Monday 28th May 2012: 7.13pm. You've probably noticed that I've wired in my Google Plus feed to an iframe on the right - I finally got round to configuring the very useful IFTTT to auto-replicate my Google Plus posts to Facebook and Twitter, so I figured why not have them appear here too? Yeah, I guess I'm about a decade behind everyone else there in getting my content to replicate around, but well to be honest I didn't have a need to spam people with my inanity, so I didn't bother. So why the sudden change of heart? Well, the big news is that I'll be relocating to Canada end of this summer to join Research in Motion's Native SDK team as a Senior Software Engineer in their Waterloo, Ontario HQ. They executed a nearly flawless recruitment process over 20+ hours of telephone and video interviewing, and if their business execution is anything like as excellent as their recruitment was then they're due for a huge bounce back if they can make it past running out of cash as they are forecast to do by the end of this year - when no doubt investors will start calling for them to be broken up rather than take on oodles of debt. I figured that as a result of the move and the new job I'll no doubt have a whole load more inanity to spout with no time to do longer updates to this virtual diary, plus people might be temporarily interested in shorter posts to Google Plus for a short time at least, so I went ahead and wired all my data feeds together so you all can keep up if you really, really want to.
On other news, my shiny new cloud infrastructure is up and running beautifully, including secure off-site automated data replication and automated download queuing and management all of which takes care of itself each night after we've gone to bed, and shuts itself down before we wake up while sending me emails daily with its progress so I know it's working okay. I'm still waiting for hard drive prices to drop considerably (see next paragraph) before implementing my RAID6 auto-bitrot healing solution for our very long lived data (and, indeed, I'm also waiting for RAID6 support to enter BTRFS mainline), but in fairness the six virtual machines spread across two hardware nodes which now operate both the ned Productions Limited infrastructure and everything in the house including the ADSL connection all work swimmingly. And I've cut our household baseline power consumption by another 50W despite the much improved, demand-on multi-terabyte shared data solution, so it's been a massive win all round!
So, how do you decide when to buy new hard drive and/or flash storage? A few years ago I did some primary research for Freeing Growth for a section on information storage trends and it's proved useful to keep that data up to date, so here's magnetic and hard drive storage capacities per inflation adjusted dollar from 1980 to April 2012:
I used the Generalised Logistic Function (also known as Richards' Curve) as the model with least-squares monthly average fitting which isn't a bad model for this sort of thing. I didn't go nuts on the regression, so don't expect Summer 2018 to be when flash will definitely catch up with magnetic. In fact, Freeing Growth (written in 2008) originally predicted 2013 as the crossover point, but what's happened is that magnetic has had a sudden extra growth spurt thanks to perpendicular recording while flash has slowed down its rate of improvement, so the catch up has run right out much later. In fact, if higher density flash is as unreliable as they think (Grupp et. al, 2012, The Bleak Future of NAND Flash Memory), the faded out logistic curve I regressed for flash storage is much more likely i.e. there will never be parity between magnetic and flash. Food for thought! If you want the raw data including its sources, I keep an Excel spreadsheet holding the full works here.
My PGCert exams start in just over a week, so my next two weeks are basically gone on finally ridding myself of the Institute of Education and the University of London. After that comes my OU Pure Maths coursework due just the week after, then I'll no doubt have to do my annual visit to Northern Ireland with my sister who is getting married next week. So, basically I'm very occupied with no freedom until the end of June, however after that I have a very exciting last summer in Europe planned. Firstly, I'm going to get my BEurtle issue tracking GUI out the door - it has the beginnings of a Redmine backend working, so I just need to add a Github backend, wire it all together and voilá, that's another major productivity improvement in my life achieved! Secondly, I'm going to make myself read all 1,100+ pages of the newly updated The C++ Standard Library: A Tutorial and Reference (2nd Edition) from cover to cover - thanks to my ISO standards work, I'm fairly familiar with the language changes in C++11 but I can't admit the same for the C++11 STL where I feel myself woefully underinformed about the new facilities (even though no compiler supports more than a subset of the C++11 STL at present). Thirdly, I'm going to finish writing up and submitting my POSIX standard changes and my C language standard changes before I have to resign my SC22 convenorship. And fourthly, I'm going to go visit as many people around Europe as is practical to say goodbye. Sadly, another installment of my Freeing Growth book series looks unlikely - not enough time, unless my Canadian sponsored work visa gets delayed significantly, as we'll have to start packing up and selling off our stuff end of August.
Last item this post: my final review of Mass Effect 3. I'm going to be lazy and copy and paste my Amazon review of ME3:
I have very high standards for computer games - most of them I don't play past a few hours as they're a waste of my time. For example, I got bored with Half Life 2 (yeah, sacrilege I know). I got bored with Max Payne and especially Max Payne 2. Couldn't be arsed with Call of Duty or anything like that. Liked Portal 1 a lot though (not so much Portal 2, it was tedious). Liked Batman Arkham Asylum. Liked Chronicles of Riddick too. GTA4 wasn't bad, the beauty of Liberty City made up for a lot of other problems. Put up with Bulletstorm which was unusual by being annoying without being tedious or boring (driving the dinosaur was fantastic, but I digress, this is a ME3 review ...)
Let me put my standards another way: I haven't played a game through more than once since Duke Nukem 3D back in 1996. No, I am not kidding, no game since 1996 was good enough. Until, that is, Mass Effect 2 which I played through no less than THREE times. It wasn't that ME2 excelled at anything in particular, rather it was just really well executed across the board with a terrifically balanced insanity mode. Lots of attention to detail, lots of variety in destinations, conversations and adversaries. ME2 felt like my own personal movie. Its only real shortcomings were lack of ability to fight with the ship in the suicide mission, and the ending was a bit underwhelming.
That made me want to play ME1. Yeah, ME1. ME1 was really rough around the edges. Got very bored in the Mako with its stupidly slow cannon and stupidly slow shield recharge. Got very annoyed by constant bugs, many of them showstopper bad, and *particularly* the jerky animation when talking to people. Disliked the stupid level designs and constant darkness. Got very bored having to run around or drive around the maps for like forever. Also found the graphics crude and ugly compared to ME2. In fact, ME1 was just crude in general - but, what it really had going for it was *moments* of spectacularness such as first contact with the Prothean beacon, meeting the Prothean VI, or indeed everything from meeting that VI onwards when ME1 suddenly became very ME2-like including nice bright, pretty level design. ME1 had a cracking ending though, best of the series by far. All in all, you got the feeling that they were really *trying* in ME1, successfully learned from their mistakes in ME2 and especially with truly superb ME2 DLC like "Lair of the Shadow Broker", ME3 was surely going to be great.
And then, you get to play ME3. It's not that ME3 is a bad game - it's like 70% of a great game. It's just that missing 30% is so terribly important. ME3 has most of the polish of ME2 and the best weapon loadout system of the series. I also don't mind the in-your-face urgency in ME3, that's appropriate. It also successfully ties up most of the loose ends from the previous two in a satisfactory fashion. It has, like ME1, some spectacularly good moments - the whole level leading up to the mother of thresher maws on Tuchanka I thought very well executed. But there's a huge difference in ME3 over ME1 - ME1 was trying its best. ME3 is just unfinished at best, lazy at worst.
ME3 is what happens when people do cost-benefit analyses to art.
I could go on at some length on exactly what's wrong with ME3, but there's no point. What ME3 looks and feels like and surely is is a game where EA management told them to deliver in six months and drew a hard line under that date while pulling off staff to other games. Some parts of ME3 are finished e.g. the random conversations in the crowd. Others are woefully unfinished e.g. there is a huge gap between the end of ME2 and start of ME3, and don't get me started on the fob off crap that is everything after you defeat Kai Leng. Another thing which really bugs me is that some of the conversations have proper film style camera angled conversation trees a la ME1 and ME2, but most are literally just pressing play on a random non sequiter one of three recorded speech options with zero interaction. That was just very lazy of Bioware/EA, and it destroys any personal relationship you have with any of the characters. In combat, your foes are all almost identical, either Cerberus or husks which gets boring quick. Romance was also much shallower in ME3 than ME2 and ME1, in fact just about everything was shallow.
Bioware/EA ought to have finished the camera angling properly and done out conversation trees. They ought to have continued the game after Kai Leng instead of cobbling together some half baked nonsense and passing it off as an ending. I don't mind how they ended it, I DO mind how they implemented the ending. I want to feel my choices in 100+ hours of play led uncontroversially to one of their three possible endings. I don't mind just three options - though the sixteen completely different endings they promised would be much better - but you need another six hours of gameplay in there to take us from defeating Kai Leng to where Shepard has tried everything he/she could to avoid that final ending based on the choices he/she made throughout the series, but he/she accepts their fate when NO OTHER OPTION REMAINS. Preferably after you've killed all your friends (paragon) or after you've murdered millions (renegade) trying to avoid your fate.
Unfortunately, because EA can only see profit, we won't get a finished ME3 with this summer's improved ending patch. We'll just get some cinematics. A real shame. Mass Effect could have been an outstanding trilogy. As it stands, I'd suffer ME1 again before I'd play ME3 again. In short: I can't tell you not to buy ME3 because I know you'll have to. But be prepared to feel empty and cheated after you've finished it, just like when the Sopranos faded to black rather than giving us a decent ending like The Shield did. And remember that feeling next time you think of EA, or when you next hear of a soon-to-be-formerly-great studio like Bioware getting bought by EA. I'm not angry any more. Just sad at the opportunity wasted.
Yeah, pretty bitter I know. That said, the Facebook campaign page is now above 66,000 likes up from 40,000 or so last entry. A lot of very annoyed customers - indeed, they got EA voted "worst company in the US" for 2012 which is quite something when US banks are so reviled, and EA were obviously a bit pissy about the accolade in their official response. However, as I said in my Amazon review, until it affects their profits their management couldn't give a sod. Customers are there to be squeezed for every penny possible, and that's all EA understands.
Well, I think that's about it for the time being. Last three months were so boring there is nothing to report from them - it was nothing but study, coursework, job interviews, applying to jobs and migrating to the new cloud infrastructure. Quite literally that was the past three months. Amazing how time can pass when none of it is free! Until next time, be happy!