I see that the definition of <length-range> has changed from 1.0 to 1.1 to
explicitly differentiate between the <length> datatype and the <percentage>
datatype. This makes intuitive sense, and the clarification is welcome.
However, there is still at least one place in the 1.1 Recommendation that
confuses or even contradicts this distinction, at Section 5.11, at the first
"Note" item (designated by a hand with a pointing finger in the PDF
version), which says: "Since a <percentage> value, that is not interpreted
as "auto", is a valid <length> value it may be used in a short form."

It is difficult to tell whether this is intended to be a general statement
for all cases, or one that applies only to the "space-before" example in the
context. However, neither seems to be appropriate. The "space-before"
property says that percentages are "N/A". And, if my understanding is
correct that "86%" should not be accepted as a valid <length> item (although
it should be accepted as a valid percentage item where percentages are
valid), then the statement is not correct for the general case either.

On a related note, it seems to me that the "Value" definitions for the
space-end, space-start, and leader-length properties (and perhaps other
similar properties) are not quite correct, or at least not as clear as they
could be:
1. It may be appropriate to clarify the definition of the <space> datatype
in a manner similar to the way <length-range> was clarified, that is, by
adding the comment that "A property may define ... additional permitted
values and their semantics; e.g. ... <percentage>."
2. My understanding of the Recommendation taken as a whole is that it is not
accurate for any of the three properties mentioned (and perhaps others) to
list <percentage> as a valid value. In all of these cases, the value (unless
"inherit") will be a <space> datatype (for space-end and space-start) or
<length-range> (for leader-length). Instead, what appears to be meant is
that the <percentage> datatype is added to the list of valid datatypes that
can be accepted for the length-related minimum/optimum/maximum components of
these datatypes. In other words, <percentage> is only a valid value for
these properties in the sense that it is a valid short-form value. If that
sense is the reason it is listed as a valid value in these properties, then
certainly <length> should be listed as well.

Of course, it is quite possible that I have misunderstood the whole thing,
in which case any clarification that you can provide will be appreciated.
Just to be clear, I am trying to make sure that a value like "86%" should be
treated as an error in places where a <length> is accepted, unless there is
an additional explicit stipulation that a <percentage> is also valid in that
place.