Duke ITAC - February 1, 1996 Minutes

ITAC Minutes 1-Feb-96

Minutes from the 2/1/96 ITAC meeting

recorded rather frantically by John Board


Mike Pickett (MP) discussed a possible site license from Sybase. OIT and MCIS met with Sybase; we have been offered a deep discount, but even discounted the annual fee involved is non trivial. Prior to making any decision, similar arrangements are being explored with Oracle and others.

MP also reported (actually at a later point in the meeting) that the preferred site in Wilson House for the new East Campus computer cluster recently endorsed by ITAC had been assigned instead to the infirmary; alternate sites are being investigated. Cost estimates for the entire project, including the dormitory printing trial, have been reduced significantly from the first cut, to about $50-60K.

Student Computing Subcommittee

Tom DiPrete (TDP) reported on progress to date. Two meetings have been held; the question as to whether to either recommend or require student computer purchase is occupying the subcommittee. Buying computers for students is being explored as an option to requiring students to buy their own computers; the financial implications are different in each case. Though no final conclusion has been reached, the current sense of the committee leans to recommending, but not requiring, that students bring computers of some minimal standard with them. The cost of requiring ownership may be prohibitive; such a sum could potentially be put to other uses which would have greater impact on the student computing environment and undergraduate education.

To aid their deliberations, especially with respect to the "best use of university resources" question, the subcommittee is assembling a list of actual and potential uses of IT in undergraduate education, and the infrastructure that would be required to enable the unrealized uses.


Betty le Compagnon (BLC) asked about a more systematic survey of IT uses in education.

Richard Palmer (RP) and John Board (JAB) both noted that they had been surveyed by someone about precisely that issue within the last few months.

TDP agreed a more systematic survey would be a good idea.

Strategic Planning Subcommittee

JAB reported on the draft Interim Report of the strategic planning group, which has been available off the OIT web page (http://www.oit.duke.edu/itac). He summarized the twelve short-term recommendations and four discussion items described in the report. (He, as secretary, will not reproduce them all here but refers the gentle reader to the web page instead.)


David Jamieson-Drake (DJD) discussed section 4.1, on off-campus modem access, noting that the current public-access modem pool was not really designed for people working for long stretches of time from home, and yet many faculty, staff, and students (mostly graduate and professional) do expect to be able to use the facility that way. He hoped other alternate access schemes including ISDN and broadband cable TV modems would be considered; conventional analog modems are one part of a spectrum of off-campus access issues.

BLC noted that she and Landen Bain from MCIS had discussed better off-campus access as a good joint OIT/MCIS project.

DJD expects that some departments would be willing to put some of their own money into a good solution for higher speed remote access.

JAB noted that the Bell South commercial modem service might be part of the solution.

John Brehm asked for the rationale behind the priority assigned to each item.

JAB responded that, though the mainframe migration issue can probably be described as first among equals and is listed first intentionally, no relative ordering was intended among the remaining items. The subcommittee felt that all items listed are important, all are feasible to implement before next fall, and many have, in fact, already been partly or fully implemented by OIT and others.

Anne Beckwith asked, relative to sections 3.9 -- 3.11 which concern information sharing, that ITAC consider how to make better use of mailing lists to distribute information.

DJD asked for a placeholder for recommendations from the data subcommittee.

JAB agreed the data and student computing committees should be mentioned more explicitly, and noted that the final report of the strategic planning committee would need to incorporate major contributions from the other two subcommittees.

Robert Wolpert (RLW) noted that the rewiring project was not mentioned and was conspicuous in its absence.

RLW and RP noted that, last spring, they had understood that Deans and other administrators were going to be actively involved in making priority decisions in the rewiring effort. This was not happening in practice; rather, a prescribed solution was being offered by OIT to each building considered so far.

Melissa Mills (MM) commented that Duke faced a potential future disaster by funding the electronics portion of the rewiring project out of a non-renewable source. This was just another example of section 4.4, the patchwork funding problem of IT at Duke.

Paul Harrod noted that section 3.12, the "year 2000" problem, accelerated the urgency of section 3.1, mainframe migration, as the impending failure of many administrative systems provides motivation to go ahead and move them off the mainframe rather than go to the trouble and expense of repairing them for one or two years' additional use.

DJD reported that the year 2000 problem is already here and affecting administration applications. Five year projections already cross the troublesome line, and the new undergraduate students admitted this spring will belong to the class of 2000.

MM noted that mainframe migration is really data migration.

TDP noted that the section on patchwork funding of IT at Duke (4.4) is different from all the others.

JAB commented that the subcommittee felt the funding mechanism for IT did have strategic implications: funding mechanics can act as incentives and disincentives for globally cost-effective IT, depending on how they work; whether or not a service is funded centrally or operated on a charge-back basis does affect the quality, usefulness, and economy of the service.

MM hoped the patchwork funding discussion could be expanded to include the patchwork of arrangements made by departments to fund local IT: some departments fund their own IT, others rely largely or wholly on OIT. The patchwork extend across units and not just within OIT.

RLW noted, in closing discussion, that Duke's historic culture was for individuals and units to use the mode of computing cheapest to them and not necessarily cheapest to the university as a whole; he thought funding was a strategic issue and should remain in the report.

RLW expects that ITAC will receive and approve a final version of the report at its next meeting, at which time it will be widely advertised.

Re-engineering of Procurement

Annette Foster (AF) presented an overview of the procurement re-engineering process which is now well underway. Copies of her slides were distributed at the meeting and can (probably) be obtained from OIT or Annette.

From a rich and detailed presentation the humble secretary attempts to distill a few key points:

  • Duke makes purchases of over $600M/year from 30,000 vendors
  • There are currently 12 distinct ways to buy something, involving 315 different steps and 8 forms
  • Current purchasing is labor intensive, slow, and yet unable to leverage Duke's buying power to get good deals on things we collectively buy lots of
  • The traditional PO accounts for only 14% of purchases; most PO's are for low dollar amounts; each PO costs $37 centrally (not counting departmental costs) to process
  • Most purchases bypass PO's via standing orders or one of the other 10 schemes
  • The re-engineering team has focused first on a new process, with only two ways to make purchases, and then on technology to implement the redefined process
  • Electronic purchasing, with electronic approvals if necessary as one mechanism; a Duke Procurement Card (special charge card) as the other
  • An electronic catalog (perhaps on-line, perhaps CD-ROM) would be created for commonly purchased items; Duke buyers would negotiate best possible deals on these items so end users would not have to
  • Design team estimates savings of $10-30M per year from buying power leverage
  • Up-front conversion cost estimated between $2-4M
  • Detailed design phase now about to start


Jim Dronsfield asked about IRs

AF: team is working on electronic IRs, but this is difficult. IRs were not included in the external purchasing numbers given in her report.

RLW asked how the short-term plans of this project affect other strategic IT planning efforts - who is assessing the interactions with other applications, etc.? What are the overlapping concerns of this project and of others?

AF: All vendors considered as finalists when a detailed design is complete will need to be screened for their interface capabilities; they must be shown to be compatible with other administrative tools at Duke.

TDP asked how he would access the electronic purchasing system as a user.

AF: directly from his desktop system. PCs, MACs, and Unix boxes are all required to be supported.

MM asked about a period of concurrent operation of the old and new systems while the new system is shaken down.

Caroline Nisbet, as head of the social design team for the project (AF heads the technical design team), said the schedule for phasing in the new and out the old was one of their responsibilities that was still being worked on.

MM asked about purchasing something not in the electronic catalog, or perhaps a different brand or grade of something that is in the catalog.

AF: The electronic order forms have a free text area where users can put anything they want; users are not restricted to buying just items in the catalog.

BLC noted that the intent was to go to the Board of Trustees for funding of the project prior to a final selection of a vendor. This is novel, but EVP Trask is behind this project so the Board might consider an accelerated schedule.

RLW wondered how the need to migrate off the mainframe was influencing vendor selection.

AF closed discussion by noting that purchasing is a university-wide need and service, and that the needs of the medical center (the source of the majority of Duke purchases) for an improved and more efficient purchasing system were unquestionably different from, for instance, some of the needs of individual faculty making small one-time purchases and the new system and process must accomodate both of these sets of needs well.