What are the concrete risks of using custom HTML attributes?

Question

This subject turned into a heated discussion at the office, so I\'m interested to learn what you think.

We are working on a web app that only targets some specific b

Answer 1

If the page is declared to be HTML 4 strict, then it should not add attributes that are not used in that HTML specifies. Differently, it is not clear what the browsers would behave.
As already reported, a way to add additional attributes is to add them as classes, even if that has some limitations.

Answer 2

Duplicate try this thread though: Is it alright to add custom Html attributes?

Also look at this: Non-Standard Attributes on HTML Tags. Good Thing? Bad Thing? Your Thoughts?

Answer 3

If its a goal to maintain valid html4.0 strict, then it doesn't matter why you want to put in custom attributes, you are breaking the goal.

I think the question you need to be asking, is why do you need to break 4.0 strict to get the functionality you want: Anything that you could use a custom attribute for you, you could use a in an existing attribute:

<span translationkey="someKey">...</span>

could be:

<span class="Translationkey@someKey">...</span>

it will be some extra cycles to parse all the class information, but so long as you don't put any css info on that class, it doesn't change display, doesn't put you in quirks mode, and doesn't get you in fights at work.

Answer 4

Or is it conceivable that some browsers will change to "quirks mode" and render the page as if it was something other than strict HTML 4.0?

No, bad attributes will not force a rendering mode change.

If you don't care about validation do what you like, but validation is a useful tool for detecting simple mistakes that can otherwise have you chasing around debugging. Given that there are many other perfectly good alternatives for passing data to JavaScript I prefer to use one of those, rather than forgo validation.

Plus, when you add an arbitrary attribute you are effectively playing in a global namespace. There's no guarantee that some future browser or standard won't decide to use the name ‘translationkey’ for some new feature that'll trip your script up. So if you must add attributes, give them a name that's obscure and likely to be unique, or just use the HTML5 data- prefix already.

Answer 5

HTML 5 allows custom attributes using a 'data-' prefix, see http://ejohn.org/blog/html-5-data-attributes/

Answer 6

There are no browser limitations/risks. Only the w3 validator will bark, but barking dogs doesn't bite.

The w3 spec says the following:

• If a user agent encounters an attribute it does not recognize, it should ignore the entire attribute specification (i.e., the attribute and its value).

IE will also not render in quirks mode or so as some may think. It will only do that on invalid/forced doctypes, not on invalid attributes.

However, keep in mind that some Javascript libraries/frameworks will "invisibly" add/use custom HTML attributes in the DOM tree, such as several jQuery plugins do. This way you may risk collisions in attributes because it "by a coincidence" uses an attribute with the same name as you do for your own purposes. Sadly this is often poorly or even not documented at all.