On April 19, 1999, the following letter was mailed to members of the SGML Server Program. It alerted them to a significant development, a new licensing mechanism for the Open Text search engine, and a proposed reorganization of the SSP to take this new development into account.
Digital Library Production Service
300 Hatcher North
University of Michigan
Ann Arbor, MI 48109-1205
Dear SSP participant:
We are pleased to be able to share some exciting news with you about the future of the SGML Server Program (SSP). We would also like to take this opportunity to solicit your guidance as we reshape the SSP.
You may recall that the SSP was conceived to support the two goals. First, we wanted to leverage our local infrastructure to make text collections available in ways that publishers such as Chadwyck-Healey and Oxford University Press were unable to provide us. Second, we wanted to leverage our production activities to help you build local infrastructure at many of your institutions (e.g., so that you could make such collections available to your constituencies). We hope that the last three years of cooperation in these endeavors has been satisfactory to you. There are currently eighteen SSP institutions, including Michigan. An increasingly large number of institutions have chosen to focus on building local capabilities, and now only six institutions use the University of Michigan as their host service. We have been able to meet most or all needs in putting collections online, and now support nearly twenty different collections.
The Open Text search engine has been both a lynchpin of our enterprise and something of a conundrum for all of the institutions. For our purposes, the most functional version of the software was their version 5, abandoned as a product in 1994 as they moved to a more expensive document management strategy with the search engine bundled. New institutional subscribers without prior Open Text agreements were forced into expensive agreements to license rights to an inferior product. Institutions that invested in hardware such as the DEC Alpha and the RS/6000 found their purchases made obsolete by Open Text. The version we liked best (i.e., version 5) was no longer supported for those who had purchased it, and no longer available for those who had not. Non-academic non-profits were priced entirely out of the Open Text market.
In an extraordinary stroke of good luck, we have been able to negotiate an agreement with Open Text that should enrich the SSP and make our dealings with Open Text much easier. The University of Michigan now has a contract with Open Text that provides us with the version 5 source code as well as the right to license the search engine to other non-profit and not-for-profit institutions. We are bound by predictable clauses to acknowledge Open Text's intellectual property and we must pay Open Text a fee for each sale or installation of the software, but we are otherwise free to add functionality and sell the search engine.
The array of opportunities this new arrangement provides us is exciting. For example, we hope to undertake development efforts that will bring XML support into the search engine. We will also be able to make the search engine available to non-profits such as university presses and associations. And, as the distributor of the search engine, we will be able to provide formal support for the search engine.
We have engaged in local discussions about how to best support these new sets of opportunities and address issues that have arisen in the SSP over the intervening three years. As you may guess, participating institutions are using (or have licensed) several different versions of the software, increasing the support burden at Michigan. The distinction between and the options offered by the "remote" service (i.e., using Michigan as a host service) and "local" service (i.e., installing the software on your local systems) has continued to confound many. We believe that we can undertake a strategy that will address these issues.
First, we would like to propose distinct programs to meet the needs of "local" and "remote" customers.
Second, we would like to propose a modest program of development of the search engine, geared specifically to SSP applications. In addition to XML support, we hope to develop support for "regular expression" searching (e.g., internal wild cards) and more document structure awareness (e.g., parent, child, and sibling relationships). We will be able to perform a thorough Y2K assessment of the software, using a suite of formal assessment tools and an examination of the code. The features of this development will be tied to those sorts of digital library applications that we have developed in the SSP and locally as we have used the Open Text search engine.
We think this is an excellent time to pause and reassess the effectiveness and directions of the SGML Server Program, and hope that you will take the opportunity to share your thoughts with us. This also calls for us to review our pricing models to reflect the new programs of support and development. Our own schedule makes it impossible for us to undertake development work on the search engine in the next few months, but we have scheduled the beginning of that process for late in this calendar year. (If you do not currently license the Open Text search engine and would like to discuss early release, please contact me.) We will work with the University of Michigan's legal counsel to fashion a satisfactory contract for the search engine, as well as for the new SSP. We will be contacting you in the next few months, particularly to talk to you about new models for pricing. We hope to have your thoughts to guide us as we begin to undertake the process of articulating the new program.
Head, Digital Library Production Services