Diary

Internal Reports

Completed Papers

Papers in Progress

Conference Presentations

Site Visit and Quarterly Reports

Main DLI Page

Web Client- DeLIver

Targeted Study of the First Group of DLI Users: The Physicists

Prepared by Laura Neumann

Semi-structured interviews were conducted with five people belonging to a research group on campus. They were asked about their work flows and practices, how they find information, why they need information, how they use computers, etc. This information provides insights into their information needs and uses and the possible niche that the DL testbed could fill in their daily activities.

This first group of users work in general relativity, and specifically on applying these equations to understanding the behavior of black holes. This rather large project is divided into smaller units, such as learning about the "surface" of black holes, how one black hole would fall into another in a vacuum, and so on. This group is co-located, in quite tight quarters, in fact. They are involved in loosely collaborative efforts with several other groups around the country, with some individuals working more closely with external collaborators on their particular piece of the puzzle. Within this group the atmosphere is collaborative, most papers are written with co-authors, and they have weekly group meetings. Everyone cited their colleagues as one of, if not their most important, information source. The group has a full range of professionals working for it; some undergraduates, graduates, post-doctoral students and professors work in this group. All of these have or are working on degrees in physics and all have (mostly informal) experience in programming computers.

Their information needs cover a full range of time periods and media. For their physics needs, some of the information that they need for their work must be "as recent as yesterday," other information might be from the 1950's or anytime since. This older work is usually found in journals. Other resources include canonical texts and a few theses and conference proceedings. Their need for recent information is often met by the Los Alamos preprint archive. This archive is central for keeping up with their field and with what others are doing. In addition, other core online resources include their own group archive of e-mail, papers, and links to other relevant information; an online bibliography of useful articles for the group; and a broader archive for the large collaboration that they are involved in. The other type of information that is heavily used is material about computer tools and programming. They have frequently used manuals or reference guides for various languages in their offices and go online to find new product information.

The means by which their information needs (for answering a question or for generally keeping up with their field) are met fall into roughly four categories. First, most people have their own "collection" of some sort at their desk. These vary in degrees and means of organization, but most people have photocopies of relevant articles; some sorts of project notebooks containing ideas, notes, solved equations, etc.; and their own collections of relevant web sites. Usually the web sites are kept track of with "bookmarks" in Netscape or through links that have been added to the individual's home page. Some people have copies of relevant theses--usually of students who used to work on the project--and conference proceedings. Material from conferences is seen as less critical: the cost of attending conferences is not worth the benefits; proceedings that consist only of copies of overheads are not useful; and full proceedings are too expensive to purchase.

Secondly, as previously mentioned, this group keeps a fairly complete archive of group e-mail, papers, and presentations, most of which they mount on their server so that it is accessible to the public. One member of the group modified some software to make it easier to attach other sources of information--such as figures or equations or links to other web sites--to documents in this archive. The importance of the server to both facilitating group work and disseminating their results was noted.

Thirdly, each group member subscribes to the Los Alamos preprint service. This service automatically sends each subscriber the author, title and abstract of articles that have been submitted each day. If an article is wanted in full, the person must log onto the archive and explicitly ftp or "fetch" the paper. Although the papers at this site are not refereed, none of the respondents thought this was a problem because the papers are on their way to a refereed journal. The physicists were also confident that peer pressure generally kept out low quality work and that they could weed out the "crackpot" submissions. Further, the people who maintain this archive are very firm about the format, length, and style in which the articles must be submitted, so only serious submitters are likely to send in their manuscripts.

Finally, members of the group have several other strategies for getting information. Respondents did mention going to the library, browsing journals, and browsing on the web. None of these was done frequently for a variety of reasons to be addressed later.

A general layout of the paths by which our informants find information indicates what is most useful to them. First, they consult their own collections of articles, notes, and web sites. If this fails, their second recourse is to ask someone else. Usually they start with those people physically located near them, then move to other people in their group who know more about their particular question. This questioning may result in the exchange of papers, ideas, or suggestions of what to try next. When individuals have exhausted the resources of their colleagues, they will then shift their search outward. Depending on the information need, they might try searching the Los Alamos archives, or getting relevant citations from articles or texts that they have on hand. As a last resort, they may go to the library; although it is only several blocks away, the physical distance is perceived as prohibitive to use.

Obviously, the two constraints on how people search for information are time and convenience. The two practices that our respondents saw as too time consuming and inconvenient (given the likely return on their investment) were going to the library and searching the web. People usually go to the library with a specific reference in hand. They do not see the library as a good place to go to find something more generally. Although the relevant information may be available there, it is too much trouble to walk there and sort through everything else to find it. The online catalog was used to find as much information about what they need from the library as possible before making the trip. Even though they saw the online system as useful, especially for determining the call number of needed items, they mentioned that it was difficult to make sense of.

As for the web, the most frequent comment was that in general, most material on the web lacks substantive and valuable content. In addition, general web searches do not usually produce relevant results. If they use things off the web, it is most often product information about a computer system or program. If someone else points out a useful site, they might go and check it out, but "surfing the web" is usually considered a "fun" leisure activity.

The digital library testbed has the potential to be useful to this group. It provides access, if not yet on their own computers, in a nearby location. Unfortunately, the content of the testbed will not serve all their information needs. The goal is to put in journals that have been published recently, from about 1994 or 1995 to the present. While information in these journals may be useful, these physicists already have a familiar source for some of this information: the Los Alamos preprint archive. The testbed will have to show itself superior or complementary to this archive if it is to be used. The archive contains an alerting service and is searchable and sorted by topic. The information there is perceived as being of high quality and relevance. It is more recent and goes back further than the digital library currently does. Articles are fairly easy to download and print out. The largest drawback is the way that searching is handled: it does not have the indexing capabilities of SGML tags. The only actual complaint about this archive related to the "characters" who maintain it. Keeping this in mind, the digital library testbed must have as many of the capabilities of this archive as possible in addition to a larger and more diverse collection and better searching capabilities. It should also allow for easy movement of material to the users' own personal and group systems. This community will be a challenge to penetrate.