The Unicode Consortium Discussion Forum

One place where block names are used is as part of regular expressions. (See UTS#18: Unicode Regular Expressions for more details).

Almost every version of Unicode adds new characters in new blocks, and occasionally the name of some blocks has changed in the character names list, or an alternate name has been recognized (over time this has affected about a dozen blocks out of well over a hundred).

What should an implementer do to track these changes?

Whenever Unicode changes the preferred name for a property or property value, the old alias is maintained. It is strongly recommended that all programs (and other standards) accept all of the aliases as equivalent.

The full list of aliases are found in:

http://unicode.org/Public/UNIDATA/PropertyAliases.txt
http://unicode.org/Public/UNIDATA/PropertyValueAliases.txt

(The above links are to the latest versions of these files; there are also specific versioned files for Unicode 6.0.0, 5.2.0, ...)

So for the block property, if you look at http://unicode.org/Public/UNIDATA/Prope ... liases.txt, you'll find under Block (blk) the following three cases where there are multiple names for a block:

blk; n/a ; Arabic_Presentation_Forms_A ; Arabic_Presentation_Forms-A

blk; n/a ; Basic_Latin ; ASCII

blk; n/a ; Greek_And_Coptic ; Greek

If a new Unicode version rolls around, when you update your software, you simply make a diff between the new and old versions of these alias files and you support any new names for existing blocks.

To find the ranges for new blocks, check
http://unicode.org/Public/UNIDATA/Blocks.txt

(Thanks to Mark Davis for much of this info).

Now that leads to the next question:

Do the ranges for Unicode blocks ever change?

The answer is "no." Once defined, they are stable.

However, I kind no formal guarantee of this stability in the Unicode Stability Policy, and some very early versions of Unicode did change blocks and block ranges, but at this time, that is only of historical interest.

If I don't want to download all existing versions of the Unicode character database, is there a quick way to find out whether a given block was defined in a given version?

Here's how I would do that. There's a file http://unicode.org/Public/UNIDATA/DerivedAge.txt which lists all characters added for each version.

If you correlate that information with the list of block ranges in http://unicode.org/Public/UNIDATA/Blocks.txt you can determine the lowest version number for which any characters are defined for a given block range. That is the version at which the block was added to the standard.

A small perl script should do the trick.