Go to previous article
Go to next article
Return to 2000 Table of Contents
3702 AD Zeist
Phone: +31 30 6982346
The aim of this study was to specify a testmodule (with visual function tests), of which the results can be used to adjust sotware to individual visual perception. The test results of the module will be stored in a user-profile. Software can then use this information to automatically adjust its user-interface and content to the individual user-profile, thus making it more easily possible for partially sighted to work with the same software as their sighted counterparts. The module (including a spectacular fine-tuning possibility) as well as the research results will be shown on the conference.
| Top |
This paper is based on a study commissioned by the Bartimeus Foundation (E.Velleman e.a., 2000) in which we examine the possibility to automatically user-adjust the representation of the graphical user interface (GUI) and the content of a digital atlas and a multimedia program to the individual needs of partially sighted users. This should be done by means of visual function tests. The results of the tests are saved in a user-profile. This user-profile can then be used as an input to automatically adjust the User-Interface, the maps and the contents to the individual visual abilities of a partially sighted user.
In this study we are working together with Wolters Noordhoff, specialised in (paper) school atlasses to produce a digital schoolatlas that will also be accessible to the partially sighted. The atlas will interact with the test modules. Only if the data are out of range, the user will have to revert to screenreaders, screen enlargers, other hardware etc. Sometimes a good combination is possible.
The user-profile that is a result of the tests can be saved to be used in other programs on the same computer.
First we looked at the definition of partially sighted, then we studied the variables that are neccessary to fit a digital mapping program to the visual (perceptive) needs of the partially sighted. On that basis tests have been specified. In this paper one of the tests is looked at in more detail. Then we look at the interactive module making it possible to interactively fine tune the mapcontent of a digital atlas after the tests.
| Top |
Blind and/or partially sighted, definition
The World Health Organization defines blindness as a visual acuity of less than 0.05 in the better eye with the best possible correction. Partially sighted are described as having a visual acuity of less than 0.3, but equal to or better than 0.05 in the better eye, or a visual field of less than 20 degrees, with the best possible correction.
Already Coolenbrander said that more people are blinded by definition than by any other cause (Coolenbrander, 1996). Therefore we should see the above classification as a formality because it is impossible to define a sharp line between blind, partially sighted and sighted people. By minimizing the viewing distance we can perceive smaller objects. This means that people with a visual acuity of less than 0.05 can even read a newspaper. We don’t like to call these people blind.
Maybe more important is the visual efficiency. A combination of visual memory, perception, cognitive and motivational factors. Because of the different mix of this visual efficiency there can be great differences between partially sighted and their visual possibilities. This confirms the need of individualisation when looking for solutions to problems for partially sighted. It also confirmes the impossibility to define a one-for-all solution. When looking at solutions for this large heterogeneous group we therefore prefer using ranges instead of presenting averages or frequencies. Offering ranges enables software companies to know what are minimal (and maximal) requirements for the settings-possibilities of their software. They can then offer accessibility to the extent that it is functional.
Cartographers and graphic designers are often worried about what a change in fontsize or contrast will do to their artwork. These questions can be answered beforehand by offering ranges for fontsize and contrast (Velleman, 1999). Sometimes the possibilities for a user can be limited in view of usability (large font covers the rest of the map).
| Top |
Aim of the current research
The aim of the research is to identify perception problems of partially sighted with GUI’s and digital maps. And on that basis specify and compute tests of which the results (a saved user-profile) can be used by programs, on the same computer, to adjust to individual perception and make programs more accessible to partially sighted users.
Because of the innovative character of the mapping software the testing and reformatting of the programs and the subsequent user-profiles takes place during the iterations of the programming process. During this process many adjustments can be and have been made to the mapping software making it more accessible to the partially sighted (Velleman, 1999). Because of the complexity of these tests and the interactivity of the variables tested and the individual mix of the person involved the tests will not always generate a 100% result. But it will in any case speed up the proces of adapting the settings of the GUI and the content of the program to a users needs. The current tests offer the possibility of fine-tuning after all the variables have been set.
| Top |
Visual functions, what should we test
It is the combination of physiological (im)possibilities and visual efficiency that defines what we can perceive with our eyes. This can sometimes even be enhanced by visual training. In general partially sighted experience problems with visual acuity, visual field, color vision, dark adaptation, contrast sensitivity and fixation. Additional problems arise for certain groups if there is too much visual information at the same time, if the screenoverview has to be built in memory (mental map), if users have difficulties scanning the screen.
In view of the difficulties of partially sighted named above, and the goal of the tests, namely to adjust user interface and content to a user-profile, we defined visual functions that would give information about the possibilities of perception. The data should be usable for the purpose. First the functions were tested the traditional way (on paper). This resulted in four visual function tests for computerscreen:
These tests were subsequently developed together with the Utrecht University. In addition to the above tests there are other functions that should be adjustable:
| Top |
The tests were made in cooperation with the Bartimeus Foundation. A literature study revealed that only very little research has been carried out on accessibility of GUI and screencontent (like fontsize and contrast) for partially sighted untill now. This is probably due to the problem of defining the group as a whole. We resorted to using ranges thus confirming the need for individualisation when looking for solutions.
The tests were done in a special office under specified conditions with calibrated materials (screen, computer, illumination etc). The screen was a 17 inch 1280x1024 screen.
A total of 30 partially sighted students in the age of 12 to 19 were tested after which the specifications were finalized.
The tests should meet the following conditions:
In total four tests were envisaged of which currently three have been finished. One of the tests will be summarized in this paper. After that we will examine the fine tuning module for the atlas program.
This program uses the developed tests to generate a map according to the user-specifications. It then offers the possibility to fine tune the different variables (if the combination does not prove ideal). The profile can then be saved to disk. Programmers can use these data to make their software automatically accessible (this is allready possible in some programs but then without testing). This process speeds up the accessibility of programs enormously. For the complete testresults and the computertests one should contact the author.
First, tests were made were users could choose fontsize, contrast etc themselves by clicking on them. The results indicated that they then chose fontsizes (10 - 35mm) and contrast settings that were sometimes three times as large as they would need to comfortably perceive information on a screen (2,5 - 6,48 mm).
| Top |
Of the three tests that have been finished untill now, we will here review the acuity test. Other tests that have been specified are the contrast test and the line and symbol test (Velleman, Bartimeus, to be finished december 1999).
A glance legibility test is best used to specify the fontsize (Dewar, 1981). A letter is then viewed for a short period of time after which the user can type the letter himself. We concluded from literature and experience that the standard Arial font was best. It is a Helvetica type sans serif font found on most computers. We found that an interval time of 17/100 seconds was best for finding the comfortable fontsize. If the time is longer, people start to focus on the screen and end up seeing all the letters. At the end of the testtrails the program computes the comfortable fontsize (size that can be read for a long time without problems). This comfortable fontsize was found by comparing the fontsize found by the test with the fontsize results of a paper reading test. An algorithm was developed.
To justify the choice of the test and to tune the testing variables like time and size it is important that:
Therefore the results of the test where compared to the results of paper visual function testing (performed specialy for this project). This comparision led to the final specification and description of the test (van Marle & Dabrovski, Bartimeus, not yet printed).
The minimum fontsize perceived by most was 5,5 mm (printed equivalent: 21pt). Less than half of the testers scored 100% on a fontsize of 2,5 mm (printed equivalent: 11pt). Most sighted people scored 100% on a fontsize of 1,4 mm. NB.: In lower resolution (800x600) the fontsize needed was sometimes two times larger. This proves that if screens are used in maximum resolution, the legibility is enhanced.
The 95% range found in testing varied between 2,5 mm and 6,48 mm. These data could be used as an input for computing the comfortable fontsize.
| Top |
Example: fine tuning a digital atlas
Bartimeus is working together with Wolters Noordhoff in the Netherlands to specify and produce a digital atlas (Bos@tlas) accessible to the partially sighted by incorporating the above tests. The project offers a spectacular fine-tuning module to fine-tune the user-profile of the atlas by realtime adjustment of a cartographic map (after testing and adjustment to the profile).
Besides the testing- and the fine-tune module, enormous changes have been and are made in the software to make the GUI, content and maps accessible to the partially sighted without the use of extra software or hardware like screenreaders, screen enlargers etc. This process of making digital cartography accessible to the partially sighted is described in more detail in "maps and atlasses for the partially sighted" (Velleman, Bartimeus, to be finished december 1999). At the time this paper will be presented a newer version of this program that can spectacularly change maps realtime to fit your eyes will be shown as will be the research behind it.
| Top |
With the specified module it is possible to measure a number of visual functions for the computerscreen. The results of these functions can be saved as a user-profile and used as input for software to automatically adjust GUI and content to the visual needs of a large group of partially sighted. The module will be incorporated into the innovative digital atlas that is used in Dutch education (Bos@tlas by Wolters Noordhoff, Groningen). Because of the different visual possibilities of partially sighted users and the interaction between the different variables the program offers a fine-tuning possibility. To make the atlas truly accessible to partially sighted, the atlas will very much resemble a Geographic Information System (GIS).
Most of the literature and research in this paper had not yet been finished at the time the paper was sent to the conference. Research is currently done by Bartimeus, Wolters Noordhoff and the University of Utrecht. Research will finish during the last months of 2000. More information about current results can be obtained by emailing to Bartimeus in Zeist Contact: E.Velleman@bartimeus.nl
Coolenbrander, A.: (1996), What’s in a name: more people are blinded by definition, than by any other cause. In: Journal of Videology, pp 13-20. Veenendaal.
Go to previous article
Go to next article
Return to 2000 Table of Contents
Return to Table of Proceedings