- Hiroyuki Miki with Oki Electric -

1.Evaluation or design work done:
1) Related to accessibility:
  1-1) Developing and user testing of embossed icons for accessible ATM  :
     - Embossed icons are designed for the blind and the low vision users
     - Two companies cooperated to unify each company's own set of embossed icons
     - Standardized set of icons between two companies are now one of the
       standard methods of ATM for the blind and the low vision people 
       in JAPAN approved by
     - Iterative designs and tests were conducted
     - Final user testing was done against 17 blind persons and 8 low vision 
       persons (20 years old upto 65 years old: average was 48.8 years old)

  1-2) Prototype development of an universal ATM for the Japanese market
     - based on some U.S. research center's proposal
     - programmed in JAVA as well as in VC++ (2 programs are developed for comparison)
     - User testing is planned.

  1-3) Prototype development of an accessible KIOSK  for the Japanese market
     - Different set of screen images are designed for each disability: total 10 sets
     - programmed in JAVA as well as in VC++ (2 programs are developed for comparison)
     - User testing is planned.

  1-4) User testing of an accessible ATM: Subjects were old people without disability.

2) Related to usability:
  2-1) design & user testing of a multi-functional remote branch terminal
       for loan companies
  2-2) design & user testing of ATM
  2-3) design & user testing of KIOSK
  2-4) a lot more designs & user testings of public sysytems

2. Issues encountered:
1) In standardization of two different sets of embossed icons between two companies:
   - Appropriate user tests were necessary to test recognizability,
     discriminability,memorability, and remindability of shapes and images:
   - We excluded "super" blind persons and "super" low vision persons,
     because they are too fast in reading Braile compared with ordinary blind 
     and low vision persons.
     The difference of speed would influence speed to recognize embossed icons 
     and thus influence statistic data so much.
   - We conducted iterative design & tests for recognizability and 
     discriminability before the final test
   - In the final test, for remindability we conducted a "two week test"
     to measure easiness to remind the icons: second test was conducted 
     one week after the first test.
     I wonder one week was too long or too short to measure remindability.

   - Convincing and enough data was necessary:
   - Some reviewer of our paper commented that we should have experimented 
     different number of icons to make sure the result -- one set of icons 
     are better than the other.
     In the real context of use, the number of buttons (identical to the
     number of embossed icons) are not arbitrary, so different number of 
     icons are out of the real context.
     Do we need still  need to consider unreal settings?  I wonder.

2) In user testing of an accessible ATM: Subjects were old people without disability.
   - Questionaire doesn't work well: Biased replies which contradict with
     measured data are more common than younger ages. This might be a cultural
     issue of Japan.
   - We consider negative comments seriously more than positive comments.
   - Questionaire doesn't work well: Old people's "can" opinion seems to be
     a little bit less than younger people's "can" opinion.
   - We don't consider old people's "can" opinion as it is.

*********************************************
Hiroyuki Miki
Group leader of Universal design team
OKI Electric Ind. Co., Ltd.  IT Lab.
550-5 Higashiasakawa-cho, Hachiouji 193-8550, Japan
Tel: +81-426-62-6721
Fax: +81-426-62-6706
INTERNET: miki445@oki.co.jp
--------------------------------------------
Also a Visiting Assistant Professor at Saitama Univ.
*********************************************