Museums and the Web 1999

Best of the Web

Archives & Museum Informatics
2008 Murray Ave.,
Suite D
Pittsburgh, PA
15217 USA

Join our Mailing List.

Published: March 1999.


Database Publishing Without Databases

Martin Warnke, Paul F. Siegert, and Carmen Wedemeyer, Universität Lüneburg, Germany

Roll On/Roll Off

»Digital data lasts forever -- or five years.
Whatever comes first.« (Jeff Rothenberg)

A big problem for anybody building up digital archives is the ephemeral nature of operating systems and of authoring tools. Having ourselves suffered from these circumstances, we decided to exclusively code anything that should last for some time in simple ASCII format and in HTML. We particularly had to convert an existing digital archive that extensively made use of hypermedia techniques from the proprietary HyperCard format to HTML and by doing so we worked out tools that were especially useful doing just this. A more detailed description of these tools follows.

The methodology we adopted was like this:

  1. Any structured information is stored in some database system. This helps keeping the material intact and consistent. Here only contents matters, no layout or design issue is tackled at this stage of the project.

  2. At an appropriate stage of the collection of material the data is exported from the database to a tab/return format, thus making it irrelevant which database system has been used.

  3. This tab/return structure serves as input to a program that converts all the fields and records of the former database to a set of individual HTML pages.

  4. This is done by means of an HTML template: some HTML page, coded by hand or by use of an HTML editor, that uses wildcards as representations of database field contents. Any HTML element could be used for this purpose: character data, href or img adresses, image map coordinates, title tags or even HTML file namesincluding pathnames.

  5. Changes to the layout or the functionality -- that is: changes to the HTML template -- could immediately be applied to the set of HTML documents, thus keeping apart content from design and functionality.

  6. The set of HTML documents could now freely be altered, annotated, and also be published an a server or locally on CD ROMs or similar media using just standard HTML technology and any web browser. Neither database run time modules nor any specific authoring or publication tool or CGI scripting depending on operating systems are needed any more.
  Fig. 1 shows this methodology in a more schematic way:

The database roll off program is the kernel of our collection of tools, although other important steps during the course of publishing digital archives are supported as well (see below).

The tools are written in Java, all the intermediate steps and the final data format are completely based on the ASCII standard, so the whole process is totally platform independant.

The Way Things Work

Let us turn to an application of the database roll off tool.

The most simple application of the roll off tool is a 1:1 match of the former database records to HTML documents. Now everything between two returns within the tab/return database export is inserted at the places indicated by the template document. This is accomplished by using a field wildcard for the HTML document filename. The filename may include the path, this means that hierarchies of directories could be generated right off the database.

Fig. 2

In terms of ASCII files and HTML templates this looks like the following:

The case opposite to the latter is an n:1 match between data and template. This is especially useful for the generation of lists or tables.

Fig. 3

Again we show this in terms of database export and HTML template:

As always, the most interesting case is a mixture of the „pure" ones above: let's call it the n:m case. Here, filenames appear redundantly within the sorted database export. Only if a new filename appears, a new file is created by the database roll off tool. Otherwise a loop is being processed within the HTML document. The program suggests possible scopes for the loop, based on the syntax of HTML: only nestable HTML structures, e.g. lists within a document or sequences of coordinates within a clickable map, could be chosen by the user.

Fig. 4

As an example for this an excerpt of the digital archive about the artistic estate mentioned above is shown: HTML pages containing clickable maps:

Terms like "0_A" or "0_B" in the example document are inventory numbers, x1 through y2 indicate the coordinates of the clickable map shapes, here being simple rectangles.

A second example shows the making of different picture catalogues:


A special application to the n:m case are index lists, where nested loops appear. The address of an index entry has to be repeated within the list of entries, which is a loop itself. Sometime we would like to have individual files for every letter of the alphabet, which means another repetition.


Needles in haystacks and other nuisances

Our collection of publishing tools also contains indexing engines, since -- strangely enough -- the standard HTML editors do not offer this basic and extremely useful philological technique. The main function of this text tool is to generate an index of HTML documents within directories. The result a sorted word list with links to the paragraphs within the documents where the word could be found. Stop words could be indicated for omission from the index, inflection lists are used to trace back the variants of a word to their syntactical roots. Besides this, the tool does some „dirty work" as conversions between ASCII and HTML, splitting up database files into seperate ASCII files and some more useful tasks. It has been used extensively for the diverse indexes of the „Umarmungen" project.

What is the use of all of this?

By use of the tools described above -- in partucular the database roll off tool -- we have been able to generate a highly complex multimedia system like the „Umarmungen" project (a sample could be obtained from the authors) in HTML, consisting of 17500 individual files, from well structured databases on CD ROM without database front end programs. This can be done without any programming skills by just the use of any database, an HTML template file generated by hand or with an HTML editor, and the tools we built that run in any Java virtual machine environment. This proved that HTML is an appropriate format for sustainable multimedia projects whithin the field of scholarly work, even for those not willing to program themselves or to restrict the structure of the documents to a rigid database format.

Researchers interested in these tools may obtain them without fee for their own non profit purposes. Please contact the authors if you should be interested in using them.


The former proprietary HyperCard version was demonstrated during the ICHIM '95 conference in San Diego (Wedemeyer, C, and Warnke, M.: Documentation and Edition of the Artistic Estate of Anna Oppermann with Computer Aided Methods - the Prototype for the Ensemble „Umarmungen, Unerklärliches und eine Gedichtzeile von R.M.R." („Embraces, Inexplicables, and a Line from R.M.R.")) and is now available as HTML version in „Wedemeyer, C. (1998). Umarmungen…/Embraces - Anna Oppermann's Ensemble "Umarmungen, Unerklärliches und eine Gedichtzeile von R.M.R.". Ein hypermediales Bild-Text-Archiv zu Ensemble und Werk/A hypermedia picture text archive on ensemble and work. In Englich and in German language. Frankfurt/Main, Basel, Stroemfeld Verlag. The CD-ROM contained in the book could be obtained for free by sending an email to A description of the project and the hypermedia catalogue of the works and of all texts of Anna Oppermann can be found under