Re: [status] weekly report

From: David Huynh <dfhuynh_at_csail.mit.edu>
Date: Mon, 27 Mar 2006 10:31:32 -0500

David Karger wrote:

> David Huynh wrote:
>
>> Last week:
>>
>> - Conducted 8 user study sessions on the UIST prototype. Although I
>> had streamlined the auto-scraping interface down to requiring just
>> *2* button clicks, some users still had to make a leap of faith to
>> click that second button, labelled "Continue," which was required to
>> confirm that the system had correctly estimated the total number of
>> items spanning several pages.
>
> perhaps "continue" is too ambiguous and something like "these choices
> are right" would have given them more confidence that it was the right
> time to click the button...

There are actually a few radio options above "Continue":

    We have estimated a total of
          90 ~ 100 items
    on 10 pages (10 items per page).

    ( o ) This estimate is correct.
    ( ) This estimate is wrong.
            Let me locate the items myself.
                [ Locate Items ]
            Let me locate the pages myself.
                [ Locate Pages ]
    ( ) There are no additional pages.

             [ CONTINUE ]

I think the cognitive barrier is due to the fact that we need to scrape
the Web pages in the first place. To 99.9% of users, a text string "Buy
now for only $139.99!" and a floating point number 139.99 are pretty
much the same and the need to parse one into another is superfluous.

If you listen to users, you'll hear "it doesn't let me do this/that"
more often than "it can't do this/that". I think the reason is that for
most people, whatever information they can see on the computer screen is
in the same goo that the computer has complete control over. There is no
concept of data type or format or data model whatsoever. If the user
cannot do something, it's because the computer is magically stubbornly
refusing to let her do it rather than because the computer is incapable
of doing it.

David
Received on Mon Mar 27 2006 - 15:30:58 EST

This archive was generated by hypermail 2.3.0 : Thu Aug 09 2012 - 16:39:18 EDT