Format spec v4 says enum member encoding is ASCII while implementation does not care and compound member name encoding description seems to be a copy-paste error from the opaque section.

**Describe the bug**

Format spec v4: https://support.hdfgroup.org/documentation/hdf5/latest/_f_m_t4.html

Section `Fields: Enumeration Property Description for Datatype Versions 3` contains the following text:

> The name for each name/value pair. Each name is stored as a null terminated **ASCII** string.

But according to [this](https://github.com/Apollo3zehn/PureHDF/pull/171) PR the actual C implementation does not care about the encoding and UTF-8 works in practice. Who is right? I guess everybody wants UTF-8 support so I suggest to update the spec to reflect the reality :-)

The second issue is with the compound member name encoding, section `Fields: Compound Properties Description for Datatype Version 2`. The string encoding is not specified at all, which is not a problem when it means that any encoding works. But it looks as if the string

> This NUL-terminated string provides a description for the **opaque** type.

has been copy-pasted from the opaque section and the wording should be adapted to better fit to the compound data type.

**Expected behavior**
Correct spec wording.