This paper describes three experiments whose purpose is to address issues concerning the applied use of spatiality as a means of representing, organizing, and retrieving information.
At that time, 1986, the dominant mode of object reference on computing system was symbolic: data files, commands, programs, etc. are initially labeled and sub sequentially referred to by name.
The experiments are intended as a base on which to compare the accuracy of location based, named based, and location+name+based object references under conditions of an expanding database.
The subjects have to file short newspaper article printed on paper. There is
no computer use. There are three different conditions. uses stick a label with a
number identifying the article on a separate sheet. The way in which the user
can stick the label depends by the conditions.
Location: place the label whenever hey wanted.
Name: added a symbolic name to the id. They have to put the labels in the place next to the ID numbers that were put on the sheet by the researchers in a random way.
Name+Location: subjects selected both name and location.
Name+Location separated: use two different labels for name and for ID and placed them on a separate piece of sheet.
After that subjects were asked to retrieve the papers after blocks of ten papers.
They tested the subjects immediately and after 30 days on 30 articles. The best performance was name+location and the worst location only.
The authors argue for using name only since te results are comparable to name+location.
I don't think that this experiment can actually say a lot about the issue of filing in a computer in a location v.s. name. The situations are so artificial and so different from a real environment. Location doesn't rely on a number placed on the space. Also, space memory is reinforced with time and this time reinforcement is missing in these experiments.
Experiment 2 replaced the bank filling sheet of experiment 1 with an identically dimensioned, but enhanced sheet containing representations of various pieces of office furniture.
Experiment 3 placed subjects in an actual three dimensional office setting.
Experiment 2 performed better than 3, but worse that name-only in experiment 1.
They suggests that inclusion of landmarks in the space clutter it and make it more difficult to retrieve elements.
They lucubrate why one methods may be better than another or limitation of experiments.
Disclaimer: these summaries are made to help me remember the content and the main ideas of the paper. Since I am interested in certain aspects, I may leave out others