[whatwg] Suggested changes to Web Forms 2.0, 2004-07-01 working

[Jim Ley]

On Thu, 8 Jul 2004 13:44:41 +1200, Matthew Thomas <mpt at myrealbox.com> wrote:
> And with the keyboard, given a sufficiently well-designed datepicker
> (e.g. those used in MS Windows and Mac OS <X), replacing an existing
> date in a datepicker would take zero extra effort over entering a date
> in an empty field: each component would be selected when you navigated
> to it, so typing the new value would replace it.

I don't think a date picker is particularly useable for picking
birthdate (it's such a common sequence of characters just typing it
into a single box is fastest)  but I agree defaulting to today
wouldn't be a problem.

> Then the datepicker could default to whichever of min or max is closest
> to the current date/time. (This would be the most appropriate default,
> for example, for airline booking forms.)

I'm not sure I completely agree with this, many airline booking forms
choose today+7 as a default, not today+1 and the return leg today+1 is
not really a logical default either.  Also we have the problem that
for me the windows date picker is pretty useless at picking a specific
date.    For example hundreds of millions of transactions have
succesfully gone through HP openskies CGI scripts, which uses a simple
double select drop down (and optional javascript approach if you need
the datetime).   This is a known UI and very successful, how does a
datetime control really improve the user experience in this case?

> Password fields are another. How do you distinguish missing a password
> field from intentionally leaving it blank? You can't.

These (text/password) are different, since they can have validation
rules which require non-empty empty input, you can't do that with
checkbox/radio.

> Checkboxes are another. How do you distinguish missing the checkbox
> from intentionally leaving it unchecked? In HTML at least, you can't.

In IE's implementation you can, and I think this is a great feature
and should be considered.

> SELECT menus -- as explicitly stated in the WF2 spec -- are another.
> How do you distinguish missing the menu from intentionally choosing the
> first option? You can't.

but the server can identify that "please select..." in the first entry
didn't happen.

Jim.