Robert Cerny mentioned looking for the historical antecedents of topic maps the other day. Subject identity and identification lie at the core of meaning, so it is no surprise that issues we associate with topic maps are woven into the fabric of history.
Collation, for example, appears in the history of Social Security, a program that was part of the New Deal in the United States.
I didn’t find this on my own (see below) but the ability to mechanically match records in two different sets, to see if they were related to each other, did not exist prior to Social Security. Jack Futterman recounts the invention of the collator:
…The machinery that we had to do the job in those days to keep records did not exist. I should rephrase that. There was no machinery that really could do the social security job before the Social Security organization came into existence….the collation, the ability to take two sets of records and do a matching to see whether they were appropriate or the same and related to one another and then to make, in effect, decisions as to whether to interfile one or to reject it was a facility that did not exist in the equipment up to that time….Oral Interview with Jack S. Futterman, January 23, 1974
Name collation used the Soundex algorithm, as Futterman recalls in other oral history remarks.
Name collation has evolved since then, see A comparison of personal name matching: Techniques and practical issues.
Topic mappers interested in expanding their toolkits for building topic maps or developing additional merging algorithms will find name collation a fruitful area for exploration.
(I first saw the information about the Social Security Administration while reading Edwin Black’s IBM and the Holocaust. Recommend reading the online excerpts with an eye towards how topic maps could have assisted with his and similar projects. There will be posts on aspects of this book in the future, please watch for them and consider posting your thoughts about using topic maps in such projects.)
In TempleScript I use “name collation” implicitly: As in the old days (XTM, CTM) topics in TempleScript can be addressed by some identifier. But I also allow authors to use “queries”:
The person ~”Patrick D.”
is-author-of = http://tm.durusau.net/?p=35 .
The ~ in front of a name allows soft text matching, soundex just being one possible regime.
The “The person” in front of the name just sharpens the query. That is to avoid confusion.
Comment by Robert Barta — March 14, 2010 @ 3:09 am