W3C home > Mailing lists > Public > public-html@w3.org > June 2007

Re: fear of "invisible metadata"

From: Joshue O Connor <joshue.oconnor@cfit.ie>
Date: Fri, 22 Jun 2007 09:56:44 +0100
Message-ID: <467B8ECC.3090701@cfit.ie>
To: Charles McCathieNevile <chaals@opera.com>
Cc: Philip & Le Khanh <Philip-and-LeKhanh@royal-tunbridge-wells.org>, Maciej Stachowiak <mjs@apple.com>, Laura Carlson <laura.lee.carlson@gmail.com>, public-html@w3.org

Charles McCathieNevile wrote:
> On Fri, 22 Jun 2007 00:07:47 +0200, Philip & Le Khanh <Philip-and-LeKhanh@royal-tunbridge-wells.org> wrote:
> 
>> Joshue O Connor wrote:
>>
>>> The problem that you seem be indicating is the difficulty in writing *good* alt text. This is an entirely different issue but an
>>> important one. The best the WG can do is provide tools that make
>>> this easy
>>
>> What sort of "tools" did you have in mind ?
> 
> Attributes and elements - which are all that this group makes :) 

Philip, as Chaals says - attributes and elements. And in many ways all
we need are the correct, supported, and backwards compatible attributes
and elements. (Copy, paste and repeat this mantra ad nauseum.)

> But attributes and elements don't "make [writing *good* alt text] easy",
> which is what Joshue is referring to, and the reason for my (genuine)
> question to him. 

I appreciate that you are making a genuine inquiry and I will try and
give a genuine answer, which is that it actually *isn't* easy. Writing
good appropriate and meaningful alt text (when required) really is a
particular skill that is acquired over time and improves over time. It
really requires thought and patience etc.  For some, this is a real
stumbling block to other automated processing of data  as it requires
human intervention.

In many ways no matter what @future_attribute/elements the WG comes up
with, it will more than likely still be time consuming, difficult etc.
There is however work being done by others (in HCI circles, I have a
colleague in Dublin who is looking at this) into sophisticated pattern
recognition in order to automatically tag images with appropriate meta
data for classification and archiving purposes - that could help
automate (or assist) in the future. Not doubt these techniques will
become more sophisticated as time goes on.

Josh
Received on Friday, 22 June 2007 08:57:05 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:01 GMT