Relation of Character Properties to Algorithms

Belloc · **Joined:** Sat Aug 06, 2011 9:02 am **Posts:** 43

In 2.3 Relation of Character Properties to Algorithms one can read the following paragraph :

Quote:

When assigning character properties for use with a given algorithm, it may be tempting to assign somewhat arbitrary values to some characters, as long as the algorithm happens to produce the expected results.Proceeding in this way hides the nature of the character and limits the re-use of character properties by related processes. Therefore, instead of tweaking the properties to simply make a particular algorithm easier, the Unicode Standard pays careful attention to the essential underlying linguistic identity of the character. However, not all aspects of a character’s identity are relevant in all circumstances, and some characters can be used in many different ways, depending on context or circumstance. This means the formal character properties alone are not sufficient to describe the complete range of desirable or acceptable character behaviors.

I'm having a hard time trying to understand what's being said here. Could you clarify (perhaps with an example) ?

asmus · **Joined:** Tue Dec 01, 2009 2:49 pm **Posts:** 172

There are several statements here. Which one do you have trouble with, in particular?

Belloc · **Joined:** Sat Aug 06, 2011 9:02 am **Posts:** 43

Quote:

When assigning character properties for use with a given algorithm, it may be tempting to assign somewhat arbitrary values to some characters, as long as the algorithm happens to produce the expected results.

Are we talking about reassigning character properties here ? Why would anyone do this ? To produce expected results ? What for ? This is not clear to me.

asmus · **Joined:** Tue Dec 01, 2009 2:49 pm **Posts:** 172

We are talking about assigning properties in the first place.

When you create an algorithm you have a choice of how you want to create your rules. These rules act on properties. It's sometimes tempting to achieve the correct results by "arbitrary" properties, instead of of properties that group characters by some linguistic or typographic aspects of their identity first, and then write the rules to make things work.

The outcome under the algorithm is the same, but in the former case, it's likely that it will be much more difficult to reuse the properties for other algorithms (or more sophisticated variants of a similar algorithm).

Belloc · **Joined:** Sat Aug 06, 2011 9:02 am **Posts:** 43

I'm sorry asmus, but I haven't got the idea yet. When you say properties, are you talking about the character properties in the Unicode Character DataBase (e.g. UnicodeData.txt) ? If so, aren't these properties already assigned ? What is the meaning of "assigning character properties for use with a given algorithm" ? As far as I can understand those properties are already assigned to each character in the UCD.

asmus · **Joined:** Tue Dec 01, 2009 2:49 pm **Posts:** 172

This is an unnecessarily static view.

New characters get added, new algorithms get defined, leading to entirely new properties - there's a lot of change that doesn't involved "re-assignment" although that also happens.

The model presented in UTR#23 intends to be applicable not only to using the data from the standard, but also to the process of maintaining the standard. In some ways it could also be usefully applied, at least in part, to properties and algorithms defined by others.

Belloc · **Joined:** Sat Aug 06, 2011 9:02 am **Posts:** 43

Quote:

The model presented in UTR#23 intends to be applicable not only to using the data from the standard, but also to the process of maintaining the standard

It's all clear now. Thanks.

The Unicode Consortium Discussion Forum

Relation of Character Properties to Algorithms

Who is online