PythonJS, Adaptive Design, User Identity, Android Studio

Post date: Apr 14, 2014 2:06:19 AM

  • Perfect XKCD explanation for heartbleed bug.
  • Started Google+ Brython community.
  • Played with PythonJS Python to JavaScript translator. It's a lot faster than Brython, worst part of Brython was it's performance, because it's not optimized code.
  • Quickly tried Android Studio on virtual machine. Even if it's early access preview, it seemd to be much better than the Ubuntu HTML5 developer packet.
  • Played with a few adaptive web design CSS templates, but those were too complex to be utilized very easily, so I sticked with simpler design for dns-hosts. Responsive and fluid designs were old names for this thinking, where pages adapt to the device being used. Unfortunately it's very common to see that sites are clearly designed for some kind of fixed end device. In worst cases, which I have encountered several during this month. Site simply says get lost, you're using wrong kind of device. Without even allowing user to select different kind of layout. I often encounter this when forwarding links from desktop to mobile or from mobile to desktop. It seems that many Finnish news services are run by totally incompetent IT staff. "You've reached our mobile app, but not from a mobile browser." - What do I or you care about the browser I'm using? All I wanted to see was about 3 kilobytes of news article text. Does the browser really matter so much. I would say no, it doesn't matter at all. But that seems to be too complex thing to the "web designers" to understand. Just give me the raw text, without any HTML as pure ASCII, if you're even capable of doing that. Domains related to this whining are: talentum.fi and tietoviikko.fi, they just fail so miserably.
    • Anyway, adaptive design is great feature for Content Management System (CMS), where it's easy to produce multiple different designs from same source data depending from required situation.
  • Wondered once again, what is the point of using email address as user identity. It's inherently very bad malpractice, yet many sites do so. Email address, telephone number, etc, can be anything. And shouldn't be considered as ultimate 'unique identifier for user'. I very much like concept where user id is just plain user id, and it's something that system generated. Like it's with Facebook. I'm saying this even I hate Facebook, but that's something they got really right. Even SQRL uses email address as ultimate user identity. It's just silly. I can change my email address daily, as well as my burner phone number. It doesn't mean that it wouldn't be me anymore. As well as the password recovery using email address is once again, one very very bad idea, but so widely used.
  • Short list of checked stuff: Amazon CloudFront + Wikipedia, Webhook (nothing new, just one way of triggering stuff), Internet in China, CNGI, IPv6 deployment in China, The Wolf Hunters of Wall Street, Riak, Redis, LevelDB, lmctfy,