Regular readers may have noticed a bit of a slow down in postings as I’ve been rearranging the scheduling of things here at Codescaling to allow for other commitments. Hopefully, I’ll be doing a regular Sunday catchup of what would have been snippets and during the week I should, all going well, be looking at a particular thing, be it software or hardware, thats in scope that week. As some may know, I’m curating HackWimbledon and may cover some of the hands on stuff there. But enough of plans… What’s on the catchup this week…
Google landed Go 1.3 this week and it does seem to feel quicker and slicker (I’m getting on with Go code myself and noticed the difference). The experimental support for DragonFly BSD, Plan 9 and Solaris is intriguing… Go on Plan 9 feels like a giant philosophical loop being closed. Also interesting is discussion of Go for Android from one of the Go team… it seems to be on course to start emerging in Go 1.4.
Big news in Python land where the PyPy team landed the first stable release of PyPy3. PyPy is a very compliant Python interpreter with a tracing JIT compiler built in. It had been stable only on Python 2.x but now there’s PyPy3 (libraries are at Python 3.2.5 level, unicode support from Python 3.3). At some point the Python 2.x->3.x transition logjam will be broken and this will be a big help.
Coin cells didn’t immediately strike one as a space for useful research but I was proved wrong on reading How much energy can you really get from a coin cell?, where different makes and models of cell were compared using an ARM controller which systematically loaded each battery. I’m more curious about this now as I just took delivery of PunchThrough’s Light Blue Beans, Arduino style controllers with Bluetooth and powered by a coin cell, but more about those in a future Codescaling post – till then check out the Surf Report Notifier.
The OpenSSL/Heartbleed fallout continues with Google’s latest move, BoringSSL, a bidirectional fork (the codebase’s separate but patches continue to flow in both directions – it needs a term, so bidifork) of the OpenSSL code. Google seem to be using bididforks to allow them to stay plugged into communities but retain control of their destiny; Webkit and Blink seems to be the first bidifork. Whether they work, we don’t know, but I suspect that its an area ripe for research and even formally recognising as an middle course for open source projects between fighting and forking.
On the Todo list – have a look at the Maynard/Wayland desktop on the Raspberry Pi, check out the OEM BeagleBoard Blacks, browse through the undocumented Swift standard library and now it’s a 1.0, checkout the WordPress REST API.