Mar
I’ve been remiss in responding to recent blog comments. This blog’s readership has a number of insightful people and I would like to acknowledge, and respond to, some recent comments.
Peter Murray commented on the challenges of incremental results. He made the point that trying to get users to do things differently than they’re used to shows that you’ve not designed your product properly. I get the point but federated search isn’t Google, and serious research isn’t Google. I think it’s ok for some software to be complicated enough that it’s worth training users to understand how it works. Google is easy but it has limitations. There are entire businesses devoted to software training and documentation.
Stephan Schmid commented on the same post. He is letting users decide how long to wait before source time out. I’m interested to know how users like that feature. Do they use it? Do they understand it?
Stephen Francouer commented that he enjoyed the basics articles and the What is a connector? article. Thank you for the feedback.
The post “Is federated search “ranking impaired? caused a bit of a stir as I flagged a statement from Toni that federated search couldn’t do relevance ranking. Tom Wilson, who Toni tried to paraphrase, clarified his points, and Abe added to the discussion. The upshot of all these comments is that federated search ranking is a very complex beast, and it is limited severely by the incomplete information available to the federated search application. The application doesn’t get to see all of the results and, within the results it does see it just sees meta data information (title, author, summary, etc.) Then you have to deal with the difficulties of comparing result summaries from one source with those of another, if you want to find the most relevant results across all sources.
Exactly three people responded to my call for reviewers of Christopher Cox’s book. I’ve mailed out the books and I’m looking forward to the reviews in the coming couple of months.
Linda, the lipstick librarian had a fun response to Which librarian blogs cover federated search the most? And, Paul Pival pointed me to LibWorm, a search engine for the library community. Libworm is an amazing site; it aggregates 1400 RSS feeds and makes their content searchable. One can also easily track search results in an RSS reader to see when there are new articles matching one’s search terms. And, LibWorm has a subject, Federated/Meta Search, that one can read via RSS. Very nice! There are tons of blogs aggregated here. Now I need to figure out how to mine the LibWorm data to see which blogs write about federated search the most.
Three readers responded to Search or authenticate first? Readers gave input on how the search/authenticate issue is being handled in the real world.
I’m grateful to the readers of this blog. Many of you have real world experience with application I don’t and I’m always happy to learn more about the industry from your comments. And, I’m delighted to be caught up with blog comments (for now).
If you enjoyed this post, make sure you subscribe to the RSS feed!
Tags: federated search
2 Responses so far to "Response to recent blog comments"
April 2nd, 2008 at 5:48 am
Sol,
One of your correspondents criticized the user friendliness of the feature by which early hits on WorldWideScience.org or Science.gov are posted and then, later, the user can update the hit list with the late arrivals.
Your response talked about the ubiquity of computer-based tools that require training to use, thus implying that it is OK for WorldWideScience and other federated search products to require training to use.
In my view, a better response would make the point that yes, the two-step hit list is different than Google, but it is nevertheless intuitive and does not require training. What OSTI seeks to do is to make all its products intuitive. We consider ourselves to have failed if users need to take a training course to use one of our products. While Google is simple to use, and we emulate that simplicity whenever we can, we think we have achieved our purpose if our applications are intuitive, whether or not they follow the Google model. We think the two-step hit list achieves our purpose by being intuitive. This is a general principle that needs to be advanced on your blog: it is essential for applications to be intuitive, which may or may not mean that they are Google-like in their architecture.
Walter L. Warnick, Ph.D.
Director, OSTI
April 2nd, 2008 at 8:27 am
Sol,
I never did a scientific study on this subject, but discussed it extensively with a couple of colleagues and friends some years ago.
Do they use and understand it?
What I can tell you from looking at the statistics: yes and no. The advanced users that enable the search information make heavy use on it. Others don’t realize this (and other) features at all.
Since it is very seldom that a source does not respond within two seconds, it is also not that important.
Greetings from Switzerland
Stephan