OK, Misha, I'm having a hard time figuring out whether my leg is being
pulled or my chain is being yanked :-)
But I'm game...
Is the problem here that oldest of network bugaboos... Sometimes a name
(title) is an identifier, and sometimes an identifier is a name (title),
and sometimes both are true?
I've taken 15,000 digital images in the past two years. They come out of
the camera with identifiers... File names with a bit of crude but
helpful semantics in them (sequence and file format).
300 of these images I spend enough time with to actually assign
"titles"... The rest just live with their filenames as titles.
What would the metadata look like? Well, it would be a mixed bag of
identifiers and 'titles', EACH of which would be useful for dragging out
those images, and each would be handicapped.
I don't think it will surprise anyone that DC specifications don't solve
this problem for me.
I *CAN*, however move towards a more coherent world by establishing a
local (my own) convention
On this topic in my own environment... Isn't this a rasonable thing for
the News community? Or are the legacy collisions already hopeless?
stu
(who is not the same stu as was formally known, and hasn't been for days
now)
-----Original Message-----
From: Misha Wolf [mailto:[log in to unmask]]
Sent: Thursday, June 23, 2005 3:32 PM
To: Weibel,Stu
Cc: [log in to unmask]; [log in to unmask]
Subject: RE: Never mind the syntax, feel the semantics
Hi Stu,
Here's an extract from a mail I've received on this subject, from a
leading member of the Working Group that developed the IPTC Core for
Adobe's XMP.
Key to abbreviations: "IIM" is an old news standard and "Descr CoCo"
is the Descriptive Common Component we are developing for all our News
standards.
<quote>
[...] discussing the "IPTC Core" this working group became aware of
problems with stating "semantic equivalence" of metadata properties.
An example:
- Adobe adopted from IIM dataset 2:05 "Object name" with an
explantion of "Used as a shorthand reference for the object."
for a field called first "Object name" then "Object Title" and
finally only "Title".
- Adobe adopted for XMP the DC title field - and synchronises data
of IIM 2:05 and of dc:title with an explantions of "A name given
to the resource." and a comment "Typically, Title will be a name
by which the resource is formally known."
- On first sight this sound reasonable. But many photographers and
agencies understood this field in Photoshop as the right place to
write down the image file name. (This was discussed extensively by
the IPTC Core group - but we found a photographer doing this
didn't break an IPTC nor an DC rule - see the explanations)
- If the IPTC states "dc:title" and the "title" of the Descr CoCo
are semantically equivalent this could happen: a software would
read the dc:title/IIM 2:05 metadata out of an image file and
transfer this to the "descr:title" - then the user will NOT see
"A label acting as a short introduction to the content."
(= explanation from the CoCo document) but a file name :-(((
</quote>
I think that the statement:
Typically, Title will be a name by which the resource is formally
known.
applies only to the library/art/museum community, but not to any of the
many communities where titles are completely transient. I suggest that
we figure out for how many years the comment has said what it now says,
and change it, for the same number of years, to
say:
Typically, Title will *not* be a name by which the resource is
formally known.
It is, after all, just a comment, so that should be OK, right :-)
Best wishes,
Misha
-----Original Message-----
From: Weibel,Stu [mailto:[log in to unmask]]
Sent: 23 June 2005 20:13
To: Misha Wolf; [log in to unmask]
Subject: RE: Never mind the syntax, feel the semantics
This seems easy to me...
Point A:
The DEFINITION is broad and inclusive, and seems to me to clearly
satisfy both the biblioheads and the webheads.
The COMMENT is just that... A comment. Intended to clarify (oops... We
might not have done the best possible thing in this case, though the
word 'typically' is a very legitimate trap door).
Definitions are normative, comments are... well... Comments.
Point B:
The tricky balance that Misha articulates below has always been hard,
and inevitably rough around the margins:
a balancing act between having definitions that are so broad
that they become meaningless and definitions that are so narrow
that they fit only one community and are not shareable.
Always we should be apply the test of common sense... In this case, its
asking the question...
What is the most title-like-object in this resource, and if I choose it,
am I likely to blow up any other community's notion of THEIR
title-like-object?
The answer seems SOOOO obvious to me that I can't understand why this is
even an issue.
Hey, Misha... We've missed you!
s
-----Original Message-----
From: General DCMI discussion list [mailto:[log in to unmask]] On
Behalf Of Misha Wolf
Sent: Thursday, June 23, 2005 2:26 PM
To: [log in to unmask]
Subject: Never mind the syntax, feel the semantics
I'll start by mentioning that I've put on a hard hat and a flame-
retardant cape, just in case I need them.
It's also worth reiterating Stu's mention of my long involvement with
DC. See, for example, RFC 2413 (Dublin Core Metadata for Resource
Discovery), dating from 1998:
http://www.ietf.org/rfc/rfc2413.txt
As I've mentioned in previous postings, the News Architecture Working
Party of the International Press Telecommunications Council
(IPTC) is actively examining the use of DC for those of our metadata
elements where there is a good semantic fit. Having been involved with
DC all those years ago, I had assumed that this would be a relatively
pain-free matter. I was wrong. Consider the humble title. RFC 2413
defines this as:
The name given to the resource, usually by the Creator or
Publisher.
The current official DC documentation states:
Definition: A name given to the resource.
Comment : Typically, Title will be a name by which the resource
is formally known.
Ouch! This comment may well work for the Library community. It
certainly does not work for many other communities, such as Web page
authors, professional photographers, or news organisations.
If I change the title of one of the hundreds of Web pages I maintain, I
am most certainly not changing "a name by which the resource is formally
known".
The same applies to a professional photographer changing the title of
one of thousands of photos on her/his computer.
And the same applies to a news story ... the title (ie headline) is most
certainly not any kind of formal name.
So we have a problem. If the Semantic Web is to work, it is not enough
to employ some common syntax or even a common abstract model.
We need to be able to share meaning. And this is obviously a balancing
act between having definitions that are so broad that they become
meaningless and definitions that are so narrow that they fit only one
community and are not shareable. Those of us working on the
architecture of mainstream news standards, perceive the comment
associated with dc:title as being on the latter end of the spectrum.
And so, as Chair of the IPTC News Metadata Framework WG, I am asking the
DC community to reconsider the text of the comment accompanying the
definition of dc:title.
Many thanks,
Misha Wolf
Standards Manager, Reuters
----------- -----------------------------------------------------
Visit our Internet site at http://www.reuters.com
To find out more about Reuters Products and Services visit
http://www.reuters.com/productinfo
Any views expressed in this message are those of the individual
sender, except where the sender specifically states them to be the
views of Reuters Ltd.
----------- -----------------------------------------------------
Visit our Internet site at http://www.reuters.com
To find out more about Reuters Products and Services visit
http://www.reuters.com/productinfo
Any views expressed in this message are those of the individual
sender, except where the sender specifically states them to be the
views of Reuters Ltd.
|