[Openfontlibrary] Font site wishlist

[Raph Levien]

It sounds like we've got a very active discussion going, about issues
including the direction the openfontlibrary is to take. I'd like to
take this time to set forth my wishlist of what I'd like to see in
such a site. Much of this material is reiterated from a typophile
thread from almost a year ago, but this wishlist is pretty
self-contained.

http://typophile.com/node/16620

My central question is whether I can use the openfontlibrary (or some
other site) as the primary site for the distribution of my free fonts.
I just have them thrown up onto static pages in my own domain now, but
that's unsatisfying for a variety of reasons. I'll categorize the
items into three baskets. Basket A is stuff that's a dealbreaker -
without it I'll be looking elsewhere. Basket B is stuff that I think
would be very useful and worthwhile to pursue, and should be done if
the site is to be truly viable. And basket C is stuff that would be
nice to see, but is a long-term goal at best.

Basket A

Thumbnails have already been discussed.

The single most important need that is not currently served is a "type
tester." All viable commercial font websites currently have such a
feature.

In many ways, doing a type tester for free fonts should be easier than
in the commercial world; a strong need in the latter is to make images
available without letting the font itself escape. Many type testers
(such as Hoefler & Frere-Jones) put bars through the image to
discourage people even from using that. In the free world, there is no
need for such games. In fact, it makes sense to make the tester
featureful enough for people to actually use it to make png's.

In the commercial world, the dominant technologies are Flash and Ajax
(with server-side rendering to bitmaps). Flash is a royal pain and not
very open, so I think it's not appropriate. Server-side rendering, of
course, requires nontrivial development on the web server side,
including of course the rendering. Perhaps the FontImage() procedure
in fontforge could be used, or it wouldn't be hard to hack something
up using FreeType.

There are other emerging technologies which may work, in which the
rendering is client-side. An SVG viewer is one possibility, which is
very standards-friendly, but will otherwise make for a pretty unhappy
user experience - it will basically be Firefox-only. Another is to use
the @font-face css2 rule, but I have no idea how well that's been
implemented either.

Quality isn't so much a checklist feature, but it's important enough
to me to mention it here. Since I'm not getting (much) money for my
free font work, it's essential to pay me in ego-gratification. The
worst way to do that would be to have my fonts mixed indiscriminately
with a lot of freeware crap. If I'm going to point people to
openfontlibrary as the main distribution point, they have to see
_good_ fonts on the front page.

Quality is also an issue for feedback. As a designer, I'm especially
interested in specific criticism (one example from Inconsolata was the
observation that accents were too small, and acute and grave were too
easily confused). I'm also particularly interested in requests, for
example for additional glyphs. Not enough feedback is a problem, and
just as bad would be lots of _uninformed_ feedback, so that it's work
to pick out the interesting bits.

Basket B

One thing that can only be done in the free font world is distributed,
collaborative development. In the short term, I expect this to be
primarily contribution of glyphs to expand the unicode range of a
font.

In the short term, such collaboration can be done without any special
tools or support on the site, but I think ultimately having the
repository for the font _hosted_ on the Web is the way to go. Ideally,
contributors would be able to upload new glyphs (or improved versions
of existing glyphs), and then users of the site would provide feedback
on whether the new glyphs were good.

For me, drawing a-z is the fun part. Doing all the glyphs to fill the
code ranges takes a lot of time and isn't as fun. A site that lets
some of that work be distributed would be considerably more rewarding
for me. This type of collaboration is also a good way for more
experienced designers to teach newcomers, especially by providing
detailed criticism.

The features needed to support such collaboration (essentially some
form of version control, where the glyphs in a font may exist in
numerous versions) could also be useful for users. If there were
several versions of the same glyph, then users could vote with their
feet and select the version they like best, "cutting" a custom TTF
file encapsulating those preferences (see the typophile discussions
entitled "altern at ive @ signs" and "Freestyle Remix Challenge" for
hints about how this may go). And of course that metadata is very
valuable as well.

Linux accomplished many things that the proprietary vendor Unices
could never touch. I sense that there may be similar potential in the
development of free fonts. One of the reasons I'm spending time on
these email threads is to encourage the tools to come into existence
to give that potential a chance.

Basket C

It would be cool to be able to apply scripted effects to fonts online.
Simple examples include stroke offset, global changes to side
bearings, and so on. If we had Multiple Master fonts, then generating
an instance would fall into this basket.

Having a font editor in the web browser would be cool. At least two exist:

[flash] http://type.gizma.com/tool.html
[safari/firefox canvas] http://erik.letterror.com:8306/glyphServer

Even without the ability to edit outlines interactively, a tool to
position accents for composed characters would be useful and would be
an easy way for people to contribute.

Conclusion

In sum, I'd very much like to see this happen. If the UKFSN grant
money would catalyze that, I'm very much in favor. I don't have any
spare time right now to do webserver hacking, but I'm more than happy
to offer ideas and criticism.

Raph

P.S. I pretty much take back my ininformed speculation about
licensing. I did study the licenses pretty carefully and came away
with the feeling that the GPL with font exception still had serious
ambiguity, but I could simply be wrong about that.