[thelist] semantic markup

Diane Soini dianesoini at earthlink.net
Thu Jun 3 21:42:14 CDT 2004


On Thursday, June 3, 2004, at 11:49 AM, thelist-request at lists.evolt.org 
wrote:
>
>
>>
>> I am wondering, however, what is there out there that actually only
>> reads a web page in its strictest semantic form? And how exactly would
>> it stumble over a <b> tag as opposed to an <em> tag. I would like to
>> try this thing out and run some tests.
>>
>
> For example, if you reckon your page will be accessed by a screen
> reader, then maybe you'll realise there's quite some difference between
> that something than simply looks <bold> and something that is <strong>

I know there are screen readers, but are those the only readers of html 
that require semantic tag? Are those the only devices that will read an 
html document in a strictly semantic fashion?

I've never used a screen reader. Do they really stumble on <b> tags? I 
would think that they would have built the ability to read such a 
common tag into the software. I'm not trying to be argumentative on 
whether <b> tags are ok. I understand the argument completely, but I do 
wonder about the assumptions made, such as that there are devices out 
there, other than screen readers, that will stumble on a non-semantic 
document. If those devices exist, what are they and where can I see 
one? I want to see what that looks like.

> Not much so far, but there's this:
> http://www.w3.org/2003/12/semantic-extractor.html

Ah, here is something, but I get a lot of java exceptions trying it 
out. On those pages I did get to work, it doesn't seem to work like I 
would expect. I tested Jeff Zeldman's site, figuring it ought to be 
pretty close to perfect. The extractor didn't extract much.

Anyway, I think it is an interesting conversation. I'm not much into 
flame wars on <b> vs <strong>, but more on questioning some of the 
dogma about certain sematic tags being somehow more appropriate for 
user interface components than others. For example, in java swing you 
use would use classes like JMenuBar or JToolBar, but in HTML the 
convention of the moment seems to be to use <li>.

Now something like JMenuBar is pretty darn clear to me what it is, but 
<li> is not clear to me to be a menu bar.  My only concern is that I do 
not think <li> is more semantically correct for a menu bar than some 
other methods because what we are talking about is not an unordered 
bullet list in the middle of some prose document, but a user interface 
component. A user interface component also has a language of context 
and meaning. Therefore to say that <li> is somehow more correct than a 
<table> (yeah, I know--please don't go there) is kinda silly. Neither 
are semantically proper or semantically perfect as far as I'm 
concerned. Neither describes a user interface element.

I'm merely questioning the belief that one method is actually more 
correct than another when the markup language was not invented with 
user interface components in mind, but with prose documents in mind.

It would be nice if we could use XUL, but we can't at the moment for 
general web content. I sure hope that a standard can come out of this 
need so that we don't have to stumble around like this for much longer.

Diane



More information about the thelist mailing list