Web font linking progress summary


Here's my reading of the state of play. It might be useful.

There is not a need for a new /font/ format. Nobody thinks it is a good 
idea - there are enough issues working with operating systems, 
applications, and TTF/OTF - let alone bitmaps, old-school Type 1s ...

There /is/ a need for a wrapper. Tom Lord made the original case and 
there is acknowledgement that this proposal stands with other related 
work outside of typography, such as Creative Commons' Mozilla plug-in 
(http://wiki.creativecommons.org/MozCC), that express the rights and 
attributions of original content to web users.

There /is/ a need to find a way to express the rights of use which allow 
the users of a web page to enjoy the enrichment of linked media. When 
they use linked media, publishers are currently unable to use 
machine-readable means to credit the contributory work of others or give 
any indication that it is not in their authority to let others make free 
use of it. They thus gag people like photographers and type designers, 
whether or not those people wish to share their work at second hand. 
This is an online culture issue, not a type issue, and it will snowball 
unless browser developers and the web development community agree that 
they should start to express those rights in a sensible way.

Right now, if you judge their feelings only by looking at what's 
shipping, browser developers do not care. To make them care, rather than 
"unnecessary extra work for no gain" (Mikko: 
http://lists.w3.org/Archives/Public/www-font/2009JulSep/0385.html), the 
feature should be compelling both technically and socially - that's how 
ideas usually get moving...

Tal and Erik have proposed a model for a wrapper that encompasses all 
the things that font vendors think they need to enter the marketplace. 
It could be rewritten in xhtml+rdf. They explained that they see the 
'allow' metadata as the expression of a grant of license and not a 
mechanism that should stop the user from seeing the font in use, even if 
it's being loaded against the expression of the license. Their 
contribution helps us all to see what the font vendor baseline 
specification is. I reckon it looks a lot like what other enlightened 
creative people in photography, video, sound, etc would want.

Drawing on Tom's proposal, it seems that to express the rights of the 
creator of a linked work, there should be notification to the 
consumer/beneficiary of that linked work. Something like an alert icon 
in the browser would be appreciated as a first step; perhaps 
right-clicking on the item would show a menu that included links to the 
license and the creator's own web presence (tricky for fonts - 
right-click on any text I guess). The license contravention (the use 
without a matching license) would typically happen if a web 
developer/designer linked to items on a server that they do not have any 
business agreement with, and this is something that web masters will 
want to want to prevent anyhow as it drains bandwidth. Sites that trade 
on making bandwidth available to host images, font files etc can do the 
exact reverse if the content creators agree (by using a permissive license).

Dave Crossland and I have explained at high level how web masters can 
currently control access using CORS and/or referer blocking on the OFLB 
wiki. Neither technique expresses a license but both can be used to show 
compliance. These techniques put the work where it belongs - at the web 
publisher's end as opposed to server or browser software authors - 
although I suspect that referer blocking will primarily be driven by the 
desire to reduce bandwidth. 

What's left?

Tom's proposal (http://noeot.com/notices.html) is currently vapour and 
it is competing for attention with others that are more or less similar, 
as Bert has pointed out (listed at 
http://www.w3.org/Policy/pling/wiki/PolicyLangReview). For me, Tom's 
proposal has the great merit of being a 'side channel' that is 
compatible with any conceivable media or software file format but I 
guess this is a red rag to people who are still weaning themselves off 
DRM (and/or the nightmare of seeing their thousands of hours' work turn 
to internet roadkill). For me, Tom's wrapper *is* Erik and Tal's 
wrapper. But technically and architecturally more elegant.

I think we are all slowly coming to see that the question of people 
downloading the font files to the desktop and disregarding any licensing 
terms is a total red herring. What we want is to express rights, 
promoting the people who did the original work and fostering a 
co-operative online culture, rather than wasting energy trying to stop 
people being naughty.

Whatever, we owe it to the web community to either destroy Tom's 
proposal, improve it, or push for its adoption.

In any case I think the discussion so far has been excellent. In 
particular we have already destroyed:
- slicing up fonts into bits
- changing font names
- shunting font table data around
- the original EOT proposal from way back when.

And left the door open to:
- efficient compression
- intelligent subsetting
- generic solutions to media attribution and rights needs
- innovation in web design, font sales techniques etc.


Ben Weiner | http://readingtype.org.uk/about/contact.html

Received on Thursday, 9 July 2009 14:27:01 UTC