Subscribe to RSS feedIs it wrong to use conditional comments?
By Martin / (X)HTML & CSS / Thursday July 3rd, 2008 / 2 comments
Conditional comments are specially formatted comments not normally parsed by HTML browsers that allows you to have certain sections of your code read by certain browsers or not. Importantly, though, conditional comments only work in IE!
<!--[if IE]>
IE only content or tags.
<![endif]-->
However, is it wrong to use these tags?
Personally, I believe so. Primarily, because they go against web standards. The HTML 4.01 specification states that "Information that appears between comments has no special meaning" - so no matter what you put between the it shouldn't change in any way the appearance of your page. Conditional comments, however, do - against the specification.
Beside this, I am of the opinion that conditional comments lead to bad coding practices. It should never be impossible to get a web document to look the same in all browsers without the use of browser-specific code, and using such code as conditional comments is a lazy opt-out.
I have also seen comments, which I believe are true, that conditional comments violate the goal of separation bewteen structure and presentation in your markup. Conditional comments are not used for structure, they are most of the time used to change the way your page appears in IE and other browsers. So if different code must really be used for IE, then it should be done through the stylesheet or scripting - not through conditional comments.
Finally, they are proprietory tags, not supposed by the majority of browsers on the market today. By supporting them, we continue to send a note of acceptance to Microsoft that we continue to work by their rules, and give less impetus for them to create proper standards-compliant browsers.
So let us look to a brighter future, and not get bogged down in the murky past of conditional comments.
∎
I hope you enjoyed reading this article. If you wish, you may view some of the other recent or popular things I have written, or subscribe to receive RSS updates. You can also add a comment, or share this article on Twitter or Facebook, below.
2 comments
n00b Then Conditional Comments are not the problem. The real problem is the HTML 4.01 definition about HTML comments in general.
A best practice in CSS is to comment as much as you can your declarations. It shouldn't be any different with the HTML. If you use comments to define your page's sections, then that's a VERY good practice.
On the other hand... is HTML 4.01 has that, in my opinion, stupid definition about HTML comments, then why where HTML comments invented and implemented after all?... Because they make sense because you can use them to describe your HTML code
Conditional Comments are based on regular HTML comments' syntax/structure. The fact that MS decided to 'play' the concept of HTML comments and create what they call Conditional Comments, it doesn't really have anything to do with if it's wrong or not to use Conditional Comments.
If Conditional Comments were used only in PHP, or only in .NET, or only by Oracle DB developers, then it would be considered 'valid' or 'correct' to use them.
Conditional Comments are as valid and it's totally fine to use them as it is to use regular HTML comment in your document, it helps the designer/developer have control over the document.
There's no problem and it's TOTALLY valid and acceptable to use Conditional Comments in HTML as it is to use the * html in CSS.
boen_robot I'd have to disagree with n00b about the "* html" CSS hack in the same sence with which I disagree with advocacy of using any browser exclusive CSS hacks*.
However, I do agree that CCs are not only perfectly valid, but also right to be used. IE is buggy, and we need a way to deal with that, period. CCs are a valid and exclusive way of doing so, making them the perfect solution. The best part is that since it's an HTML extension, you can use it to add/remove/alter not just HTML, but also CSS and JS.
As for being in contrast with the specification - where did the specification leave any way for implementations to allow implementation defined markup? I see no way in HTML 4.01, and XHTML too (all versions) for that matter. Some newer languages (like XProc; a draft at the time of this writing) have such elements (<p:pipeinfo> in XProc's case), but that doesn't solve the problem for older languages.
In other words - the HTML standard is broken in the fact that it expects all implementations to implement it and it's acompaning standards flawlessly, and doesn't provide any hooks for extensions and/or bug fixers.
* For me, there are also "browser inclusive CSS hacks". For example, IE7- position elements off when you set "position:absolute". All other browsers (IE8 included) place it at left:0;top:0. For IE7-'s sake, I'd specify those explicitly instead of using CCs. If fixing this problem required something more though, I'd rather use CCs instead.