Skip to: Navigation | Content | Sidebar | Footer

Weblog Entry

Notes From All Over

September 26, 2003

On CSS redesigns, ISSN, semantics, and FIR revisited.

A note from Paul Hammond to those who would rebuild existing sites in CSS:

CSS can do so much more than tables ever could, and your work is showing none of this. In fact, it’s hinting at the opposite, that CSS doesn’t give us any new toys to play with and people can probably get away without learning it, as long as nobody looks at the source.

Read the rest. Thought-provoking, and touches on a very good point: consider your etiquette. No one likes to be told they’re wrong. §

Requisite link time: Joe Clark dismisses the increasing rejection of weblog owners who apply for ISSNs. You may remember an article on mezzoblue a few months back relating my own problems in obtaining one for this site. First-hand experience on this matter suggests Joe is entirely correct in both his assertions and conclusions. (via Zeldman) §

Marking up a title. This is a simple book title, folks, it should be open-and-shut. That it’s not (and don’t get me wrong: it isn’t) suggests a big hole in general understanding of what is so semantic about semantic XHTML.

I’d argue that true semantics require rigid definitions, leaving little room for grey area. But then I realize I don’t know the first thing about the Semantic Web, and browsing through the source is far from enlightening: RDF is covered every which way, but I haven’t seen a single mention of XHTML.

Do we have an ‘Elements of Style’ for semantic XHTML yet? Will one exist? Does XHTML even matter in the context of a semantic web, or is RDF the ticket?

These are questions. I don’t know the answers. §

In bringing back online some of the still-missing content last night, I added this roundup of alternative FIR methods that came to light around the time I published my Digital Web article. I never publicized it, because it seems there really isn’t a perfect alternative yet. Phark is nice code-wise, but doesn’t work in IE5 and doesn’t solve the images off, CSS on problem. Levin/Gilder is code-heavy and rather ugly, but does solve the accessibility issues… though it introduces a new problem with transparent GIFs.

We haven’t cracked this one yet, but there has to be a perfect solution out there somewhere… §

Reader Comments

September 26, 01h

about the ISSN issue: sure, joe clark - by all means a sacred cow of the (accessibility) web community - says that the registrars have no right to refuse ISSN registration of blogs. well…my first reaction would be: and so what ? a blogger defending bloggers’ rights to register…that’s certainly going to put pressure on the registrars, eh ?

Gabe says:
September 26, 02h

Wait a sec Patrick, on first glance I thought your comment might have some merit, but then I realized: who defends anybody’s rights to anything? The people themselves of course! Sure, sometimes a mysterious and powerful benefactor may take up someone’s cause, but that is hardly the driving force behind societal progress. You ought to take the merits of an argument based on the argument itself, not your opinions about the author.

September 26, 02h

what i’m saying is that it is naive, in my opinion, to think that just because everybody links to joe’s article, and makes a big fuss about it by whatever means (emailing the registrars ? calling their government representatives ?), that things will automatically change. i’m getting a slight deja vu here, thinking back at the “free jaws for web developers” discussion…

i’m all for the idea, and no, i don’t necessarily have anything against joe…i just don’t like the feeling of “joe said it, so iit will have more clout” subtext i get from many articles on his writeup…

MikeyC says:
September 26, 03h

“Joe Clark dismisses the increasing rejection of weblog owners who apply for ISSNs”

I’m sure that talk about ISSNs will pop-up at the next meeting…looking forward to it ;)

I think that the policy is to reject blogs outright unless the owner complains. Realizing they have no defense for this type of discrimination against a new publication medium they eventually will grant the ISSN. This process most likely filters out about 95% of the blog requests.

I was initially rejected…complained…then granted an ISSN a day later (ISSN 1708-2935). Two emails is all it took to get an ISSN (plus a third to confirm that I had received the second one granting me the ISSN).

So my advice is to save all the arguments on why your blog deserves an ISSN for the 2nd email as the 1st will simply get you a form letter response. The 2nd email is the one they will be forced to read if for no other reason than that the continued correspondence begins to irritate them and it’s just easier for them to issue you the damn ISSN!

I posted my initial thoughts on the whole ordeal:

andreas says:
September 26, 05h


1. The link to your revised image replacement page works fine from your blog’s frontpage, but not from the permalink page. Dashes in the URI are replaced with %E2%80%93.
2. There’s a glitch in the first and second textareas’s content; the string ‘background-image:’ is mentioned twice.

Else, nice overview.

James F. Cerra says:
September 26, 06h

Re: rebuild existing sites in CSS

I don’t recode to show somebody they’re wrong; I do it to practice my CSS skills. I often figure out interesting and new hacks, compromises, and insights for my CSS “tool-belt”.

James F. Cerra says:
September 26, 08h

Re: Roundup of Alternative FIR Methods

Is there really a problem? Lately, I’ve been leaning towards pseudo-RDF for assigning IDs and classes. For example:

<div class=”title”>Foobar</div>
<div class=”subtitle”>What does it mean? </div>

The h1 tag indicates that we have a heading, the first div is the title of the heading, and the second div is the subtitle. This would be like RDF triples:

heading1 title “Foobar”
heading1 subtitle “What does it mean?”

Is there anything inherently wrong with this technique?

Eric says:
September 26, 10h

A request: Would it be possible to put items as separate posts? It’s kind of daunting to post a reply on one particular thing when there will likely be comments on other unrelated parts of the post.

Valette says:
September 26, 11h

the gilder/levin method is repeating the background image vertically - i see two copies of the image. win2k firebird0.6

Seamus says:
September 26, 11h

Good to see you updated the roundup page from a two months ago (I found it through backtracking on refers). Maybe we need to get the screen reader makers, web designers, web browser designers, and accessibility experts in one room to figure out how to clear up all this mess and to make everbodies lives easier.

Lach says:
September 27, 01h

Dave, just a small quibble.. XHTML isn’t, and never will be, semantic. It does not tell you what anything means. It’s purely structural. It’ll tell you everything you need to know about what a piece of text *is*, but nothing whatsoever about what that piece of text *is describing*.

In the end I think the best solution to achieving both will be to markup all documents up using *two* markup languages. XHTML to inform tools of a document’s structure, and another one (hopefully easier to get than RDF) which sits in the same document identifying meaning.

Then again, there are only a small subset of documents that will ever need both of them anyway. Structural description is more than enough for most ‘content’ documents around, and documents which require semantic description are more storage / database style files than content.

September 27, 03h

I’ve bene thinknig of applying the new Skip navigation method recently discovered as a pseudo FIR replacement. You can read the full story at
but in practice this allows screenreaders and google to fully access your headers without any penalties for hiding information (as opposed to h1 span, hiding the text in span and using FIR), as well as any CSS disabled browsing device. CSS enabled browsers that have images off can be catered to by putting the same information in the image?s title attribute. This way it is possible to accommodate all users.

September 28, 10h

If RDF is ever the ticket, I’m going to refuse the ride.

If we’re going to move toward semantic markup, there needs to be something that provides semantics. I have two more questions without answers, though: do we really need to move toward semantic markup? And if so, is it enough to have a human- (but not machine-) readable profile that defines the semantics? That’s what HTML provided, and it seemed to work well enough. Everyone knew what the ‘p’ element meant, even if they didn’t use it.

Danny says:
October 06, 12h

Re. Semantic markup - good questions (I’ve added a link to your post on the ESW Wiki, maybe someone will take note).

Personally I’d lean towards Lach’s suggestion of separate languages for structure/meaning, but these certainly don’t have to sit in the same document. XHTML (+CSS) + RDF would I believe be a good choice. RDF isn’t difficult, it’s just different from most markup so people tend to get wrong-footed. RDF/XML can look horrid, but it’s designed primarily for machine, not human consumption. If you look at the model first (which is the good bit!), it’s not far from familiar relational database ideas.