March 10, 2013No Comments

Moved this blog

Click away now if you're not interested in web servers.

Having spent much of the last few years making sure clients’ websites are fast and responsive, I finally got round to sorting my own blog out. It was letting the side down, being quite sluggish and unoptimised in its shared hosing environment. So I've moved this over to my own server which is running Varnish, nginx and also PHP 5.4 with APC.

WordPress, which this blog runs on, works out of the box in Apache + PHP environments but getting it running in nginx (with PHP-FPM) was a bit of a pain in the arse. This is mostly because .htaccess files are an Apache thing so don't work in nginx. nginx has its own syntax for setting up redirects and there are tools out there for converting between the two but none of the ones I found converted my existing rules successfully.

Frustratingly, following these official instructions to the letter did not work either. Getting the main WordPress installation working was fine but the problem was WP Super cache which relies on some specific rewriting logic. Certain valid post URLs were throwing up 404 errors depending on the state of the cache and what characters were in the address. At that point it had reached 1am so I decided as an interim to put Apache on the server too. I'll return to the nginx configuration sometime soon. I have other sites running off this server through nginx, it's just that WordPress-plugin-specific problem that needs solving.

Running Varnish on the server means it's now easy to route traffic to either Apache or nginx as required based on the request hostname and of course it caches the returned documents so speeds things up even more. Ideally I'd have assets going to a CDN but given that I only get a handful of visitors each day it doesn't really seem worth the effort.

Anyway, that's it for dev-ops news. Hopefully this blog will be a bit more responsive from now on.


P.S. OK, so signing in to leave comments is currently broken. This is due to my Varnish configuration. I need to make exceptions for cookies. I'm sure you'll manage in the meantime. <-- Now fixed

February 22, 2013No Comments

Bitcasa having teething troubles?

*See update at the end of this post*

Bitcasa_Logo

Having read several favourable reviews I signed up for a Bitcasa account yesterday. The desktop app for OS X looks good and works well. The user-experience is not entirely dissimilar to Dropbox in that you have a special folder which is then synchronised to their cloud storage. However the the Bitcasa drive is a mounted network drive, so doesn't take up space on your hard disk, whereas the Dropbox folder is local.

bitcasa-infinite-drive

"Infinite storage" sounded appealing and they're running a $69.99 offer for a year, so I went for it.

However... I'm now finding it incredibly slow to upload files. Speedtest.net tells me the upload bandwidth of my internet connection is 13Mbps, or ~1.88 MB/s. Bitcasa (mirroring a folder through the app) is just about sustaining 190KB/s which means it's using only around 1/10th of my available bandwidth.

Dropbox by comparison on the same machine and connection manages around 750KB/s, which is about 4 times as fast.

I've tried to get some help from them but at the moment it's like talking to a brick wall. 24 hours on they've still failed to respond to two support tickets I raised, nor have they replied to my nudge on Twitter. And there are at least a couple of threads in their community forum filling up with comments from what sound like extremely dissatisfied customers - such as this one.

I hope it's just teething trouble and they sort this out, but the apparent stony silence from the company doesn't bode well to me. As one forum user commented (here):

"The bottom line is it came out of beta too early and we're paying the price now."

Surely the first and most important thing to do is talk to their new paying customers. When people part with their cash they want service!


Update - 23 Feb

Last night I received a reply from Bitcasa on Twitter:

Also, they've responded to my ticket and have commented (and apologised) in a number of the forum threads, such as this one:

"We definitely are inundated with tickets, and are working very hard to respond to everyone as quickly as we can, in order to best address everyone's help needs."

It sounds as though they've had a huge influx of users and that's taking its toll on both customer support and the system itself. Plus there are seemingly a few bugs in the software that need ironing out. From their explanations, due to their protocol the data transfer rate is affected by the ping time to the server. They're using Amazon Web Services so are currently restricted to US, Japan and Ireland data centres.

Being in the UK I'm getting a ~30ms ping time to eu1.api.bitcasa.com, which is pretty good, and my Bitcasa upload speed seems to have vastly improved overnight without me having changed anything. I'm now getting around 800KB/sec according to their app, which is not too bad. I hope it stays that way.

I'm sure they will get things sorted out but, as ever, customer support is key. If things are going wrong 24 hours without a response can feel like a long time for a paying customer. Even to acknowledge that there's a problem and to say "we're working on it" goes a long way in these situations.

January 24, 2013No Comments

Technology is getting ahead of us

20130124-074820.jpg

The clock in the station I commute from each morning was for a few weeks displaying the wrong time. I think it was about five minutes fast. And then they stuck a handwritten Out Of Order notice over it, which remained for a further two weeks, until they finally took it away, presumably to have its network interface replaced or its firmware flashed or something along those lines. Three weeks hence it has still not been reinstalled.

I can't help think that with an older clock someone could have simply gone up a ladder and pushed the minute hand back by five minutes and that would have been the end of the matter.

Are we becoming too dependent on high-level technical systems?

November 5, 2012No Comments

Relativistic game engine from MIT Game Lab

A Slower Speed of Light is a first-person game prototype in which players navigate a 3D space while picking up orbs that reduce the speed of light in increments. Custom-built, open-source relativistic graphics code allows the speed of light in the game to approach the player's own maximum walking speed. Visual effects of special relativity gradually become apparent to the player, increasing the challenge of gameplay.

http://gamelab.mit.edu/games/a-slower-speed-of-light/

June 27, 2012No Comments

Leap Motion demo

Great demo video of Leap Motion's new device in action:

And here's the full accompanying article.

Thanks to @kerridashanne for the link!

June 27, 20124 Comments

Overshare

I'm no great fan of Facebook, though over the past few years my hatred for it has mellowed into more of a grudging acceptance of its presence. I killed my account once and didn't log in for a whole year. During that time one person actually ended up calling me to ask whether I was OK; apparently I'd "dropped off the face of the Earth". I eventually ended up coming back to it for work reasons. We get asked to make Facebook apps, and to do that you need an account.

Things were more fun before Facebook and, working (as I do) in the online marketing industry, briefs are generally far less interesting than they used to be. Almost everything is now a Facebook app of some variety, and everything else at least features "social integration" which is a pseudointellectual way of saying it has Facebook and Twitter share buttons, and maybe a comments panel. And it'll probably prefer (or sometimes only permit) single sign-on with a Facebook account. Basically creativity has by and large gone down the toilet at the behest of the planners whose measure of success is the number of Likes their client gets, which in digital adspeak translates conveniently as "the size of their community".

The worst and most idiotic crime that gets committed to this end is the insertion of what's being termed a "Like Gate": Users are forced to click "Like" on a brand's page before they're permitted to use whatever app it is they went there to use in the first place; hardly a surefire way to build up a strong and loyal fan-base I would say.

None of this is surprising and the blame doesn't lie with any individual (except perhaps for Mark Zuckerberg); it's just the business basics of misguided quick-win marketing, but hell is it bland.

Of Facebook itself, I accept that it does have benefits: Most notably, I can just contact a person without having to know whether they have changed their email address. That feels contemporary, but then I'd like to think that anyone who considers me a friend would drop me a line to let me know their new address whenever they change jobs. And anyone else, I don't really care so much if they drop off my radar.

And in terms of content, if Twitter is my constant source of personally-tuned news, industry chatter, comedy and expertly crafted witticisms, then Facebook is my Daily Mail and ITV. It's low rent material. The follow model is much more interesting than the friend model.

But, oops, none of this is what I set out to write, I'm just ranting.

What I meant to say is that something I find increasingly irritating (and it seems to have stepped up a notch in 2012) is the amount that apps are all-but forcing me to share my every interaction on Facebook. And it's now starting to include largely passive actions. This morning I went to read an article on the Guardian's website and was presented with this:

Welcome Ade, Sign into the Guardian with Facebook

...and next to it the names and avatars of some of the 62 friends who "use The Guardian".

I understand that these are desperate times for the traditional press, but I don't wish to know whether a friend has merely read an article (for that is one of the Guardian Facebook App's features), it tells me nothing. If they "Like" it then arguably that's more useful, and even better they might post a comment about the article along with a link to the original, but the simple fact that they've read it is just utter noise. Likewise, Spotify wants to actively tell the world whenever I simply listen to some music!

And this is all down, of course, to the great fallacy about social media: That companies are doing it for the benefit of the community, for their users. When I play Bejeweled Blitz on my iPhone, OK it's nice that I can compete against the two of my Facebook contacts who also play the game, but then after every single go, it wants me to share on Facebook the fact that I've had that go. No thanks.

Please, please, please share me!

And people are doing it. My Facebook news feed (or whatever it's called at the moment) is strewn with: "Mike read an article on The Guardian", "Luke 'Likes' Volkswagen", "Tom was awarded a 300k medal on Bejeweled Blitz!!!", "Matt listened to [some album] on Spotify". Yes I can opt out of these things by hiding updates from various apps or users, but really I'd rather just not opt in in the first place.

The number of sites and apps using Facebook as the default form of registration and login, and then defaulting to sharing everything that happens within that site on Facebook, and the fact that so many people are simply going along with these schemes, apparently unthinking, is to me nothing short of a slow motion car crash for the Web. The idiots are winning. Facebook, let's not forget, is a proprietary and privately owned platform and the Web was founded on quite opposing principles.

As users the power is in our hands to reject all of this. So, for as long as it remains possible I'll be reading my news anonymously if that's quite alright. And no, I won't be linking my Spotify account to Facebook, though thanks all the same.

June 11, 2012No Comments

Doing things right

MacBook battery compartment

Steve Jobs on attention to detail:

‘It was important, his father said, to craft the backs of cabinets and fences properly, even though they were hidden. "He loved doing things right. He even cared about the look of the parts you couldn't see." ’

Steve Jobs, by Walter Isaacson

May 29, 2012No Comments

How big is a billion?

mi = 1, bi =2, tri =3, quad = 4....

Great video about the differences between the "long" (old British, and arguably more logical) number system and the "short" (US) system, which was adopted officially in Britain in 1974. RIP the milliard...

March 30, 2011No Comments

Timeless

Timeless

It's been a while since I made a pointless technical demo. For most of the last 10 years I was working with Flash. Prior to that, I'd learnt how to write basic procedural JavaScript (at the same time as teaching myself HTML and CSS), but I'd never really fully got to grips with design patterns and more structured or object-based JS.

And more recently, when not busy writing emails or sitting in meetings, I've been more focused on back-end stuff: PHP frameworks, and fumbling around on Linux servers (with the help of much Googling). So my JavaScript abilities have become quite rusty.

I had the idea of making a broken 7-segment display clock after walking past one every morning, getting off my train at London bridge. There's something interesting about the unfamiliar symbols, and it's vaguely reminiscent of the self-destruct device in Predator.

A couple of weeks ago I was fortunate enough to have ten hours to kill, on a flight. I had my laptop with me so decided to have a crack at building the clock with JavaScript, CSS and a few PNGs. It took me around 4 hours, including the Photoshop bit.

This is the result. It is kind of pointless but it looks quite cool in my biased opinion. Clicking on the clock gives you the real time, which is arguably more useful.

Feel free to download the source code. Sorry it's not on GitHub, I'm not down with the cool kids yet.

UPDATED: You can find the code here on GitHub

August 21, 2008No Comments

Screengrab Remover

If you: a) work on a Mac, b) make a lot of screengrabs (I'm talking to web designer types here) and c) have a tendancy to be a bit untidy, then you might, like me, often end up with a desktop full of images called Picture 1, Picture 2 and so on. And my desktop gets pretty cluttered so I don't get round to deleting them all often enough.

Well, I was just messing about with Automator, and I know this is nothing revolutionary or anything, but I made a simple batch process for cleaning them up. And, I only just figured out that you can save your batch processes out as applications. So I thought I'd share one.

This simply finds all images on your desktop whose names start with 'Picture' and moves them to the trash. Quite a handy little 5 second saver to put in your dock...

Download Screengrab Remover

Disclaimers:

i. I take no responsibility if it goes wrong and deletes your entire world. I don't think it will though.
ii. I think it only works on OS X 10.5 (Leopard).
iii. If it doesn't work, oh well it only took 30 seconds to make.

Contact

© 2023 Ade Rowbotham Ltd