unicode-org Files

The beta version of the UCD includes nine files which are intended to
help beta testers evaluate the changes since the previous
version. Those files are not part of the UCD, and will be present only
during the beta period.  Those files only informative. While we
believe that those files present accurate data, there is no guarantee
that the data is indeed accurate. The files are not intended to be
machine readable.


All files cover:

- the code points which have been assigned
- the properties which have been added or removed
- for each property, the code points for which the value of the property
  has changed
- the changes to blocks
- the changes to named sequences
- the changes to normalization corrections
- the changes to standardized variant


The *.summary.diffs files give only a summary, in the form of the number of
things which have changed; the *.changes.diffs file add a list of the new
characters, and a lists of the property changes for existing
characters; the *.all.diffs file adds the property changes for the new
characters.

The *.nounihan.* files cover only the properties not in the Unihan
database. The *.unihan.* files cover only the properties expressed in
the Unihan database. The *.all.* files cover all the properties.

---
<x>-<y>.nounihan.summary.diffs contains only the number of things
(code points, blocks, etc) which have changed between versions
<x> and <y>. The format is

<before> + <added> - <removed> # <changed> = <after> (<ignored>, <undefined>)

where:
- <before> is the number of things in <x>
- <added> is the number of things added in the <y>
- <removed> is the number of things removed in <y>
- <changed> is the number of things changed in <y>
- <ignored> is the number of things which are not accounted
- <undefined> 

---
<x>-<y>.nounihan.changes.diffs gives a bit more details: the code points
which have been assigned are listed, together with the character name; for
each property, the code points with a changed value are listed, as well
as the old and the new value. The new or removed blocks, named sequences,
standardized variants are listed.

---

<x>-<y>.nounihan.all.diffs does not ignore the newly assigned
characters. The most important consequence is that if a property of a
newly assigned code point has changed, then the change is counted and
listed. However, this does not means that all the values of the
properties for all the newly assigned characters are listed: for
example, the combining class of unassigned code points is 0, and it
remains 0 for most newly assigned characters. Hence only a subset
of the newly assigned characters are listed for the changes to 
the combining class property.

---