as >>plain text. One way to do that would be to use a chess fount which is >>mapped to my collection of code points for a chess fount. I would be >>interested to know of how, if at all, that could be done in a document >>produced using Word 97 running on either a Windows 95 platform or a Windows >>98 platform using regular Unicode or XML. > >For someone unfamiliar with Windows and Word 97, are these proposed >operations (inputting PUA character sequences, saving them as UTF-8 plain >text, opening and correctly reading the document produced) technically >possible on the systems mentioned, assuming you have an appropriate font? >
Well, what I say can be done can indeed be done, yet not quite in the manner suggested in your question. The Private Use Area characters are input by using Insert from the menu bar, then choosing Symbol from the drop down menu. A dialogue box is produced. The symbol can be selected from a display of small boxed representations of all of the characters available in the current fount. There is a cursor which shows a magnified view of the character currently being considered for insertion. It is possible to set the system up for characters to be custom linked to the keyboard. For example, I once had Word 97 on a Windows 95 system set up so that I could enter the twelve accented characters used for Esperanto by using such combinations as Alt+c for c circumflex and Alt+Shift+c for C circumflex. I found that I could key in Esperanto about as fast as I can key English (30 words per minute, which is not fast as keying goes). Saving is to a file type known as "Unicode Text". Getting the file back may display wrongly. One needs to specify the fount to read it properly. For example, Word 97 on a Windows 95 system may be set up to have Times New Roman as the default fount. Suppose that that fount on that machine does not have the fi ligature in the Private Use Area. Suppose that one creates a file using the Tahoma typeface using the fi ligature from the Private Use Area. One then saves the fount as "Unicode Text". Closing down Word 97 then starting it up again and reading in the "Unicode Text" file brings back the information, yet does not display it properly as the fount being used is the default fount with which Word 97 on that PC starts up, namely Times New Roman and that fount on that machine does not have the fi ligature in the Private Use Area. However, displaying using the Tahoma fount displays the fi ligature correctly. As a passing note, please note that Microsoft Corporation very generously provides more modern versions of various founts for free download from its website. The Times New Roman fount on the PC in the above example has just not been replaced with an updated version. The format "Unicode Text" is not defined within the Word 97 help as far as I can find, yet I have tried an experiment of producing a file as a Word 97 Unicode Text format file, then changing its file name extension to .raw and reading it into Paint Shop Pro as a grey scale picture then zooming in on the picture and deducing the bit patterns from the colours. I entered the word actor using an experimental fount of my own with a o r from the ordinary places and a ct ligature at U+E707 as in my golden ligatures collection of code points for ligatures. http://www.users.globalnet.co.uk/~ngo/golden.htm The colours came out as follows. 255 254 97 0 7 231 111 0 114 0 13 0 10 0 After that, all of the grey scale picture which I produced is colour 0, yet I am unsure whether there were none or more 0, 0 pairs after the 10 0 in the file. I do not know the format of that Unicode Text file, but I think that it is not UTF-8. However, one can save in that format and read back in from that format. Also, I have found that one can read in such a file into Microsoft WordPad under Windows 98. One can even, in WordPad under Windows 98, get the ct ligature character into the program using Alt 59143. An interesting document on that technique is as follows. http://www.users.globalnet.co.uk/~ngo/pai04200.htm That page also contains a link to the Microsoft free founts download page. William Overington 21 June 2002
