[Accessforall] Minutes of AccessForAll Meeting on 2012-01-31

Liddy Nevile liddy at sunriseresearch.org
Thu Feb 9 21:10:05 UTC 2012


I am sure that we will have a problem because some of us see this as a  
metadata problem and some don't. I don't know another word or way of  
describing what we are doing as I think we are trying to use  
descriptions of resources and services to adapt what is delivered to  
the user to their needs and preferences.

Anyway, I am wondering why this group wants to reinvent how to do  
metadata? URIs are for machines - words are for humans. If you look at  
the ISO standard for metadata (N19788) you will see that there are  
words for humans and URIs for machines ????

As for what Gregg is calling tags - I am really confused. I don't know  
what they are. I think what he wants (what you want, Gregg) is an easy  
way of looking at the property-value pairs once we have them - so a  
person can have a collection that is relevant to them and forget the  
rest? or so they have the ones that are useful to them, even though in  
another context they might be useful to someone with a very different  
set of needs? For example, if my problem is difficulty with hearing  
but I have good eyesight, I want a collection of things to do with  
sound to work on but I won't need special things for sight. People who  
work with metadata create sets like this by having what they call  
faceted views - or they have application profiles - maybe a set for  
mildly deaf users, and a set for severely deaf users??

I think that fortunately many of the problems being raised here have  
been worked on for many years now by metadata experts and the  
solutions are not as far away as they seem.


On 10/02/2012, at 3:59 AM, Clark, Colin wrote:

> Hi Liddy,
> On 2012-02-09, at 1:39 AM, Liddy Nevile wrote:
>> I think we should remember that from the perspective of JTC1, a lot  
>> of work has been done to make a standard for metadata that is the  
>> simplest and most used way of providing and implementing metadata.  
>> I think at least for the purposes of JTC1, this expert advice  
>> should be noted. Beyond that, as most major entities in Europe use  
>> the sort of metadata advocated now by JTC1, I see no reason for  
>> trying to invent a better way of doing things?
> We're actually 100% in agreement here. I perhaps should have been  
> more explicit, but I see ISO 24751 as exactly an example of that  
> "judicious use of containment" that I was referring to in my  
> previous email.
> What we do need is to come up with a reasonable extension mechanism,  
> which is representation-independent (i.e. it can't depend on, say,  
> XML namespaces), so that communities can contribute their own  
> properties easily and in a way that won't collide with others.
>> Eric Miller has pointed out that the magic of URIs, as Tim BL says,  
>> is that they can always be disambiguated and they make it easy for  
>> people to add their own. I am not an expert in optimization but I  
>> do know that the use of URIs does not mean that people get lost  
>> because everything is URIs instead of words. The usual  
>> implementations have clear instructions and wizards for humans that  
>> 'speak' about the definitions of the terms represented by the URIs  
>> in ways that are appropriate to the audience using the wizards -  
>> and these vary enormously without any effect on the resulting  
>> metadata.
> I get the magic of URIs. My point is about human users of the  
> specification. Compare Gottfried's example (http://myurc.org/TR/res-prop-vocab1.0/ 
> ) to the ISO 24751 spec. Which one is easier to understand, at a  
> glance?
> The optimization question is straightforward: we just need to come  
> up with a non-redundant technique for solving the very important  
> issue that Gottfried raises with his proposal: namespacing.
> Hopefully this clarifies,
> Colin
> ---
> Colin Clark
> Lead Software Architect,
> Inclusive Design Research Centre, OCAD University
> http://inclusivedesign.ca

More information about the Accessforall mailing list