University of Waterloo
                    Electronic Data Service Task Group
                      Minutes of Meeting May 18, 1994


Present:  Boris Bruder
          Albert Kemp
          Sue Moskal
          Doug Morton (Secretary)
          Bill Oldfield (Chair)
          Richard Pinnell
          Shabiran Rahman
          Carol Vogt
Guests:   Ron lambert
          Keith Warriner
CC:       Bruce MacNeil
          Boris Bruder
          Mike Ridley

1.   Minutes of Previous Meeting
     No changes

2.   Business Arising
     To be covered below.

3.   Additional Agenda
     Sue Moskal - layout of the EPAS Gopher

4.   Mission Statement
     The mission statement for the EDS Task Group was reviewed as
     background for the guests.

5.   Ron lambert & Keith Warriner - User Requirements
     Our two guests discussed how they have used data in the past
     and how they would like to see access to data provided in
     the future;  they also provided some insight into just what
     the files we have are and how they would be used.

     Unless the user has a SAS or SPSS masterfile of the data
     (unlikely to be distributed because these will look
     different for each platform) the user has to create their
     own from the data files and the SAS or SPSS "stub" files. 
     Because of the problem of the differences between user
     platforms, it is unlikely that we could provide such
     masterfiles ourselves.

     A researcher looking into a topic tends to ask about for
     information on where the data might be; it often happens
     that the exact dat are not available but inferences can be
     drawn from other available related data. Our guests
     preferred to come to a central location to check printed
     codebooks to see which aspects of the data tables they wish
     to retrieve and work with.  (A good codebook will have
     information about the sample population, size, response
     rate, sample selection criteria, and other factors
     indicating the quality of the survey as well as information
     about individual questions).  

     Codebooks can be several hundred pages long and it was
     suggested that it is impractical to have each user print out
     a personal copy. Our guests also discussed the practicality
     and difficulties of making and using machine readable
     codebooks. They agreed that searching codebooks might be
     useful and that access from their offices would be nice but
     the vocabulary shifts from survey to survey, making
     effective searching difficult.  Serendipity is a definite
     factor in using codebooks and this may not change easily. 
     There was also concern that computer neophytes are already
     facing a steep learning curve to retrieve and use the data,
     let alone retrieving and using the manuals electronically.

     Asked if students had access to computers that could connect
     to files mounted on other computers on campus (such as
     Watserv2 or Library) our guests responded that it was
     possible. Students now connect to UWInfo and WATCAT, for
     example.

     After some discussion the following model sounded feasible:
     The EDS would mount information on a menu structure (such as
     gopher) with information about titles, files, codebooks,
     etc...  There would be pointers to programs such as the
     EPAS software.  The files we have been calling the "stub"
     files (the ones to be combined with the data to become a
     system file) would be modified to include an NFS pointer to
     the data on "library.uwaterloo.ca" and made available for a
     user to download and use.  (it is apparently possible to
     restrict users to UW-only with the NFS software.)

     Ken Warriner stressed the need for a human presence in the
     process.

     The guests were thanked for their time and thought. They
     were able to answer many questions from a user's prespective
     with background and authority few in the group have.

Next Meeting: May 25, 1:30 pm LIB 428