Re: Review of draft-ietf-justfont-toplevel-05

["Chris Lilley" <chris@xxxxxx>]

On 2016-12-09 22:10:14, Dale Worley <worley@xxxxxxxxxxx> wrote:

Major issue:

Section 5.3.4, Collection Font Type (font/collection)

Fragment Identifiers A string, no longer than 63 characters and
restricted to the printable ASCII subset, codes 33 - 126,
except for the 10 characters '[', ']', '(', ')', '{', '}',
'<',>
'>', '/', '%'. If this string matches one of the PostScript
names (name ID=6) [ISO.14496-22.2015] in the name table, that
font is selected. For example, "#Foo-Bold" refers to the
font
with PostScript name Foo-Bold. If the name does not match,
or
if a fragment is not specified, the first font in the
collection is matched. Note that the order of fonts in
collections may change as the font is revised, so relying on
a
particular font in a collection always being first is unwise.

However, the characters '"', '#', '\', '^', '`', '|' are not
admissible in fragment identifiers, per RFC 3986.

It appears from ISO 14496-22 that those characters are allowed in font
PostScript names, so you probably want to enable the use of %-escapes
in fragment identifiers to represent those six characters.

Chris Lilley: 

Thank you for spotting this. I examined the grammar in RFC 3986 and agree that these six characters, which are allowed in PostScript names, would need to be percent escaped.

Because this subsection was getting a bit long (and appeared in two places) I created a new section on fragments for font collections. I added a list of characters to escape and added a second example with an escaped character. i clarified that the fragment is un-escaped prior to comparing with name table entries.

I also added a normative reference to RFC 3986.

I have submitted a new draft -06 with this change.

Minor issue:

>From my correspondence with the author, I think his intention is that
the parameter values should be case-insensitive. It seems to me that
RFC 6838 section 4.3 makes the values case-sensitive by default, so if
they are intended to be case insensitive, that needs to be specified.

Chris Lilley: 

I left these as case-sensitive per RFC 6838.

--

Chris Lilley

@svgeesus

Technical Director @ W3C

W3C Strategy Team, Core Web Design

W3C Architecture & Technology Team, Core Web & Media