Kids these days with their portable smoke machines. #jamesblake (at The Music Hall)
Went to Costa Rica after the exams, both the Pacific and the Caribbean coasts. The latter was decidedly superior, although the trip took much longer due to less developer curvey roads. Here are a bunch of pictures of beaches’n’boats’n’bums.
Going through the Polaroid stash. Need to order some 600 or 779 before its too late :/
Web apps in the browser are a frustrating experience. There are many usability problems (on the desktop, anyway) that prevent web apps from joining the ranks of native applications, e.g. a lack of a standard platform UI, unpredictable and incomplete support for keyboard shortcuts, and navigation that necessitates using the back browser button.
However, I believe the biggest problem with web apps on the desktop is the lack of a change of context. Switching between tabs or windows does not result in the same conceptual switch as switching between native apps, by virtue of all web apps running as part of the browser process(es). As a result, when working with multiple web apps, one has to constantly stay aware of which window or tab the corresponding app is running in.
One of the solutions is to open each web app in a new browser window, but this breaks down when using Alt-Tab and other application switching schemes since all browser windows look the same. Another solution is something like Fluid, which is the domain of the more advanced users, and requires installing an application. Finally, there are attempts to integrate web apps into the desktop, such as those in Unity on Ubuntu.
This solves the context switch usability issue: it conceptually places the web app alongside other native applications, and makes it immediately obvious which application the user switches to when using Alt-Tab, Exposé etc.
While the solution isn’t perfect, I believe it’s low-hanging fruit when it comes to improving the largely non-existent integration of web apps into the operating systems of today. Comments are highly appreciated.
The downside of storing data in an API is the lack of any established domain-specific data protocols; a problem that never existed for files. Ironically enough, RSS is one of the few agreed upon web formats, which is now hiding behind the Reader API.
WebKit is the only browser implementing the File API, which Google is pitching to W3C. Has anybody else noticed how unsightly it is?
PERSISTENTconstants and other methods are all floating in the global namespace. I can see why Mozilla is promoting the IndexedDB API instead.
Went to South Western U.S. in January. Just picked up the rolls today. Scanner/VueScan decided to freak out and only let me scan 2 negatives at a time, and now I want a dSLR even more.