Difference between revisions of "Talk:Vision2 and Unicode"
(added howto) |
Colin-adams (Talk | contribs) |
||
| Line 3: | Line 3: | ||
*Do the Unicode stuff first as an outside library that takes STRING_GENERAL instances as argument | *Do the Unicode stuff first as an outside library that takes STRING_GENERAL instances as argument | ||
**The implementation should be as compact as possible for the tables, and of course very efficient. This is actually a quite difficult work since it requires some reading of the Unicode standard. | **The implementation should be as compact as possible for the tables, and of course very efficient. This is actually a quite difficult work since it requires some reading of the Unicode standard. | ||
| + | ***I would say it requires reading the text of the entire book once, plus the standard annexes, and then re-reading (maybe several times) the sections relevant to the particular task you are engaged on. At least, I found this approach necessary when I was implementing geuc, and UTF-16 and UTF-32 serialization, in order to get a good grasp on all the issues involved. (Colin Adams) | ||
*Implement helper classes using RAW_FILE as argument to read text file in a special unicode encoding | *Implement helper classes using RAW_FILE as argument to read text file in a special unicode encoding | ||
**Looks like there are some kind of standard out there using BOM | **Looks like there are some kind of standard out there using BOM | ||
*update editor to accepts STRING_GENERAL instances as arguments and thus enabling user to enter unicode characters | *update editor to accepts STRING_GENERAL instances as arguments and thus enabling user to enter unicode characters | ||
Revision as of 22:09, 28 April 2006
Howto
The basic ideas are the following:
- Do the Unicode stuff first as an outside library that takes STRING_GENERAL instances as argument
- The implementation should be as compact as possible for the tables, and of course very efficient. This is actually a quite difficult work since it requires some reading of the Unicode standard.
- I would say it requires reading the text of the entire book once, plus the standard annexes, and then re-reading (maybe several times) the sections relevant to the particular task you are engaged on. At least, I found this approach necessary when I was implementing geuc, and UTF-16 and UTF-32 serialization, in order to get a good grasp on all the issues involved. (Colin Adams)
- The implementation should be as compact as possible for the tables, and of course very efficient. This is actually a quite difficult work since it requires some reading of the Unicode standard.
- Implement helper classes using RAW_FILE as argument to read text file in a special unicode encoding
- Looks like there are some kind of standard out there using BOM
- update editor to accepts STRING_GENERAL instances as arguments and thus enabling user to enter unicode characters
