"I pressed F1, but you didn't come over to help."
"If they are clicking they are looking for information. If they are typing we tell them to stop because they are using Hotmail."
"The most important issue in electronic delivery is printing."
...Just a few quotes from the American Library Association conference in Washington DC at the end of June. Why was I at ALA? Well, like a lot of you who go to the Online Exhibition in December I entered various free draws without much thought for them. Thus I was surprised a little later to get a call from Information Access Company to say I'd won their draw, but even more so to learn the prize was not the usual bottle of champagne but a trip to the ALA Conference.
This event is not like the typical UK conference with a couple of hundred people on a university campus and with a general theme. It attracts around 20,000 people and has hundreds of events in various venues across the city, organised by the divisions and interest groups of the ALA and covering all aspects of librarianship. Some events start as early as 7.00 (not the ones I went to!); others don't finish till 10.00 or later in the evening and there are dozens happening simultaneously. Though there is a thick printed programme some activities are known only to members of the particular groups involved and it is really only possible to sample the conference. This report similarly is illustrative of a range of presentations, poster sessions, discussion groups and so on.
Some interesting work at the University of Arizona [1,2] brought familiar conclusions: users don't read instructions or subject guides; their needs are not what librarians think they are; explanatory text on pages may be ignored; there must be incentives to participate in focus groups, with a going rate of $10, or a free lunch for faculty. Other points of interest were: graphical icons can substitute for words, so a picture of a magazine or newspaper can stand as a link to indexes; the positioning of options on a screen may imply their relative importance; have a "where to find" drop down menu, for example where to find maps.
At the University of Wisconsin  they used the QUIS questionnaire, while a study of the JSTOR full text journal collection at the University of Michigan using Web Assisted Self Interview (WASI) showed: a decrease in use over two years (though it was accepted that meaningful longitudinal studies are difficult); higher use by economists than historians because more of their core journals were included; printing was important, but difficult to achieve because of poor technical support; users like the search facilities.
A study on the preference of library school students for the Web or telnet interface to the Dialog system used the concept of emotive aspects of searching  - stress, frustration, surprise - as a way of evaluating interfaces. It showed familiarity with techniques and skill levels as being more important than the actual interface, as long as support is given.
Hotbot's Marketing Coordinator told of their extensive user testing, based on observation, focus groups, online questionnaires, e-mail feedback and query logs. He concluded that: people don't know how Web search engines work and this earned a cheer of agreement from the audience; the average query uses 1.7 terms; the 'Expert' label had deterred people, so was changed to 'More options'; people do not use the help files; and it is hard to get an interface that will satisfy all 8 million users.
An overview paper on methods divided them into those like usability labs, interviews and surveys which are largely retrospective, and those like transaction logs and content analysis which are more real time. Most studies rely on the former.
A general issue that arose from the session was of privacy violation in identifying users from transaction logs. One study logged IP addresses but gave each a random number in order to track usage, but without knowing who the users were.
A plea for more user based online help suggested a need for hints, examples, error details and auto detection - word processors can detect spelling errors so why cannot search services? In addition librarians should provide roving assistance. Although users may seem to be coping, they are often just reluctant to ask for help, so staff should be more proactive. Other user studies are cited in the bibliography .
The librarians who spoke on this topic were sceptical. One urged the audience to call them "publishers' DOIs", although the Director of the International DOI Foundation  suggested "authors' DOIs" was more appropriate.
The Foundation now has offices, one member of staff, has raised some money from members such as publishers, abstracting services, Microsoft and copyright bodies and will soon issue a White Paper. Other world bodies like ISBN International, ISO, NISO and Internet organisations are also getting involved.
There were a number of queries raised: about what is unique - HTML and PDF versions would be different; how usable DOIs will be; is it all about rights management?; can DOIs cope with the same information in different locations, such as an electronic journal? There were also assurances that free publications, such as those from governments, could be accommodated.
The project is huge given the potential number of objects. Time Inc has a collection of 20 million digital images, but in 3 years only 500,000 have been catalogued. Mmovies are analysed for every 5 seconds of image and music. Given the size of the task, retrospective processing may be limited.
What will be the impact on libraries? the answer was how do libraries want to use them?
The talks dealt with collaborative projects statewide and in other ways where the aim, at least at present, is to make the partner library catalogues available to all (usually with OCLC SiteSearch) and to have general access to things like core full text journals and FirstSearch databases, rather than to bring into one institution the range of national and local material that UK hybrid library projects are attempting. Galileo in Georgia  is the longest established and covers universities, public libraries and schools. Key to its development has been political support from the State Governor and $10m from the Georgia Lottery.
CIC, or the Big Ten , is an upper Mid West collaborative scheme with similar features to Galileo. They set up a Centre for Library Initiatives to coordinate development and saved $1m by negotiating licence fees centrally.
A third presentation was on MnLINK , a project in Minnesota which is still at the planning stage.
The Library of Congress Web page has a feedback facility. Questions range from the sensible to the ignorant: send me a catalogue of everything you have; Isn't everything in the Library electronic full text? It was also stated that as enquirers can be impatient, sometimes mailing again after 10 minutes to see how their enquiry is progressing, the scope and limitations of such services should be made clear to users.
There were also: a comparison of the coverage of full text journals by different services ; two attempts at guiding users through the information research process [11,12]; an outline of a programme to promote electronic resources within a university ; the tale of the university library director who did not know how to maximise his browser window in Windows95. And just to show it wasn't all PCs, the recently restored domed reading room at the Library of Congress is wonderful.