On Oct 2, 2008, at 1:14 AM, Kristof Zelechovski wrote:

I hoped to see an example of an actual semantic improvement of introducing INPUT[type=SEARCH]. Maciej described a conceptual improvement for the ideal world; in the real world, the developers will have to use the heuristics anyway because there are search fields in the wild not marked as such and
most of them will remain that way of course.  Please describe how the
potential presence of INPUT[type=SEARCH] will improve the user experience (ignoring presentation) based on the assumption that the present heuristics
will be used anyway.

According to the Chrome developer who spoke up here, Chrome currently uses OpenSearch metadata. If many search fields used <input type="search">, then user agents could identify them without the need for OpenSearch metadata. I would imagine many smaller sites with a site-scope search box would not have such metadata but could easily adopt <input type="search">.

Regards,
Maciej


Chris

-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Maciej Stachowiak
Sent: Thursday, October 02, 2008 2:10 AM
To: Nils Dagsson Moskopp
Cc: whatwg
Subject: Re: [whatwg] native styling for search input boxes


User agents and assistive technologies could use the
knowledge that a field is a search field in all sorts of helpful
ways.
What exactly were you imagining ? In the end, it's a text field like
any
other.

For example, Chrome will keep track of search fields that the user has
used on various pages. I assume they currently use a heuristic, this
would be a clear signal of search-fieldness. (I do not speak for the
Chrome team here and I do not know if they would want to use it.




Reply via email to