Niall’s virtual diary archives – Monday 19th March 2012

by . Last updated .

Monday 19th March 2012: 5.00pm. Wow, a full six months between diary entries! Unfortunately it's been a combination of both being incredibly busy and not having a massive amount of anything interesting to say which has been the cause of my tardiness. Much of September went on getting the World Economics Association launched, in particular upon writing lengthy email replies, and getting all editions of the Freeing Growth manifesto into print. In October I was appointed as ISO SC22 mirror committee convenor for Ireland by the NSAI's ICTSCC committee which is the mirror of ISO JTC1. In November I was contracted by the WEA to write a virus scanning plugin for the PKP Open Journal Systems framework they use for their journal management (and which meant I had to touch PHP again, which is always unpleasant), and my copy editing of that book manuscript on Economics I mentioned last entry turned into a co-authorship of a book with Handelsblatt correspondent Norbert Häring called Economists and the Powerful: Convenient Theories, Distorted Facts, Ample Rewards. That book manuscript then sucked up pretty much all my free time until it was delivered to the publisher mid-February. In between came an avalanche of IT contract work, so much so that for the first time in a long time I had to actually turn paying work down which always sucks royally.

Since mid-February, I have finally begun to get around to a major systems upgrade of our computer systems. Before the end of its tax and accounting year in January, as normal my company had ordered its annual hardware parts which this time included a bottom end Sandy Bridge server which was amazingly cheap for what it is, despite having an 80 PLUS Gold power supply (>=90% efficiency) and therefore sips just 28W which isn't much more than my old Intel Atom 220 internet gateway despite being umpteen times more powerful. My plan is to solve a long-running headache in my server deployment: right now, there are three hand configured servers on the public internet which use a manual rsync for backup onto what is effectively Megan's television as it switches itself on when used, and off after a timeout. As time has gone along, that television has ended up storing not just backups, but all the shared stuff between my computers, all the GIT repos, all my company stuff etc. Yet it runs pretty much without backup - every now and then I dump copies of things onto an external 2Tb USB hard drive, but overall this setup - while somewhat secure and flexible - is extremely manual, and very time consuming. I also have the current problem that only my main workstation has a copy of VirtualBox on it for temporary OS deployments for testing, and that gets hand configured each time too. If I want to do a quick server config change, right now I have to chance my arm, take down my services or do a lengthy rsync to VirtualBox, distill a set of upgrade steps and repeat those on the live server as my ADSL outbound is too slow to upload images. All in all, none of this is ideal. For example, some months ago I had a major outage of all my websites and email due to a late night misspelling in /etc/network/interfaces which required a technican intervention as the server was no longer able to boot. That's expensive, never mind inconvenient. The fact that email, web and everything else is all the same server is particularly unhelpful.

What I really want is a cloud platform with virtualised OS instances, so to test you simply clone a running instance and employ a whole instance per service so each service stands alone and independent from the others. I also want to be able to deploy temporary OS installs much more quickly, so instead of always having to run test suites on my workstation using a fake localised config in VirtualBox, I can instead run them against a real server which could at any moment be deployed onto the real internet. A cloud platform lets you run a proper automated backup solution which auto-syncs parts of itself with the other servers as an off-site backup of the really important stuff, plus runs anti-bitrotting sweeps on the really long-lived data. That means that the public servers automatically are backed up locally, and the local servers are automatically backed up remotely. Furthermore, instead of almost all my big and important data living on an external USB hard drive which needs to be plugged in for me to use it (I don't leave it permanently plugged in, lest it get deleted or damaged), I could have all that data on demand secured with RAID redundancy and a snapshotting filing system like ZFS or BTRFS so no one can delete anything whether accidentally or otherwise. That alone would be extremely useful.

Anyway, I've got that cloud infrastructure working and here are the instructions on how to do it, but I haven't finished data replication yet. The slow pace is once again due to lack of free time, but also because the cloud stuff isn't a priority until Ubuntu 12.04 LTS has gone gold as most of my deployments are Ubuntu Server LTS, so with the imminent release of 12.04 there seems little reason to rush. In the meantime, I have been investing a lot of effort into a once and for all distributed bug tracking solution which lets you coalesce multiple sources of issue tracking into your GIT repo via both a RESTful HTTP API as well as a programmatic API. This subproject is called BEXML, and it's an extension of my BEurtle issue tracking GUI though it's also a fully standalone library. Getting this working - and I'm not far from finished - would be a major boon to not just myself personally who has to handle issues arriving via email, public bug trackers as well as privately when I spot a problem and need to remember to fix it some day, but also to several major open source software projects and those who run and coordinate those projects as they face exactly the same management problem as me.

Adding to the demands on my time are taking two university distance courses at the same time: the PGCert, which please God will be over this summer, as myself and the University of London/The Institute of Education have definitely parted ways after their never ending dreadful customer service (they simply don't care about the student experience [they go through the motions, but it's all hot air promises and nothing ever changes], and every staff member always says "it's not MY fault/responsibility". So if no one will ever take ownership of any problem not strictly within their personal remit, no wonder there's such appalling customer service!). I'm also taking that OU Pure Maths course which is surprisingly fun with all sorts of abstract puzzle solving, and as much as Open University courses are pure spoon feeding and regurgitation I am certainly not complaining in Pure Maths where spoon feeding is just fine by me! What irks me in courses such as Education is that I don't see why my well argued, supported and referenced opinion isn't as worthwhile as most of exactly the same within the Educational literature. I'm paying for the course, so my academic arguments ought to have equal initial standing with other non-peer reviewed arguments (they may well, and indeed probably will falter after analysis, but the point is that argument is argument). What happens instead is a NIH response, so if it's Not Invented Here then it's obviously no good, especially if the so called "educational expert" has never heard of anything outside their field and are too bloody lazy or ignorant to bother doing a god damn Google search before deciding that if they haven't seen it before, it must be lies. This is not what one would expect from the Institute of Education, one of the premier educational research institutions in the the world, but there you go. I, as the paying customer, am not paying £64 per ECTS credit (about €75 per ECTS, expensive by European standards) to be told my arguments are worthless relative to the course material - especially when the course material is constantly banging on about how student's sociocultural understandings are to be treated as valid-in-themselves and not to be dismissed out of hand just they they're doing with my arguments. I dislike hypocracy at the best of times, but I especially dislike being discrimated against for pointing their hypocracy out when I'm the one paying their wages. So definitely, the sooner I can get away from those robbing bastards the better, because what they advertised for their course has little to do with its practice and if UK consumer law applied to universities I'd submit a product mislabelling complaint (it doesn't, so you have no redress except via the quality regulator). Caveat emptor I guess. I just wish it hadn't cost me two thousand pounds to find out I've been conned.

So that's basically what I've been up to during the past six months. I had to can the Oxydérkeia project sadly - without a set of students to test it upon, and now it won't be part of my Masters in Research thesis with the IoE, rationally speaking it had to get chopped. It's a real shame though - that technology would have been extremely useful in a multitude of tasks, in everything from evaluating prospective new employees down to time and motion studies of computer using employees. However, the Institute of Education won't play ball and seem mired in their own navel gazing and ivory tower as their precious government funding gets cut (they have very little experience gaining research funding from industry), and I can't see recruitment agencies paying for its development given how unbelievably technologically backward recruitment is (it's because, of course, technology will eventually make much of their utility obsolete and they know it). Better then that I focus on other more productive-to-me uses of my time - hence my focus on solving that issue tracking problem as even if nobody else uses it, it would be bloody handy for me personally.

What as to the near future? In the immediate future, finishing what I started above is the plan. In the slightly longer term, chances are good that me and Megan are leaving Ireland for economically greener pastures. As much as I'm busy and doing things like turning down paying work much to my chagrin, Megan won't see a salaried job that has anything to do with teaching any time in the next five years. If we're going to have kids, we need a stable income. Ireland can't provide that for both of us, nor will she anytime soon and neither can most of Europe in its present state. So since the start of the US H1B visa season (March) I've started applying for jobs in the US and to a much lesser extent, Canada. So far interest has been very good - to date more than eight out of ten applications I've made have resulted in telephone interviews, and it makes last year when I was applying within the UK and Europe look dreadful in comparison. Surely one of these ought to result in a visa sponsorship, and off to (probably) Silicon Valley it'll be for us. If I get a window of at least a month this summer before we emigrate, I'd like to start a second Freeing Growth mini-book this time on the natures of growth, so asking things like What is growth? What forms does it take and how can these be measured using today's tools? Very 19th century, I know, but amazingly I can't find such open ended questioning since Jevons in the 19th century believe it or not. We're well overdue an update given oh, the rise of computer analysis, the development of new statistical mathematics and such. It's about knowing what we don't know.

Almost certainly it won't be another six months till the next entry! I'm playing Mass Effect 3 in the rare occasion I have the time, and I understand from the internet that the endings are shockingly bad with over 60,000 votes in a poll on the Bioware forums and over 40,000 votes in a Facebook campaign page calling for the endings to be fixed to something like what had been repeatedly promised by Bioware since the start of the trilogy. From my own perspective, so far into the game ME3 looks rushed: there are as many graphical, camera and gameplay glitches as there were in ME1, they stupidly made a core team member into a paid first day DLC because they clearly had nothing else to hand, they haven't bothered with most of the interactive person-to-person conversations they had in ME2, and I'm struggling to understand how they came up with this storyline given the events in ME2 as there appears to be a missing bit in between the two. I know that many of the lead storywriters quit Bioware about a year ago, and there has been a general talent exodus since Bioware were bought out by EA, never mind some dreadful ME book releases since that talent exodus one of which had to be recalled it was so bad. ME3 looks just as you would expect if the core talent left or stopped trying half way through development. The endings, I understand, appear to have been bolted onto the end of an unfinished story arc, so basically you're nearly there storywise and suddenly you get some long cutscene, get given three choices each of which chooses pretty much the same ending and supposedly that's the trilogy into which one has invested a hundred hours of your free time over four years done?

So expect, once I'm finished, to express my bitter disappointment here. Mass Effect 2 is the only game I've bothered to play more than once through since the 1990s, and I can count on one hand those games which I have ever played through more than once. The fact that all those decisions I took would have absolutely zero effect on the final outcome is a travesty, and even a company well known to hate customers such as EA must surely realise that these sorts of asshole move will impact profitability (which is the sole thing that EA management understand, let alone perceive). It's real unfortunate - what a lost opportunity so close to the finish line! Hardly the first time in human history to happen due to dreadful management though. So, till that next entry, be happy!

Go back to the archive index Go back to the latest entries

Contact the webmaster: Niall Douglas @ webmaster2<at symbol> (Last updated: 2012-03-19 17:00:00 +0000 UTC)