At 12:00 AM 4/5/2006, Richard Czeiger wrote:
After all, SOMETHING has to go in the <dt> and it makes sense to
make it the one thing that is MOST appropriate.
Taking your logic to the nth degree you might as well put everything
in a <dt>.
Not really -- if everything's in DTs there's no point in using the DL
in the first place; you might as well use an unordered list. As you
know, the DL has at least these advantages:
1) its semantic structure is that of a list of item/description pairs
which matches up well with many of the lists we present;
2) it's an HTML list with two types of child element, making it easy
to style them differently without using classes. Unlike OL and UL,
each DL list item can consist of multiple DT and DD tags, increasing
the markup & styling possibilities.
When you're marking up a literal dictionary you usually know which is
the term and which is the definition: the term is usually the short
one, is usually listed alphabetically, and usually comes first; the
definition is usually longer, is usually not alphabetical, and comes
second. Or the term is in vocabulary A and the definition is in
vocabulary B. The definition is often thought of as describing the
term, but really they describe one another, collaborating in a common
web of meaning. Which is the DT and which is the DD depends on the
author's purpose for the list and their perspective on its content.
With other types of collection that stray farther from the
"dictionary" model, it's my assertion that which term goes in the DT
and which in the DD can be even more arbitrary. If you're marking up
a collection of items in which no one element is definitively The
Term but instead in which several items collaborate to form a single
gestalt, then yes, you could put more than one item in DT and/or more
than one item in DD. (Multiple DTs are common even in literal
dictionary lists, as illustrated in the W3C spec.)
Even a "description" can be a DT:
DT a large, quadrapedal, grazing ungulate
DD cow
DD deer
DD horse
If you'll grant me that flexibility, then I'll conclude that it's not
incorrect to freely choose the one item for the DT that will make
styling the list a breeze:
image title
description
price
Doesn't the title describe the photo as much as the photo describes the title?
True: A photograph isn't a 'definition' but it's also not the
primary identifer for most scenarios outside a Photo Gallery. The
reason I know this is because if you pointed me to a URL and I first
asked what I'd see there, you'd tell me "its a Mbira Dzavadzimu",
you might even say "its a photo of a Mbira Dzavadzimu" if it was an
especially spiffy photo but you wouldn't just say "a photo" which is
exactly what <img /> is interpreted as. It's technically correct but
not especially helpful. Textual data, in this case and most others,
seems to be the most appropriate way of initially identifying an
object/concept, at leats in the telecommunciations medium.
No, I wouldn't likely say "a photo" any more than I would say "a
headline." An IMG tag by itself doesn't convey much meaning, but
neither does H1 or P. It's the content of those tags that conveys
the most useful meaning to us.
What I hear you saying is that the crucial difference between text
and image in this context is that text can be encoded in a defined
character set and is therefore parsable but images are transmitted as
bit-streams without a similarly simplistic symbolic encoding and are
therefore unparsable. I can imagine this argument for machines --
although image-search engines are going to blow that one out of the
water in the next few years -- but more to the point when sighted
humans look at a page they read, comprehend, and remember the
meaning, content, and context of both the images and the text. If
I'm bookmarking pages about mbira I'll grab the ones with photos and
drawings and scans of musical transcriptions as well as the ones with
plain text.
According to your argument (I think), if I scan in text or produce it
in Photoshop, then the meaning of that text is lost and it no longer
qualifies as something that can go in a DT, unless the DL is, say, a
gallery of font samples.
Imagine browsing through a catalog in which the webmaster has used
Photoshop to render the product names as images of text but failed to
provide alt attributes: you're still reading the text if you're able,
it's just not ASCII-encoded. Do the product names no longer go in
the DT because of the method used to render the text? But then if
you add alt attributes they can go back in the DT? Hmm.
I feel that an image can be a term to be defined even when it's a
photo or drawing of a product in a catalog. Imagine browsing through
a clothing catalog: you see an outfit that catches your eye, then you
read the item's name and description. The text defines and describes
the image. Imagine you're looking through a web developer's
portfolio: you see a thumbnail of a website that looks interesting so
you read the blurb. Imagine you're looking through any list
whatsoever that includes images and text: you see an image that looks
interesting and you read the text that describes it.
Now if you're text-searching those same listings, you'll locate items
by their titles or blurbs or alt-text; in that case you may already
know the name of something and the image will act in a supportive
role to help describe the item. It's the same list, though, and I
doubt many people would propose that we swap DTs & DDs dynamically
based on our assumptions about how the visitor is using it in the moment.
I'm curious: why are you applying different standards of
meaning-interpretation to images and to text? If you regard images
as anonymous without regard to their pictorial or alt-attribute
content, then wouldn't it follow that you'd see <dt>Mbira
Dzavadzimu</dt> as an anonymous text block without regard for its
meaning as well? Of course you wouldn't (even though you may not
know what "dzavadzimu" means or even if it means anything at
all). We attribute meaning to the symbols we use. As you know,
words in English are spelled with letters each of which is derived
from a picture of a concrete object, you know, like alpha/aleph
started out as an ox's head. Text is made up of strings of these
letter-pictures that have, over time, been assigned new
meanings. But it's not that different with what we call images. A
photo of a dog may seem like just a photo of a dog, but if a
particular breed of dog is urinating at the base of a particular
national flag then a whole flock of additional meanings take
flight. A bear in a cartoon can say "Russia." A T-shirt that says
"bum" across the chest can mean very different things depending on
what font it's in. The Firefox logo brings up associations of foxes
even though the software itself has nothing to do with the
critters. The Internet Explorer logo includes a big blue e. When is
an "image" a "word" and when is "text" an "image"? Certainly the
symbologies are different, but is the type of information conveyed so
categorizable and so categorically different that it can never
qualify as a definable term except in an art or photo gallery?
How about a non-verbal dictionary of international symbols:
DT symbol for no smoking
DD picture of a man smoking
DD picture of a custodian scolding the man and pointing to the symbol
DD picture of the man putting out the cigarette
DT symbol for men's toilet
DD picture of a man drinking
DD picture of the man walking through a doorway marked with the symbol
DD picture of the man urinating
But maybe you're not talking about meaning theory but instead about
web accessibility. In five or ten years when Jaws can read a page of
text including "image of a Golden Retriever catching a frisbee," will
your argument about images and text necessarily differ?
Warm regards,
Paul
******************************************************
The discussion list for http://webstandardsgroup.org/
See http://webstandardsgroup.org/mail/guidelines.cfm
for some hints on posting to the list & getting help
******************************************************
