27 people says things!

1. Semantics let you attach meaning to your content that was previously not there if you just through your design/information into a table. Semantic code is not just for text-based browsers, but for accessibility and for alternate-display devices as well.

   If the header of a site is stuck in a div inside a table, how is that any better than putting it into an H1 where it belongs?

   I think this guy is just spewing troll trash.

   By Mike on September 1, 2005 1:49 pm

2. His argument is so 2001. Who, besides this joker, is honestly defending table-based Web site layouts?

   By Jason on September 1, 2005 2:42 pm

3. We’ve heard this arguement before, and from better writers. Regardless of your take on the visuals, the focus of handicap accessibility should be a top five in your reasons for focusing on semantic methods, just as Mike pointed out… if not, then the flexibility to adapt to multiple devices, or update your design at will without touching your content.

   In the end, just because it looks good on your text editor, doesn’t mean it sounds good through your screen reader.

   By Brady J. Frey on September 1, 2005 4:15 pm

4. I think a lot of people are defending table-based layouts. (No statistical data to back up my claim…).

   I do think that he was missing the point for semantic coding. If you semantically layout your website, then you can drastically alter the design without changing any of the html. By taking the extra time to create a semantic document, you save yourself a lot of time in the long run by allowing redesigns to be done with a lot less modification. Just switch out the stylesheet and you’re done. (think of the zen garden). It also allows you to offer different layouts to different display devices (which would be harder to accomplish in a table-based layout).

   I think semantics is underrated…

   By Kyle Posey on September 1, 2005 4:17 pm

5. Maybe this is just a “shock” piece? I’m mean, really, in what world does table-based design make more sense?

   By Dan Wolfgang on September 1, 2005 4:18 pm

6. Semantics are overrated. Good, clean, flexible, maintainable code is underrated. I agree with the author.

   By Mike D. on September 1, 2005 4:45 pm

7. Comedy gold. At least if he’d used Lynx, a text-only browser more than three people in the world have heard of, his argument would have been somewhat compelling. Of course Lynx (which, more relevantly, renders similar to plenty of mobile browsers) totally reverses his findings, but hey. They got a link out of it right?

   By Dave S. on September 1, 2005 4:53 pm

8. Well, if we’re going to talk about semantics, it’s important to really define what “semantic” really means. I get the feeling that the word “semantic” gets tossed around far too much as a “buzz phrase” without actual understanding of the meaning of the word.

   According to a dictionary definition of the word “semantic”, the word means “of or relating to meaning in language” or “of or relating to semantics”. The meanning of “semantics” is “the study of meanings”.

   So really, all semantic really means is “with meaning” - when we say divs are more correct semantically than tables, all we’re saying is that divs have a more appropriate *meaning* than tables. But do divs really have any meaning?

   My point is - semantics are useless (like the above guy says) but not in the same way that he argues. Let’s not focus on whether tags actually mean what we want them to mean - whether a certain header tag is more appropriate than a certain other header tag, or whether we should use <div>s or <span>s - semantics have nothing to do with accessibility. Accessibility has everything to do with accessibility, and tables aren’t accessible. End of point.

   By francey on September 1, 2005 5:26 pm

9. …the website of one of the most influential CSS designers in the history of the Internet.

   Is that quote going to go on Dave’s next book, I wonder?

   Anyway, not a very good article. Let’s disregard the accessibility arguments for a moment.

   Yes, a simple table-based looks a lot like a simple CSS-based page. Maybe it even uses less code. But if you do a more useful real world comparison - say, Microsoft’s old site compared with Doug Bowman’s CSS-based redesign - this argument would be shown to be completely fatuous.

   By Christian Watson on September 1, 2005 5:42 pm

10. The guy just doesn’t know what he’s talking about. His criticism of Mezzoblue’s “semantic design” mainly consists of complaining about his browser’s DEFAULT indenting of sequential headings.

    By Craig C. on September 1, 2005 9:37 pm

11. I find it funny that way down at the bottom of his page he has this:

    “Comments exist, but only registered [users?] whom are logged in may see them.”

    ^ As far as accessability goes, he’s not marching to his own tune. While I have always been a fan of the attitude expressed in Mike’s Article, and can see where he’s coming from, I agree with the other Mike on this one:

    “I think this guy is just spewing troll trash.”

    By Nathan Smith on September 2, 2005 1:01 am

12. Ouch guys, hate the article, not the magazine. Either way, it is contribution based and so I had no issues posting his viewpoint.

    “”Comments exist, but only registered [users?] whom are logged in may see them.”
    ^ As far as accessability goes, he’s not marching to his own tune. While I have always been a fan of the attitude expressed in Mike’s Article, and can see where he’s coming from, I agree with the other Mike on this one:
    ”

    I actually developed the site, he just wrote the article. So that problem, feel free to take up with me… article issues, well, feel free to leave a comment with an opposing viewpoint :) I knew it was going to be contreversial when I published it.

    Phil Brodeur
    Neversidian Director
    p.brodeur@neversidian.com

    By Phil Brodeur on September 2, 2005 1:18 am

13. Sheez, some people take an oppositional opinion and just go wooo.

    The point of this ezine here is not to get links, traffic or anything of the sort or near the contrary. Getting a link here is like ‘oook’. However, it is a contribution-based ezine and depends on the viewers to submit content to be published.

    Noel’s article was passed around and many people who ‘are’ into semantics and so forth, and they agreed with many points. This isn’t to discredit or say another’s opinion is of equal or lesser value, it is merely to express a perspective on the entire subject.

    That my friends, ‘is’ the point of this ezine. To allow anyone/everyone to submit, and use quality control to manage it.

    Now, in all fairness… Noel wrote it merely to spur the life in your panties, nothing more or less. While I knew such would be condemned by the semantics community, it had some valid points and interesting things to say. We look forward to any ‘valuable’ contribution or even input as this is a new venue for us.

    Thanks for the link!

    Shawn R. Lockheart
    Director of Neverside Community Relations
    locke@neverside.com

    By Shawn R. Lockheart on September 2, 2005 1:55 am

14. One of the biggies I’ve come to realise over time about good markup is that it is then very easily scriptable. With a simple page, it should be possible to add various behaviours with .js without touching the original html templates. This can only be done easily if “semantic” stuff is used, at least to the extent of shoving classes and ids all over the place.

    By Gregory Dyke on September 2, 2005 2:57 am

15. it had some valid points and interesting things to say.

    This is the bit where I have to disagree - it doesn’t seem to say anything that hasn’t already been beaten to death multiple times over the last few years.

    By Mathew Patterson on September 2, 2005 2:59 am

16. I think the kernel of good in this article is getting lost in how it was written. Sure it’s over-the-top radicalism to suggest that tables are actually philosophically better than CSS layouts, but it is indeed true that CSS, as the language stands today, does not fully address the needs of developers, designers, and electronic publishing in general. There is still much innovation in this space just begging to happen but the state of the web currently prevents it (or at least severely slows it). No, I’m not just blaming the spec writers here. It’s a combination of spec writers, browser makers, OS providers, and even users which conspire to slow things down.

    So… is the author right that CSS ain’t “all that”? I think he is. But where he’s wrong is that it’s pretty good, and it’s the best we’ve got — and therefore we should use it while continuing to push forward.

    By Mike D. on September 2, 2005 3:13 am

17. Yes, a simple table-based looks a lot like a simple CSS-based page. Maybe it even uses less code. But if you do a more useful real world comparison - say, Microsoft’s old site compared with Doug Bowman’s CSS-based redesign - this argument would be shown to be completely fatuous.

    You are comparing a badly coded table based site with a well coded css based site. I think with a well coded table based site using css can be just as clean and have minimal code as a pure css based design.

    BTW I really love the design of Neversidian.

    By Tom on September 2, 2005 7:28 am

18. After reading this article I am starting to question the logic of some people.

    Not that the author hasn’t thought this out, however the logic and thoughts behind all this are skewed.

    And someone please tell me…

    “…help organize the code of a page more effectively…”

    WTF tables are these ones that he is using?

    By Ryan Latham on September 2, 2005 9:54 am

19. “I feel that it is important to take a new look at using tables to design websites, and perhaps even make recommendations to the W3C to create new tags that will help designers organize and manage websites even more efficiently than tables can.”

    Aren’t they already doing this?

    I do tableless design wherever possible, but floating divs isn’t the greatest either; let’s be honest. While I do it, it always strikes me as a bit of a hack. I hope the w3c recommends something better.

    By Mark Priestap on September 2, 2005 10:59 am

20. On second thought, I retract. My first comment was in reaction to the pullquote that appears in this entry (which isn’t too bad in my opinion). Not to the article itself. Serves me right for commenting before reading the whole piece.

    <crawls back under rock />

    By Mike D. on September 2, 2005 11:21 am

21. Comedy gold. At least if he’d used Lynx, a text-only browser more than three people in the world have heard of, his argument would have been somewhat compelling. Of course Lynx (which, more relevantly, renders similar to plenty of mobile browsers) totally reverses his findings, but hey. They got a link out of it right?

    Mr. Shea, Mr. Shea… so, your justification for this is, essentially, since you’ve never heard of links or links2, my argument doesn’t count. I said the majority of browsers render tabular layouts better. I said in the FAQ that not all text-based browsers will render the sites the same way. You have to realize that every major operating system/distribution has a package, port, or for links. It’s not unknown to the Internet. Try a real argument next time.

    For the rest of you… understand that I don’t actually hate semantic code. I know that it has its usage — if your website is aimed at, for instance, blind users, semantic code might be just what you need (screenreaders don’t describe tabular layouts so well, as one of the comments to my article demonstrates).

    I find it funny that essentially the worst argument against what I said is that I’m complaining about default indentation of headers. Yes, I am, but if you work your way around that half of a paragraph, you might see some of the other things I wrote.

    By Noel Forbes on September 2, 2005 3:50 pm

22. Mr. Forbes, you cited default indention as a failure of semantic markup, betraying your ignorance of the very matter you’re bitching about. Many of the other things you wrote are bunk too, and also seem to stem from your basic misconception of what web semantics is about.

    You complained about the appearance of Mezzoblue’s ISSN registration in the site header, simply because you didn’t seem to understand what it meant. That’s not Dave’s problem, nor the browser’s, it’s just yours.

    You complained about not being able to find the sidebar, which comes at the end of the document as supplementary content (that is, content separate from the main content, of lesser importance). “Sidebar” is a publishing term and doesn’t neccessarily refer to a spacial location. Perhaps Dave could choose a different name for it to avoid just this sort of confusion, but had you actually selected the link you complained about you would have been directed straight to the sidebar. That’s why the link is there.

    In web standards parlance, “semantics” has jack-all to do with the visual layout of a page. Semantic markup means taking an element’s inherent meaning into account, sometimes injecting additional meaning through attributes and context, and using the most appropriate element for the content it delineates.

    Tables have their own semantic meaning, which is to mark up tabular data. Calendars and train schedules are best organized in a table where the data can be sorted into related columns and rows. Tell me, please, what is the semantic value of a table used purely for visual/presentational arrangement of elements? What column of data is correctly labeled with a transparent spacer gif?

    Using table elements to trick text into taking on a visually pleasing layout is an inappropriate abuse of HTML. It’s not the most semantically correct element to use, and the only reason your text browser renders it so well is because web developers have abused HTML for far too long and browser-makers have chosen to accomodate it.

    The web is, at its core a textual medium, not a graphical one. Imagery is secondary to content, CSS rules are suggestions at best, and even the most “freakin’ awesome” design should not damage the integrity and availability of the content. More and more of us are coming to understand this every day.

    Please get out of 1997 and join us here in the new millennium where we no longer treat our meaningful content like misbehaving pixels to be beaten into a grid.

    By Craig C. on September 2, 2005 6:25 pm

23. Everyone is entitled to an opinion, the article was rather confuesing though.

    By rob on September 3, 2005 6:02 am

24. This one is mezzoblue.com, the website of one of the most influential CSS designers in the history of the Internet. Some would say it looks “freakin’ awesome”

    … why?

    Mr Shea is responsible for the creation of the CSS Zen Garden yes, however from my knowledge he’s known for little else. mezzoblue.com showcases nothing new in CSS design and there’s far better designs out there.

    By Chris Gwynne on September 3, 2005 8:24 pm

25. Mister Forbes,

    have you ever wondered why everyone understands you when you are using the word ‘computer’ to describe a computer?

    It’s because you are using semantics. In other words : you are using conventions to make sure everyone understands you. Even more : your parents and peers thaugt you that these conventions are the most efficient way to pass along a message.

    Same goes for any language, markup or not. Now, using <p class=”list-item”> is much like using the word ‘prostitute’ to describe a nun. Not only do you cause confusion, you also set a bad example for those trying to learn the language.

    That’s what the semantics debate is all about. The css-positioning-versus-tabular-layout debate is a much more specific issue.

    By Tim on September 5, 2005 4:39 am

26. Oh, man. When I first read that, I thought you were recommending trashing tabular design.

    I was just about to hire a therapist for your upcoming mid-life crisis, but then I realized you were just quoting some other poor bloke.

    I like cheese.

    By Alanna on September 6, 2005 10:39 am

27. I think with all probability, Noel it just a tosser whos complaining because he doesn’t get the point.

    He compares freebsd.org to mezzoblue.com in a text based browser, while freebsd.org shows as three cols of text, mezzoblue.com, shows as one.

    This means that there is more room in the text browser for content to fit accross the screen, on freebsd.org any long lines of code will be split and placed on the following line. While mezzoblue.com will also have the same thing happen, there is more room for the content.

    He also makes the point that the table layout requires less CSS to render the same as the semantic layout. This is a valid point, but he needs to recognise that you’re never going to change the table layout with just CSS, you will have to edit every single page of the website, and edit the CSS.

    With the semantic layout, all you really need to edit is the CSS. He also placed some of the styling of the table layout in the HTML.

    This guy just doesn’t understand the reasons for content and design seperation.

    By Rowan Lewis on September 8, 2005 10:18 pm

28. Subscribe to comments via RSS!