Just updated - it's quite noticeably faster, and feels much more native than their previous shortcut of lots of UIWebViews. I commend the Facebook iOS team! This app has regained its throne as the model of iOS UX.
Let this be a lesson to us all: when putting user experience as the first priority, the nirvana of writing your UI once in HTML and having it work universally still isn't there.
Are you sure? It's better, yes, but it's not a model of UX and on an iPad it seems not much thought was taken into making it a truly universal app. There's a lot of unused space and everything is too large.
The Google+ app with its combo of Path and Flipboard is a much better model to follow. Fast, and it takes advantage of the screen size in both versions.
> This app has regained its throne as the model of iOS UX.
I wouldn't go this far. Sure, the CoreData backend and table view controller implementation is beautiful and awesome. However, the root level navigation could be improved. The pseudo-popovers for messages, notifications, and friend requests is inconsistent with the rest of iOS, even though it might be consistent with past versions of the application and the versions of Facebook on other platforms.
> Let this be a lesson to us all: when putting user experience as the first priority, the nirvana of writing your UI once in HTML and having it work universally still isn't there.
And it will never be. HTML just sucks for application development.
Why on earth exchange the benefits of using the hardware resources properly for the mess that HTML+CSS+JavaScript is? To mess the customer's experience?
Don't be afraid of touching C, C++ or any other language with native code compilers and at most rewrite the UI part. Or shell out some money for tools like Mono or similar.
Although I am forced to code Web applications at work, I am a firm believer that web should only be used for communication protocols and documents as such.
I love the energy itself that people are putting into making the web more of an app development platform. It's commendable how strongly and energetically they push to get there, building cool demos and frameworks that on a relative scale (i.e. from the web of ten years ago to the web of today) seem magical. But, frankly, more and more I feel like you, and more and more all that effort seems wasted on bringing everything up to par with old desktop patterns and techniques when we could all instead push everything somewhere new.
Like you I write web apps, but I'm old enough/lucky enough to have had some other experiences, and to be quite honest I find even Gtk+ is more pleasant overall than the MV* framework of the month. That we have folks now trying to write AutoCAD or Photoshop in JavaScript just like a loss for our industry.
Let this be a lesson to us all: when putting user experience as the first priority, the nirvana of writing your UI once in HTML and having it work universally still isn't there.