We learned along the way
- Generating data dictionaries. We started dealing with the
issues surrounding the publication of a large body of data element
documentation when the IRM was first formed in 1993. A lot of
the information we were gathering about the Payroll/Benefits system
at the time was free-form text that was expected to change over
time. We collected the information in SAS datasets and then generated
a 600 page data dictionary using WordPerfect. Problems included:
getting the information in the first place, translating the jargon,
the SAS limit on variable length (200 characters), the difficulty
of page formatting in SAS, the temptation to edit the data once
it had been formatted in WordPerfect.
- Gopher. Our first attempt at electronic publishing was a gopher
- SURDS. Experience with semi-structured text. First draft
- SIS. First try with search screens.
- Web Conference in June 1995 suggested many more possibilities
and was the occasion for writing the very first draft of this
Web Management Manual.
- Usability tests became necessary when we had several draft search
menus for the SIS data dictionary application. Having a systematic
approach to looking at our Web site from a users' perspective
encouraged us to be much more experimental.
- Data warehouse lite. delivering actual data, not just metadata.
- Learning Perl. All along, since beginning of SIS
- Writing this manual. First as outline for session in June
- Lots of issues raised by our January 1996 Conference on "Organizing WWW Information"
Resources to support a large-scale Web site
- Need to have the information to publish. Although the
cost of generating the information may already have been expended,
this is probably the big ticket item.
- Special HTML editor
- An overall resource design.