DLI Spring 1997 Partners Workshop
April 3-4, 1997
Minutes
ACTION ITEMS:
*DLI will improve functionality and look
and feel of Web client.
*DLI will identify a subset of publisher partners to test the Web
client.
*Publisher partners can help encourage CIC members to utilize the
testbed when it is deployed.
*DLI partners need a project plan from the DLI for development by
August 1997.
*Need to plan for continuation of testbed after end of project.
*Bill Mischo will provide plan of action for deployment of custom
client to UIUC Civil Engineering Dept./ASCE will help with
publicity.
*The DLI is encouraged to present at the ‘98 WWW conference
in Brisbane.
DLI FINAL PLAN (May 1997 - August 1998)
AIP will set up first repository in summer ‘97, cloning DLI testbed at their expense.
The custom client will be deployed on 6 workstations in the UIUC Civil Engineering department lab, available to both undergraduates and graduate students. Bill Mischo will provide a plan of action for the deployment and ASCE will help with the publicity.
Multiview Client (by Eric Johnson) will work with real sources (SGML Testbed and Concept Spaces for SGML plus thesauri for INSPEC) in selected labs and for selected users (careful usability tests with trained users)
DLI team will add functionality and improve look and feel. Multiview client and custom client may be ported to the Web.
TECHNOLOGY TRANSFER
Bob Wedgeworth, University Librarian, discussed the issue of technology transfer. In particular, he has been encouraged by some of the DLI publishing partners to create a commercial entity to transfer technology developed by the DLI to our publishing partners. The university is currently looking at setting up a commercial entity, hopefully by Fall '97. They have been actively setting up the premises of the company (to be called IDLT), with a preference for a start up company rather than licensing to the outside. The DLI sees this as collectively advancing our research. 1 million dollars has been slated initially for first couple of years before there are products on the market, with the company headed by Wedgeworth, who will act as president to get the company started. After start-up, a CEO will be recruited immediately who knows the database/publishing industry and software technology. Bruce Schatz will still head the research and development unit and continue to transfer that technology to the company for stabilizing and packaging of products. Investors have already been found for the initial startup phase but once startup is completed, DLI partners will be invited to have the opportunity to invest. The University is required to have an external assessment, which is currently being conducted. When the University makes its decision, the company will be opened up. Mr. Wedgeworth/DLI is investing in the future and would feel comfortable with our publishing partners to be involved as well.
When the project ends in August ‘98:
1. Q: What happens to DLI people the day after the project ends?
A: DLI personnel on soft money will move to other related government funded projects, assuming funding. Please note that there are currently only 3 library personnel on soft money at this time. Key library employees such as Bill Mischo and Tim Cole are supported by the UIUC Library.
2. Q. What happens to the DLI collection?
A. As agreed contractually in the beginning of the project, the collection goes back to the partners. However, Robert Wedgeworth, University Librarian, attested that the testbed will in no way be "dumped" at the end of the project. The library will continue to run the testbed on an operational basis for at least 6 months, at which time we will discuss with our partners what should/can be done with the collection. Mr. Wedgeworth is committed to a long-term relationship with our publishing partners for the purposes of collection development and expansion of full-text for the UIUC Library. In conjunction to IDLT, the library is planning on starting an industrial partners program to continue supplying materials and to do some intensive focus groups. The library is very interested in having a standard service that will do more focus group studies.
3. Q. Would the DLI consider a 6 month extension to the CIC distribution?
A. NSF would be unlikely to support this. Looks likely there will be another DLI but topics won’t be the same, even if we are funded again.
4. Q. Will the DLI be adding any more materials to the testbed?
A. AIAA Journal will be added and will be the last new journal title added.
5. Q. What about technology transfer to the publishers?
A. Most of the scripts are in pretty good shape with some of the processing in Visual Basic and C++ on Unix needing to be cleaned up somewhat. The DLI is willing to give our programs to our partners and we welcome our partners to schedule single day trips to Urbana for intensive one-on-one time with the testbed team.
SGML DISPLAY/DLI WEB CLIENT
Concern was expressed over the minimum functionality displayed by the Web client which will be deployed by the DLI later this spring. It was argued that this factor, combined with hesitancy expressed by CIC library directors to put up a resource that is temporary and has a difficult-to-use client, will render the research data of the fourth year of the project useless due to low numbers of users. DLI members responded by promising to make a concerted effort to improve the functionality of the Web client, including the look and feel of the client. DLI would also like to identify a subset of publisher partners to test the Web client. Publisher partners can help encourage CIC members to utilize the testbed when it is deployed.
The issue of HTML display was brought up. The DLI will continue to follow progress made by commercial applications but does not have plans at this time to spend resources on this topic. Of particular interest at this time is the move towards XML (see http://www.mcs.net/~dken/xml.htm, http://www.textuality.com/xml/ http://www.w3.org/pub/WWW/MarkUp/SGML/Activity for more information on this topic). In particular, the following was brought up:
Netscape and Microsoft may have XML-aware clients in 1998. Some advantages of XML, which allows for documents and applications over the Web with more sophisticated viewing capabilities include:
*Ability to extend HTML by adding tags and
attributes without moving toward a proprietary coding scheme
*Ability to create tags which can be used to facilitate searches
based on document structure and data elements as well as to
create a screen display (Document structural information intact)
*DTD not necessary
The DLI is encouraged to present at the ‘98 WWW conference in Brisbane.
PARTNER REQUESTS
*AIP needs project plan for development by
August, not by NSF’s timeline to plan for next fiscal year.
*Need to plan for continuation of testbed after end of project.
SOCIOLOGY REPORT
The sociology team reported that were not surprised that more computer literate users (CS people) don't exploit the functionality of the SGML/testbed. They believe that factors contributing to low usage of the testbed include:
-the system is currently available in only
a few places
-the system is buggy and goes down. Users should not be expected
to try system again once it has failed.
-"testbed" is not descriptive enough to encourage users
-there is not enough depth and breadth to the testbed
-users are confused in general by the variety and proliferation
of technology.
More usage is expected once the Web client is available and a publicity push is being planned.
SOFTQUAD UPDATE
In August of 1995, the DLI held a workshop for their partners, and during the meeting it was emphasized that SoftQuad's Panorama was a critical component to the success of the DLI project. The participating organizations noted that they still required additional capabilities in formatting mathematics. In particular, that there was a requirement for a single agreed upon SGML math fragment that all parties could work with to create and render math on screen. The sense was that a single SGML fragment could be treated like standard encoding. As a result of this workshop, the UIUC DLI held an SGML mathematics workshop in May of 1996 with invited experts from the publishers, mathematicians, software industry and the SGML community. Throughout the discussions at this workshop, SoftQuad indicated that it intended to continue to develop Panorama as the foremost browser on the market, but that there was not a compelling business case, at that time, to invest heavily in mathematics at the expense of other capabilities.
At this meeting SoftQuad was approached by Tim Ingoldsby (AIP) who made it clear that AIP was interested in making a financial commitment to SoftQuad for the further development of an SGML browser that could successfully render all of their mathematical documents. A meeting was then convened in New York with an attendance of the foremost scientific publishers. A business plan was then prepared by SoftQuad, with an additional distribution to five academic publishers. This plan proposed that each member partner contribute to the cost and expertise of a program to be developed by SoftQuad. This program was agreed upon by the member partners and is to be implemented by SoftQuad with an estimated completion date of an accepted product by the end of the year and a commercial product by mid to late 1998. Any publishers wishing to participate in this process can contact Pete Spence at SoftQuad (416) 239-4801.