Inventory of Possible Lead Users

This page inventories possible lead users. The purpose of this document is to list possible lead users and then to write a paragraph or two about that possible lead user's fit as an early adopter and collaborator. Many otherwise welcome customers may be inappropriate as early collaborators.

Key questions to ask of every potential lead user:

• What services would you commission if you were to be one of the initial user/collaborators?
• What would define the community of submitters to the system?
• What would define the community of retrievers from the system?
• What is your primary purpose in providing a repository for these publications?
• • a teaching tool
  • a publicity tool
  • archival, historical information
  • other
• What metadata tags are required?
• Which fields must be indexed?

Identification of a particular department, lab, or center to collaborate with and host a preprint service seems most appropriate at this time. Ideally the preprint service should allow for a follow-on project to house the source data underlying the papers preprinted.

These are the various lead users under consideration.

The current retrospectively scanned collection of MIT Thesis pages held on theses.mit.edu represents what customers perceive as an interesting and complete enough collection to use. At this time, however this collection should take lower priority than outreach to customers producing new content. This collection may prove useful to adopt at some point to give the infrastructure credibility.

There is an electronic thesis submission project that does represent new content. Unfortunately, the base of submitters is the diffuse community of graduate students, each one with unique workflows. It seems like this community will be too difficult to reach out to as a lead user community, and that the best success is to have a facility available that graduate students will gravitate to as the infrastructure gains credibility over time.

The current, retrospectively scanned collection of Computer Science Technical Reports held on ncstrl.mit.edu represents what customers perceive as an interesting and complete enough collection to use. As with the theses.mit.edu collection, the preference is for new works not retrospective acquisitions. Also as with the theses collection, there is the potential to adopt this collection in the infrastructure as a credibility enhancing step.

There are active communities in the Artificial Intelligence Laboratory and the MIT Laboratory for Computer Science who might be appropriate lead users of this system. The deciding factor here would be in identifying a cohesive group of people with similar enough work flow to produce an acceptably complete collection with reasonable efforts.

It is to be expected that there will two classes of adopters of the Digital Archive infrastructure from the CSTR community:

• Early adopters wanting to try out our system.
• Very late adopters wedded to their unique workflows who will wait until the system proves itself very stable and credible.

The preprint service held at www.arxiv.org offers some interesting possibilities. Although the collection is not strictly MIT content, this can turn into a credibility builder because patrons often perceive an MIT-only collection as incomplete.

At one time there were rumors that the maintainers of this system were looking to hand it off to others to maintain. At the present time maintenance has shifted from LANL to the California Digital Library.

So our involvement with arXiv should be to make sure that our metadata can be sent to them and thereby facilitate searches that refer to our collections.

A search on keyword "publications" (Within MIT) on the web.mit.edu web site turns up many departments labs and centers that have elected to place some aspect of their publications on the web. Perhaps a well-documented site of record would attract these sites to a central repository.

Services observed on the various sites:

• Pretty front page introducing the particular department, lab or center.
• Search for publication. (Pretty basic, Author, Title, Keyword, sometimes within a controlled vocabulary of specific terms.)
• Display of item's bibliographic metadata (often including Abstract)
• Ability to order the item from the department, lab, or center (which might involve credit card processing).
• Display of some or all of the actual publication. (In various formats, PDF seems popular.

Perhaps the way to proceed here is to pick a single department, lab or center that uses some or all of the observed services, and build a host site that parallels (with an eye to replacing) their site.

Over a year ago MIT Libraries was approached by Forbes Dewey, professor of Mechanical Engineering and Bioengineering, asking us to create and manage a database of biomedical images for use as a teaching tool. His group has developed a prototype database of 17 MRI images as a demonstration project. http://miro.mit.edu/servlets/DcmDemo

The demo seems to be sponsored by the International Consortium for Medical Imaging Technology (ICMIT) which includes non-MIT members. (http://icmit.mit.edu/projects.html) Forbes envisions a 5,000 image database (50 GB) containing images produced by several labs in several departments. HST and Brain and Cognitive Sciences would be major contributors, in collaboration with several Engineering Departments.

Advantages: experience with image files and different types of metadata, the group has already experimented with a prototype implementation, and the user group is very eager. Additionally, this is an area that is receiving a great deal of external funding which might factor into the business plans of our service.

Contact Info