SIGIA-L Mail Archives: RE: SIGIA-L: future directions for IA
RE: SIGIA-L: future directions for IA
From: Louis Rosenfeld (lou_at_louisrosenfeld.com)
Date: Wed Aug 29 2001 - 16:14:09 EDT
Christina Wodtke wrote:
> [stuff deleted]
>
> So, to turn my attention to the Lou post--
>
> Lou and I been bantering off list on his blog entry. My feeling
> is that his
> list included a lot of the meat of what we are doing already, but not
> talking about on the list for whatever reasons.
>
> *Distinguishing users' information needs
> If one is practicing user-centered IA, and you are doing content
> architecture, you are practicing this. Who is the user, and what are their
> needs in retrieving the information on a given site is core to the work of
> any content-rich site's IA, be it IHT or epicurious. Carbon IQ is doing a
> lot of this for our clients, I'm going to guess we are not that unique.
Yes, you are right, sort of. We all do this to some degree of course.
But how many of us could parachute into a web design war room today and say
"OK, it's clear that we need to support users who know what they are looking
for. So we'll use a site index and a search engine. And hmmm..., there
looks like a number of users are here to learn about a topic. So we'd
better use some sort of thesaural capability to help them grow their
concepts, as well as supporting some means of iterative searching and
browsing so they can refine their concept." Et cetera, ad nauseum.
Yes, we're getting at what's inside users' heads to one degree or another,
but without any sort of communal discussion of what actually constitute
different information needs, and what architectural components work best to
address those needs. If the field is going to become... well, a *field*,
then we're going to have to come up with some shared expertise, not bits and
pieces of experience stored in the heads of individual information
architects.
> *Determining content granularity and
> *Understanding and using metadata
> There is a reason XML talks keep showing up at the IA summits.. Hopefully
> someone is making the connection right now.... Controlled vocabularies are
> key to search, especially search/browse cross use such as yahoo has been
> doing for years.
Partly because I keep lobbying for it. ;-) But I'm sure many attendees
wondered what the hell an XML person was doing on the program.
No, the connection has still not been made. This is partly due to the fact
that we're a new community that hasn't figured itself out yet, which makes
it hard to approach another, more established community. And partly because
the XML community seems to be a little too fixated on tools that provide
syntax to information, while paying too little attention to the semantic
side of that challenge, such as where those controlled vocabulary terms to
populate all those new fields are going to come from. Ergo they've not
figured out that they need us. Yet. Although I've been saying "yet" for a
few years.
> *Developing hybrid architectures
> The nature of the web is almost all architectures are hybrid. Take
> egreetings: one had to design 1st and information retrieval system that
> allowed a user to successfully find a desirable card from one of multiple
> mental models of seeking (occasion, recipient, mood) then *send* it, track
> it, reply to it... from content architecture to interaction design with a
> fine veneer of information design on top we have a fine hybrid. I
> can point
> to everyone's favorite example Amazon, which mixes browse architecture,
> recommendation engines, search and susceptible moments... We're doing it.
>
> (mini-topic drift: does Amazon have IA's? How much great IA is
> being done by
> non-IA's?)
Yes, these exist. Yes, many of us are working on them, whether we realize
it or not.
But again, it's a topic that I don't believe we've explored as a community.
I don't think it's in our communal vocabulary yet. Which is why I'm trying
to agitate here...
> *Presenting search results better
> Is Avi in the house? And does this belong to AI's in their information
> design role, or does this belong to designers?
Certainly part of this belongs to visual designers.
But what I'm getting at is what ways we should be clustering and ranking
results, and what content elements we should be displaying per result.
Should we be using visualization tools, and when?
These aren't really the domain of visual designers. They should be areas we
take on, but we haven't, largely because we're typically stuck with the
defaults of whatever search technology the IT folks have deemed will require
minimal maintenance and take the least toll on their server and network
infrastructures. So the tools have typically driven these decisions, and of
course the converse should be true. What studies are we doing to understand
what users would want and need after they execute a search? *That* is what
should inform tool selection.
> *Rolling out enterprise-wide architectures
> And now we are back to what some folks call big IA....
>
> I don't always agree with this idea that we need a CIAO, but I do
> believe we
> need a CUXO-- we need one person who holds the vision and assures a
> consistent user experience across the company's properties in order to
> protect the brand. And this is so much bigger than our little slice of the
> pie....
I think one person isn't enough. I'm working on a model that involves a
centralized team working providing services to distributed content owners.
This is what Microsoft is already doing, but while we can learn from their
example, it obviously won't work for everyone else. The key is to develop a
broader model--or framework--that would help us come up with the best
variations for our own organizations, whether we're Microsoft or Mike and
Sam's Sporting Goods.
> but we might be a cherry in that pie.
This archive was generated by hypermail 2.1.2
: Sun Nov 23 2003 - 22:54:48 EST
|
