jenkins <firstname.lastname@example.org> wrote:
> On 6/27/97 4:23 AM, Martin J. Duerst (email@example.com) wrote:
> >I am very familliar with the feature that conforming Mac applications
> >automatically change the input method if the cursor is moved to a
> >different place (with a different Mac script). Based on my personal
> >experience, I'm not at all convinced that this is the right feature.
> >These are my reasons:
> [Martin's reasons snipped]
> Actually, we've been getting a lot of feedback from users that they don't
> like this. We're probably going to drop it as a recommendation.
Please don't. Make the system smarter, and allow the user who wants to
manually control input method switching to specify when it should and
shouldn't switch in a control panel. Of course, one option in the
control panel would be to turn off automatic input method switching
completely and operate on "full manual".
Martin's reasons were mostly like the good arguments against autofocus
in cameras a few years ago: it's too slow, it's sometimes inaccurate so
the pictures get blurred, etc. Like Martin's reasons, these were very
important points, but they don't necessarily lead to the conclusion that
the technology referred to should be abandoned.
Consider the first of Martin's reasons:
>> - The user is not in control. The system is second-guessing.
Another name for "second-guessing" is "automation", and it's what
computers were made for. The problem is not "how do we make sure that
computers don't make decisions for us" but, rather, "how do we get them
to make the *right* decisions for us?"
Martin's additional reasons are observations of cases where the computer
is making a different decision from what he intended. These are
excellent points, and I could add more to the list, but they aren't good
arguments against automation. They're arguments for better, *smarter*
automation. They're also arguments for the need for a manual override.
With most of the bugs worked out, autofocus has become a feature that
most photographers (including a lot of pros who assumed it was only for
incompetent amateurs) would not want to live without, though pros still
demand a manual override.
Nothing guesses correctly all of the time, but there is always a certain
accuracy threshold above which people prefer to use the automation as
the default, and only override the occasional exceptions, rather than
operate in "full manual" mode all of the time.
Automatic input method selection can and should be made smarter, not
abandoned. I'm working in an old version of Filemaker (Japanese) that
doesn't switch the input method when I jump between English and Japanese
fields. I'm constantly tripped up by the system not doing what I was
expecting, e.g. trapping my English words in the kana converter, etc.
You know you have an interface problem when you find yourself making the
same mistake several times per hour, despite having used the product for
years. (Either that, or I'm just not very bright...or perhaps both. ;-)
Mac design, unlike the PC world, has always been based on blaming the
design instead of the user for interface difficulties, though, hasn't
When I put the cursor in the middle of a Japanese sentence, I'm almost
certainly thinking in Japanese and expecting Japanese to roll off my
fingers as soon as I start typing. (A useful system is supposed to at
least *act* as though it knows what I have in mind.) Same for English,
Chinese, Korean, and even sub-scripts like katakana. (If the cursor is
placed in the middle of a katakana word, I probably intend to write
something in katakana. I wouldn't mind if the system automatically made
that assumption in that case and I only had to manually switch my input
method for exceptions.)
Please don't do anything to make text input less smart, just allow for a
local manual override to handle exceptional cases and a systemwide
method for disabling the automation (completely or on a
feature-by-feature basis) for people who think the current state of
automation is more trouble than it is worth.
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:35 EDT