[whatwg] <blockquote cite> and <q cite>

Benjamin Hawkes-Lewis bhawkeslewis at googlemail.com
Fri Jan 19 09:17:46 PST 2007

Leons Petrazickis wrote:

> On LibraryThing, I use book title or sometimes author + book title. I
> only pull out ISBN for disambiguation with old books that have too
> many editions, but where I want the right cover to show up.

People may vary. I use ISBN in such cases every time. If you were
quoting from the book, you'd generally need the exact edition so I
suspect you'd find yourself making even more use of ISBN. As far as I'm
concerned, any citation UI should allow entry with either method.

> ISBN is not an intuitive concept. 

I'm not sure about that. It's not a very complex concept. It's just a

> It's like being asked your account number when you call your tech support. 

That's intuitive, it's just not very memorable.

> Account numbers are not human-readable. Neither are ISBN. You can't
> proofread them for typos, 

You can, but not as easily. But with a lookup UI you don't need to, the
result will be either right or wrong.

> I have nothing against citing by ISBNs, but if I ever cite, it will be
> by book title and author, by movie title and director, by song title
> and singer. HTML5 can support both, but it should at least support the
> latter.

You can recommend books by author and title, but for most books it is
insufficient to /cite/ them by author and title alone, especially if
it's a citation for a quotation. More generally, you seem to be talking
about what a citation should /look like/. I'm really talking about how a
citation /should work/. I don't propose that ISBNs are the key part of
how citations should be displayed, but merely that they are clearly one
of several useful components for how citations can be used by machines
on behalf of people, e.g. as part of an OpenURL.

OpenURL documentation is not easy to understand, but from what I've seen
current OpenURL formats are primarily designed for querying, not for
local extraction alone. That is, you cannot extract "full" citation data
(e.g. working out which part of a second author's name is the sorting
part) from current OpenURL formats, without querying an
OpenURL-supporting registry. I suspect that specifying a richer OpenURL
format that could reduce latency problems due to lookups could be part
of hCite work, perhaps in cooperation with OLCL (that is, hCite could
define an alternative serialization in the form of an OpenURL). Rich
URIs which contain full citation data offer an efficiency saving over
Ajax-style network requests and a measure of data protection again
serious network catastrophe by encoding fuller information within the
document itself. (I'm thinking here of how classical scholars
reconstruct ancient writings from scholia.)

Having said that, I emphatically do not believe we need richer URIs than
OpenURLs already provide for OpenURL to be useful (e.g. first author
last name, title, journal name, page start, page end). In particular,
OpenURLs of this sort encode sufficient information for it to be
possible to /automatically/ rework old URIs to contain fuller
information if a still richer format is devised in the future.

It's worth stressing that it's possible to use such URIs in ways which
are intentionally vague (e.g. author, title) or ways that are very
specific (full citation data with edition and page number). A good UI
and spec should allow both.

I would also contend that it's not actually all that /bad/ if a spec
encourages the checking of citations and quotations; indeed, it's a
positive good. Human beings are extremely fallible, and mistakes in the
citation of authorities are intellectually and socially pernicious. The
trick is to make such checking as automated by UIs as possible.

The general idea here is to exploit the ever faster digitization of
books, archives, and music to create an increasingly seamless hypermedia
experience where you can move directly from citation to work, rather
than merely recreating a labour-intensive footnote that evolved within
the limitations of the Age of Gutenberg. URIs like OpenURL can act as a
bridge between the old world and the new. By making it easy to cite, we
will encourage people to back up what they say and for others to check
what they have said. By making it easy to cite, we will make it easy to

Benjamin Hawkes-Lewis

More information about the whatwg mailing list