Sorry for joining in naively to a conversation I've not been following, but
reading Karl's remarks on the facilitation of metadata entry for users, some
discussions in the vicinity of the recent SVGOpen that concerned usability,
accessibility, and metadata made me think the following (that I suppose is
rather outside the realm of HTML):
Suppose the user or author (since in an app the distinction is blurred
somewhat) is building something like a graph (in the discrete math sense),
an image repository, or even a diagram (though the categories of content
here are heterogeneous, making the argument a bit more tenuous) using a
guiwebapp (like inkscape for diagrams or
http://srufaculty.sru.edu/david.dailey/svg/graphs30.svg for graphs). Let's
say there are n basic entities (like graphs or images) for which metadata is
required. Let us furthermore assume the metadata description language is of
order 0 1 2 3 or 4 * and that the minimum number of user operations
required to complete the metadata description for a single entity is bounded
above by k.
We then may plot a user performance function that estimates the probability,
p, that users will actually succeed in entering data (as a function perhaps
of not only n and k, but of the user's investment in the process). Clearly
as n and k grow and as the user's investment in the process declines, so
does p. We are interested, through, interface, in maximizing p.
I have a hunch (in math it is called a conjecture, but in CHI it is more
like a hunch) that not only how, but also when, this conversation between
user and software takes place affects the probability. For example if an
artist were using Inkscape to draw SVG, then mandating a conversation about
metadata each time a curve or gradient is completed is likely to drive users
to AutoCad for their diagrams, even if wine is served.
In certain cases, it makes most sense to build that conversation as an "exit
interview". If we will have k phrases to enter (using a grammar of graph
theoretic phrases) for each of n objects, then we may wish to build a very
comfortable GUI to facilitate that for all the affected entities upon
closing the app: Dear user, you have just completed a schematic drawing for
the Intel i-Chore 42x processor, would you now like to a) save b) enter
appropriate metadata c) save and enter data d) drink wine. The notion is
that a GUI enabling such, could if it were viewed as a stage or mode of
development a) rely on the visualization of the opus as thus far created b)
be appropriately rich to the order of the metadata description language and
c) make the data entry process unbundled from the creation process, hence
allowing diversification of the assignments of tasks to workers (e.g. the
familiar phrase of the assessment revolt of 2028: "let the bureacrats do the
bureaucracy!"). That isn't to say that we should not also facilitate the
entry of data at each stage of the drawing process, with a sub-interface of
the master metadata editor, but given the complexity that some metadata
editors may have to convey, the nature of the conversation between user and
software may not be allowed to remain entirely casual (that is, wine may
need to be upgraded to tequila).
/fwiw
David
(by the way, an Intellectual Property/provenance description language such
as the library and visual rights communities work with might be an
interesting overlay for the web, provided both free and corporate models
(together with ample graph theory) are included)
* define the order of a metadata description language as 0 if it consists of
simple non-delimited strings, 1 if it consists of delimited strings (with a
single delimiter), 2 if the delimiters are parentheses (required to match),
3 if the delimiters act like parentheses of multiple flavors as in XML, and
4 if the language is fully graph theoretic (parenthesized strings plus cross
linkages -- footnotes).
----- Original Message -----
From: "Karl Dubost" <[EMAIL PROTECTED]>
To: "Henri Sivonen" <[EMAIL PROTECTED]>
Cc: "Ben Adida" <[EMAIL PROTECTED]>; "Paul Prescod" <[EMAIL PROTECTED]>; "Ian
Hickson" <[EMAIL PROTECTED]>; "WHAT-WG" <[EMAIL PROTECTED]>
Sent: Sunday, August 31, 2008 11:20 PM
Subject: Re: [whatwg] Creative Commons Rights Expression Language
Le 29 août 2008 à 23:04, Henri Sivonen a écrit :
Also, having more metadata leads to UI clutter and data entry fatigue
that alienates users. In the past, I worked on a content repository
project that failed because (among other things) the content upload UI
asked for an insane amount (a couple of screenfuls back then; probably a
screenful today) of metadata when it didn't occur to system specifiers to
invest in full text search. More metadata isn't better. Instead, systems
should ask for the least amount of metadata that can possibly work (when
the metadata must be entered by humans as opposed to being captured by
machines like EXIF data). See also
http://www.w3.org/QA/2008/08/the-digital-stakhanovite
hehe. This was a-good-try-but-mischaracterization-from-the-ministry-of-
truth to associate this article with the rants on metadata :) Let's
clarify.
What I explain in the article is not the volume of metadata, but the
volume of items and the context of usage.
1. Extract anything you can from the data itself (exif, iptc, xmp,
modifications, date)
2. Give a possibility in the UI to modify or add data.
In a business environment, you might have to give metadata about a
work. I do it in my every day job. I give titles to my emails, I put
comments in my cvs commits, etc. etc. These are all constraints. Not
adding the data would still work technically.
For my own personal photo, I don't (want/have) time to put plenty of
metadata. And that's fine. I do though bulk metadata at a regular
pace, for location (ex: all these selected photos have been taken in
Taiwan with the help of GUI tools. Yes tools save my life).
Having a UI cluttered with fields to enter is not a failure of
metadata, it is a failure of the project in the social and business
constraints of the project.
--
Karl Dubost - W3C
http://www.w3.org/QA/
Be Strict To Be Cool
