Hyphen, namespace or none of the above?! about mdoml

[jslegers]

See also http://lists.w3.org/Archives/Public/public-webapps/2015JanMar/0003.html

[jslegers]

An overview of use cases where IMO non-hyphenated custom elements have great added value :

• Polyfilling, where only modern browsers offer support : I've considered it for polyfilling the <dialog> element.
• Polyfilling bleeding edge specs : It would also allow both CSS frameworks or customer projects to already implement elements defined by future HTML specs before there's any official browser support.
• CSS framework innovation : CSS frameworks like Bootstrap and Foundation are popular enough to actually set a standard that browsers are unlikely to break in the future. In fact, it would allow CSS frameworks to experiment with new standards that can be adopted by browsers at a later stage without the need for polyfills.
• REALLY custom stuff : Tags like <car>, <friend>, <hobby>, <swag> or even <chatbox> are unlikely to become part of the HTML standard because they're too specific, too non-technical, etc. If I control which third party code I use in my project, such tags should be considered safe.

An overview of use cases where I would prefer hyphenated custom elements :

• Tags like <combobox>, in projects other than CSS frameworks. This, because they are likely to break future specs or result in conflicts with third party code.
• When standards for CSS frameworks do not exist, it might be a good idea to start with something like <bootstrap-combobox> and organicly evolve towards <combobox> if there is popular support for any particular implementation. This would allow CSS frameworks to use development strategy for custom elements similar to the CSS prefix approach used by browsers.

[mdo]

You seem to share my opinion on this matter, as none of your custom elements use namespaces and some of your custom elements do not contain any hyphens. Why did you make that decision and how do you intend to approach any issues involving code validity or spec compliance?

This is an experiment, so whatever happens can be easily dealt with. I chose to rock no dashes because it felt right for a longer term vision of how HTML and frameworks could evolve. Code validity isn't an issue here for me right now—there are a few things that don'e validate in Bootstrap as well (browser hacks).

I might re-evaluate that later, but for now this seems perfectly okay to do to demonstrate the point.

[jslegers]

I chose to rock no dashes because it felt right for a longer term vision of how HTML and frameworks could evolve.

We're on the same page here. Unfortunately, the idea won't be able to move past that stage unless it's picked up by the W3C.

If you'd like to help push this idea through at the W3C, here's our chance. Do you have any experience with their mailing lists?