In a message dated 2010.05.25 12:52 -0500, NoOp wrote:
May be we should speak about the three Liberation fonts,
created to have the same metric properties than the classical win fonts ?
never got any problem on different O.S. when using those fonts
Maybe we should - but that begs a number questions:
(1) ... Liberation Serif/Sans/Mono of course only cover the most
basic font requirements, and even then do not provide a match to
satisfy everyone. For example, the slab-serif Courier New is
probably the most common monospace font for Windows; Liberation
Mono is a good match in character metrics but not in appearance. ...
(2) How does OO do the font matching? A "replacement table" can be set
up to substitute fonts by name, but that just seems to be an override
for whatever algorithm OO would otherwise use. What is that algorithm?
Where does it reside? How does it work?
Studying VCL.vxu, it seems that fonts are matched, not on typeface
attributes and metrics ..., but by name, with a list of
substitutions for each name, and a fallback specification of
"serif" or "sans-serif". This is pretty thin stuff, and it begs
the question of how those font substitutions lists are generated.
Most font matching questions, such as posed in this thread, would
probably go away with answers to these more basic questions.
Can't answer your questions, but this might be of interest:
<http://www.oooninja.com/2008/02/metrical-equivalent-fonts-and-font.html>
Thanks! - it was helpful for the explanation, and even more for the links.
One of those links, to OpenDocument spec section 15.4.15, on font
matching
<http://www.oasis-open.org/committees/download.php/12572/OpenDocument-v1.0-os.pdf>,
shows that what I described above (from studying VCL.vxu) is pretty
faithful to the ODF spec. But another
<http://qa.openoffice.org/issues/show_bug.cgi?id=15687> while discussing
the problem of non-Roman glyphs, acknowledges weaknesses of that
name-based system. In 2004 the lead engineer described it as "an area
that needs a lot of work"; I gather that is still mostly true.
However, it was nice to see the intention to adopt the PANOSE structure
[a successor to the WPDL structure mentioned earlier] for classifying
and matching fonts, which would do away with most, if not all, of these
problems. Unfortunately, it is (A) not envisioned in the ODF spec and
(B) non-trivial to implement, not least because of its requirement for
accurate metadata tags on the installed fonts.
- (A) should not be an insurmountable problem; after all, specs are
made to be improved.
- (B) is the problem that hindered WPDL font matching: Font suppliers
could not be bothered to characterize their fonts properly, and
WordPerfect did not have a group to enforce its font description
standards (at a time when its market share might have allowed such
enforcement). As a result, most users (those insufficiently obsessive
to correct the descriptors and metrics on their installed fonts) just
lived with poor font classification and matching. I'm afraid the same
will be true for PANOSE, if there is no group to normalize/enforce font
metadata. [I've tried to think about a market-oriented way to facilitate
such normalization, but I'm afraid that market pressures may be even
weaker than they were 20 years ago.]
John
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
