Good stuff… and agree particularly that the inconsistent and often clunky approach to validation across browsers currently means its better to suppress HTML5 validation entirely and design something that works.

I’m puzzled though about your opposition to the use of the tag. I’m seeing the slide out of context, so perhaps I misunderstand your point, but the tag has become meaningful with HTML5 where it was previously presentational (and therefore inappropriate). To mark something as ‘small print’ seems pretty useful to me, particularly to screenreaders who have no other way of telling by its visual presentation that something is legalese etc

Good presentation. Regarding sliders, are you saying using them as the ONLY means of input is not recommended, or using them at all? A slider that updates a text box would still allow you to put in exact numbers.

Joe, I read your HTML 5 book on vacation to Mexico and enjoyed it very much. My next focus has been on UX so blog posts like these are really helpful. Hoping to see more UX lessons in the future. (thanks for the free education)

I developed UIs for native/desktop applications using toolkits like Motif, VCL, MFC, AWT, Swing, wxWindows/wxWidgets, and Qt for many years. I have since been developing web UIs for the past several years, however.

Even after a few years of front end web development, I am continuously amazed at how shoddy the entire infrastructure is. HTML5 is not helping at all.

Everything about it is horrid and inefficient. The markup is obtuse. JavaScript is a terrible programming language. CSS is awkward and inefficient to use. And like your presentation shows, support for creating basic data entry forms, with elementary validation, is extremely broken.

Even using an ancient toolkit like Motif, while using a low-level language like C, were easier and more efficient to create software with than HTML5/CSS3/JavaScript. It pains me to say it, but the resulting UIs were often more user-friendly and usable than many web-based UIs today.

It’s true. HTML5, CSS3 and JavaScript are a big step backwards even from Motif, which is generally regarded as one of the least-desirable UI toolkits that achieved significant use.

Hi Joe,

Interesting presentation. Glad people like you are looking at the UX side of HTML.

I am going to take your assertion that ‘a designer who doesn’t understand how people behave is going to be no more successful than an architect who doesn’t understand physics’ a little further and say that based on some of the examples in your presentation, they do not understand the tools available to them.

For example…

The example web sites you provide has nothing to do with HTML5, but rather bad design – this would be true regardless of the tool. Better tools could be developed to better align to UI guidelines, but I believe that itself is also problematic.

I do also hate the inconsistency in the way browsers support HTML5 features and standards. However, some of these features are also not finalized i.e. there is no defined standard – some of the features are cutting-edge. The different implementation, rendering engines of each browser, etc, also makes this difficult i.e. -webkit-, -moz-, etc…

I hate * as well. But, I think you was a bit harsh with required=”required”. This is just a boolean that some JavaScript can query. In fact, this is an awesome attribute, because it better enables MVC.

I also think you were harsh with the jQuery datapicker. The ordering of the days of the month can be changed i.e. .firstDay. A webdev would have to do a lot of work to effectively internationalize this though.

So, in conclusion, designers and web devs should better understand the tools they have available to them and use a bit of design sense – I thought you illustrated this nicely.