Ralph LeVan said

<<As part of the VIAF project, we spend massive amounts of time trying to parse names.  They are *not* broken into separate fields.  Last name?
One part, two part, more parts to that last name?  Is it Beethoven or van Beethoven?  Names in some cultures are *really* complicated, but they are all jammed into a single subfield with some minimal punctuation to separate them.  Good enough for a card and usually good enough for a display, but really tough for smart machine processing.

That's just the most common composed field.  There are many others.
Break them apart.>>

We need much more granularity in personal names so they can be made machine-actionable. There are many more personal name forms than forename surname. What about patronymics/matronymics, clan names, the various elements of Islamic names (such as Arabic kunyah and  laqab), etc. 
Once broken apart they may be reassembled however a particular system, set of rules, display convention, or individual researcher requires. We have already seen the problems with MARC in the move from AACR2 to RDA where $c creates problems because of the various types of names which have traditionally been inserted in the subfield.

And quoting Ralph again  " That's just the most common composed field.  There are many others."

Wayne Richter
Asian Materials Specialist/PCC Liaison
The Libraries
Western Washington University
Bellingham, WA 98225-9103