I went to yet another conference last week, but this one was a little different. For one thing it wasn’t a web conference! Mostly not. At Geekyconf there were a load of talks on a load of subjects as well as games and food. It was more about delight than take-homes and it’s hard to draw together a writeup, so I’ll just point to a few pics I’ve put on my Flickr and touch on some highlights:

Expectedly good:

  • Sex, Love, Dwarves, Bacon
    Dr Esther MacCallum-Stewart
    About “Love” or emotional attachment in computer games, which is hard to manufacture, basically, giving odd abstractions like how the affections of most NPCs can be bought with the right bribe.
  • The Wonders of Stuff
    Zoe Laughlin
    Materials science. Invisible balls, gecko tape that only sticks to smooth surfaces, regenerating concrete, optic concrete; SCIENCE!
  • Scandinavians hitting you with sticks
    James Wallis
    Larping, like tabletop RPG, is something I like to think about but could never provide the level of commitment needed, so it’s nice to hear other people talking about it so knowledgably.
  • How to turn Drunk Ideas into Books
    Mr. Bingo
    That guy with the nasty postcards. Seen a similar version of this talk a couple of years ago but it’s still funny.

Unexpectedly good:

  • Omnibus Racing for Beginners
    Simon Abernethy
    Did you know there used to be such a thing as pirate buses? And because individual buses competed, they used to race to stops? RACING PIRATE BUSES.
  • Determining the fastest person on two wheels
    Edd Sowden
    I begin to switch off when cycling’s brought up as I can’t even ride a bike! But Edd gave an excellent collection of nerdisms and infographics.
  • On The Counting Of Sneezes
    Peter Fletcher
    Hillariously deadpan talk on a frivilous exercise that boiled down to a few profound reflections.
  • Five Facts About Smell
    Alice Bartlett
    Fun, funny, interesting. Explaining how smell works. The answer? NOBODY KNOWS.

It was great, so I was surprised that the attendance seemed to be pretty low. Maybe the vague premise of geekdom is too hard a sell? The (surprisingly lovely) hall could have held double or triple the attendees and looking at their website some speakers, sponsors and caterers must have pulled out or been downsized, leaving when MAC met CHEESE to feed everyone and Giddy Up Coffee to dole out continuous coffee. Maybe it wasn’t a bad thing though as both were really excellent (and free for attendees).

Bottom line is I was in dev / gamer / geek / spicy-mac-n-cheese / deep-fried-oreos / good-coffee-on-tap heaven and I hope it happens again next year!

Full Frontal


Earlier this month was Full Frontal, a javascript web conference in Brighton. It’s been going for a few years but this is the first I’d attended. Though I’m not as hardcore into JS as many front end devs, I actually found it much more accessible than the jQuery UK conf earlier this year. Basically it was excellent, with good coffee, fab freebies, a nice vibe and above all excellent talks.

ES6 UNCENSORED by Angus Croll

I’m not so much a programmer so I didn’t get the fine points of this, but thanks to clear examples it was easy to see that Angus was excited about the upcoming ES6 JS upgrade enabling terser syntax for a few common tasks, and I do love me some dry code. Honestly though my main take home was that he gave a great talk around a stutter, which made my public speaking anxieties seem awkwardly trivial.


Robots and rabbits and live demos galore! Andrew builds and programs robots to interact with his super cute rabbit. It didn’t think much of the laser pen robot but was pretty keen on the automated food pellet dispenser. Things got real when he fired up the quad copter controlled with an xbox controller, got it send back a video feed of the audience, and then used facial recognition technology to replace everyone’s faces with troll face. Awesome. Oh, right, yeah and it was all done through Node, which is cool. Soon we’ll be running drones from our browsers!


Seems a conference isn’t complete these days without a talk on the importance of mobile, and this one had all the standard impressive charts and stats about the growth of mobile. Joe took it a bit further however, talking about the extended mobile ecosystem from tablets to wearable devices wearables without a screen, and even cars. It tickled me to think of a “mobile device” being so large but apparently cars with built in connectivity is already fairly standard and becoming more so.


This talk was right on the money as I’ve been pretty reluctant to use JS on mobile sites due to performance issues, but apparently mobiles being slow to parse JS isn’t an issue anymore! It was also nice to hear that we shouldn’t waste time modularising as I’ve never gotten on with Require. Lots of easily digestible takehomes, great stuff.


It was funny after a full day workshop on the exciting and fast-moving word of chrome dev tools to hear one of the speakers bemoaning the state of in browser inspectors, though superficially they’ve not really changed since IE dev tools back in the mid 2000s. Still, even relying on Chrome can leave you in hot water when it comes to developing for other browsers and asking devs to actually develop in chrome can be controversial, a dev’s choice of IDE a uses is super partisan, though Sublime Text seems to have been in the ascendant for a while, many will never leave Eclipse, Visual Studio, *cough* Dreamweaver. Anyways his idea was simple, and the solution he’s working on is it connect up Chrome’s api to the IDE and various browsers. He live demoed manipulating a FireFox DOM from chrome devtools to many gasps of delight from the audience, followed but a too-quick plug of his project RemoteDebug.


As a CSS nerd I’ve marvelled at loads of these migical JS-free animations on Codepen and such places, but any time I looked under the hood I ran a mile; waffly code and endless lines of keyframes scares me. It seems I need to take another look though because the special sauce here (apart from some basic trig) is getting around that scary verbosity with SASS, and almost programmatic shortcutting. Definitely something I should tinker with.


Angelina Fabbro
Some mozillians making a framework of mobile app widgets called Brick (not X-Tags). I think I was suffering from a combination of some cake-pop related sugar coma (I might have had 3) and CSS animation mind-splosion because most of this drifted over the top of my head.

TIME by Jeremy Keith

Jeremy Keith’s been one of my favourite speakers as long as I’ve been going to these, and it was a real treat to see him give a lovely talk about the relative ephemeral nature of web work, cut with an old video about powers of ten. The main points were that the old adage about stuff on the web being there forever is just patently false, that we should be thinking about the longevity of our work, and at the moment the closest thing we have is html, mostly because it gracious error handling and human readability. which makes me worry a little about my blog: it totally falls apart if the back end fails. Maybe I should be looking into transferring to Jekyll?

So yeah! One of my favourites this year (the others being Ampersand and Responsive Day Out). Recommended.

“Debug & know thy tools”


As a long time Front End Dev I’ve used Javascript a lot but my focus has always been CSS and I often find debugging JS awkward. So naturally I jumped at the chance to go to a workshop on debugging by Remy Sharp, one of those names that keeps popping up in the field of JS. (You wouldn’t think it to look at him but he’s been a developer since the 90s.)

Though tools like Weinre or Dynatrace were mentioned, “Thy tools” turned out to be mostly Chrome Devtools, but even with that narrower focus there was loads to get through. Remy rattled through dozens of these tips and tricks almost faster than I could follow!

Debugging tips

  • Disable plugins and cache, use incognito mode and create a test user for a clean bug-free slate.
  • Develop right in devtools! Right-click > “map to file system resource” to be able to save to disk!
  • Shortcuts: ctrl-o: open a file, ctrlshift-o: open a function from that file and in DOM inspector press H to hide a given node, clicking “{}” prettifies the JS.
  • For element-specific breaking, right-click simply write click on it and select “break on”. You can also set event listener breakpoints in the right hand nav of the Sources tab.
  • Hover over functions, variables etc to get more info on them.
  • Chrome inspection works both ways: type simply type “inspect this” in the console to jump it’s location in the dom. Genius when you’ve got scope issues.

Bleeding-edge Canary devtools

Once you get past the obvious much can depend on what devtools you’re actually looking at; Chrome autoupdates regularly but the latest beta features are only found in Canary, and even then there’s additional experimental features that are switched off by default. For example to make devtools ignore certain files, you first need to navigate here in Chrome Canary and enable developer tools experiments, then go to Settings (the cog icon) > Experiments (left hand nav) and click “Enable frameworks debugging support”. Close and open the inspecter, head to settings again and under “General” you should now have an option called “Skip stepping through sources with particular names”, where you can type “jquery” or whatever other plugins you’re not interested in debugging. Simples!

Bottlenecks to watch out for.

Connection speed
The network tab is a pretty easy one to parse, looking for too many simultaneous http requests, too many in general, too much content in terms of file size (normally images) any 404s or text files not not g-zipped. Basically you want to streamline everything so you get a quick “Time to glass” because if nothing shows people lose interest fast.

Local memory
When the browser gets slower over time it’s probably down to memory leaks, caused by programmatically setting more and more event listeners and variables. In devtools go to Timeline > Record > Memory, play with the page and watch memory go up. When you force garbage collection it should go back to where it was. For a more detailed view go to Profile > Heap snapshots > Compare snapshots and click around looking for red highlights which mean something garbage collection couldn’t get rid of because they’re still being referenced in the js.

“Jank” is something close to my heart: Google’s name for visual distortions caused by animations not keeping up with the 60fps refresh rate, particularly on scroll. It’s recommended to use setTimeout 16ms, or better yet requestAnimationFrame whenever doing something onScroll. You can check it with Timeline > Frame to see if things are cool. Another problem can occur when small animations cause the entire page to redraw, which can be checked with settings > general > rendering > click things to show what’s redrawn. The only solutions to this that I saw were hacks though, namely selecting an appropriate frame and giving it the style transform: translate3d(0,0,0);

Commands cheatsheet

  • $$ grabs all elements that match
  • $0 grabs the element you just clicked
  • v$_ gives whatever was last returned by the console
  • $_ + copy copys…
  • console.trace() find what calls a thing
  • console.clear() clears, obvs

At the end of the workshop we split into groups and used what we’d learned to check a site randomly allocated from a collection we had submitted; it was surreal to have people baffling about a site some colleagues at VML had built, but then it was also bizzarre how totally thrown a couple of people were by responsive sites. “Whoa! It looks different!”? I’m disappointed if a site doesn’t! I know that it’s a web devs’s conceit to jiggle the portal size on any unfamiliar site but I didn’t think it was purely a front end thing…

Anyways, lots of tips, lots of take homes. It only went too quickly, I honestly could have done with double the time!

Social photo metrics


I’m totally sold on new Flickr but the still-old “stats” functionality is super weak. I assume they’re phasing it out as it’s only available if you have the discontinued “Pro” membership, which is sad as it’s rewarding to see the background noise of hits, especially when people stumble across the account and spend a while clicking through dozens of photos. Still, the fact that it effectively splits referrers into “Flickr” and “Dunno” is laughable. I’d love to know where else my pics are linked from!


That said, although Facebook’s “Insights” system looks far more sophisticated (and heavily monetised) there’s no data about where hits come from either, because everything comes from Facebook. I’ve checked out 500px but it seems to be built around a meritocracy algorithm that while great on paper ends up being gamed. As a USP it’s flawed.

For now I guess I’ll stick with Flickr, build my portfolio there, maybe put some pics in a few groups and see what happens. I just wish they did some fun stuff with data, they’ve got enough of it!

“Make it like the PSD”


It’s disheartening to hear this after weeks spent strong-arming HTML/CSS into near-perfect shape; “It looks different” they say. I want to reply “Of course it does, this is a website! That’s a picture of one!”

While the early design process likely involved wireframes, sitemaps, user stories and days of client meetings the thing that’s given the ostensibly binding “sign-off” is always the Photoshop documents; essentially “artists impressions” of what the site should look like. That they’re used to sell to the client is unsurprising, what’s baffling is that these idealised screenshots are still given to developers, project managers and QA as instructions of intent. As web technologies progress the adamant expectation that the site match the PSDs to the pixel is only getting more absurd.

Where PSDs fall short

  • Portal sizes A PSD has a single set of dimensions while the size of the browser varies greatly. It’s more prevalent now with mobile devices and RWD but it’s far from a new problem.
  • Things move In our blindingly obvious category: interactions, hover/focus states, animations, “content choreography” transitions between different portal sizes; none of it is even guessed at in a PSD.
  • Fonts oh god fonts They render slightly differently in OS and Windows, in Firefox and Chrome, and certainly in a browser and Photoshop. This is maybe the most overlooked issue; the number of times I’ve had a bug raised to the tune of “the font is wrong” when it’s exactly as specified in the PSD is just crazy. Webfonts need to be tested and tweaked in situ.
  • Copy it saddens me that designers have to spend time awkwardly putting real copy into dozens of PSDs but the fact is it’s very likely to change before going live. Designing around the size and shape of copy is doomed, especially if the site is to be translated, CMSed or responsive.
  • Un-transferable rigour When you see a page with 5 font weights and 11 font sizes with fractions of pixels, you know something’s going wrong.
  • Mistakenly inferred rigour small variations in margin, colour or font size that turn out to be oversights or mistakes rather than part of the design intent. Not worth spending hours hunting down but they’ll still be raised in QA because it’s not like the PSD.
  • Impossible stuff. Design problems that only come to light at the build stage. It happens.

You say “pixel-perfect”, I hear “brittle”.

So what’s to be done?

  • Designers should learn to code is often an end point of these discussions, but it runs against agency siloing, ignores how much time they spend selling, and to be honest I’ve had as many problems with designers knowing a bit of markup (e.g. being stuck in the tables mindset) than designers knowing none.
  • Get rid of Photoshop! It would be lovely if it was used more appropriately, like for sketching or jpeg imagery, wouldn’t it? But this would be a drastic overhaul and clients want to see shiny things, so if you need to use it, this guide on Photoshop etiquette can help.
  • Style guides In dozens of PSDs and hundreds of amends it’s easy for inaccuracies to creep in, so having a concrete reference point for developers (and new designers) can be invaluable: defining the palette of colours, font sizes, buttons and link behaviours.
  • Design QA This is certainly the most easily implemented; make the original designer part of the QA team and making sure things are in line with the vision, assuming they have availability and the project budgets for it. But that’s just the first step to…
  • Collaboration Front end developers and designers need to work together much more in a tight prototype-and-tweak cycle, moving closer to the browser and de-emphasising the PSD from the position of ill-suited absolute authority it currently holds.

I’d go as far as to say FEDs should be considered a downstream designer under the direction of the Creative Lead. We might as well admit it as it’s already happening, and though FEDs are awesome at filling in the gaps left by the Photoshop-driven design process, we’re still wasting a lot of time chasing down unimportant details and guessing at things that might be critical: We can’t replicate the vision or the understanding of the designer that’s been in back and forths with the client for weeks or months.

Quit treating PSDs as sacrosanct canon. Work closer and prototype together. The creative process continues into the browser whether you like it or not.

Agency Siloing


I wrote recently about my love of conferences; among other things they’re a great for getting a sense of what’s going on outside the office in the wider world of webmongering. One thing that always surprises is they’re chock full of designer-developers, exotic cross-disciplinary creatures that represent a distant ideal to me. Where do they all come from?

I’ve worked in agencies for nearly 8 years now, home of the design-development divide where capital-C Creatives never leave Photoshop, back-end developers have little interest in look and feel, and those with wider skill-sets are outright banned from practicing. It can get pretty bizarre: I’ve seen inexperienced freelance designers brought in at great expense because the guy who’d been designing and developing HTML emails for 5+ years had the wrong job title (in the end he had to redesign them anyway) and similarly tech PM’s forced to negotiate for a dev’s time to change some copy in a line of HTML.

In the ill-defined no-mans land between the design and development lies a handful of Front End Developers, the awkward go-between bridging the gap between form and function, coaxing Creatives into giving a crap what happens post sign-off, explaining to project managers that the PSD can’t be the canon representation of the site, and fixing the layout after the templates go through the CMS meat grinder.

Don’t get me wrong, I love the job and most of those problems can be fixed with a quick chat (few more blog posts in the works about that), the odd part is the identity crisis. Ostensibly as a senior dev I’m supposed to be an aspiring software developer, but I feel no affinity there. Being succinct and inventive with js/jQuery is sweet, but so is being terse with SASS, clean with markup, word-crafting a tweet or curating photos. I work daily with code, but that doesn’t make me a programmer any more than the fact I designed this site makes me a designer.

So I guess that’s part of the reason I take comfort in the designer developer conferences, and reassure myself that my position isn’t so strange. I’ve even reworked my CV to be more unapologetically unsiloed, because like fellow FED Brad Frost says, maybe there’s no des/dev divide at all.

Europa Universalis IV


I was one of those kids who thumbed through medieval history books and coloured in atlases with empires imaginary or real. Carolingian vs Byzantine was always a good one. A couple of decades later I’m more a gamer than a colourer-in so I’m quite excited about both of the grand historical strategy game sequels out recently: Europa Universalis IV and Total war: Rome 2.

They’re both set in Europe and revolve around building up provinces, infrastructure and technologies, forming alliances and armies and carving out an Empire. Rome is the flashy graphical one, set in europe from the Punic War, centered around epic battles and not too bothered about being historically plausible, whereas battles in EU4 are almost abstractions and the focus is firmly on the strategic map, which has the wider scope of the whole world from 1444 to 1820. These are the obvious differences but what’s not so straightforward is how differently they play:

Rome is a battle RTS with a turn-based strategy game tacked on, if you’re not fighting a war you’re heading towards one. EU4 is a real-time world-wide historical sim, where there’s lots to do in peace-time and even in the quiet periods it can be fun to crank up the game-speed and watch the world evolve. The closest thing I can compare it to is The Sims in that it’s more about writing your own narrative than winning.

It’s absurdly compelling to take control of a small proto-nation and create an alternative history. In my last game I played the Baltic merchant republic The Hansa, conquered Denmark and went on to found Germany. This time around I played an alternate reality where Burgundy survived, thrived and when on to displace France as the nation of the franks.

Hopefully I’ll get time to play some more TW:R2 but first I need to form Italy three hundred years early and become Holy Roman Emperor.

Introverted conference junkie


I’ve been to a fair few conferences since my first in 2007; this year alone I’ll have attended six and I had to stop myself from adding a couple more. I love them for the ideas and the inspiration, every one is like a uni lecture on your favourite subject by your favourite lecturer. It’s all pretty awesome but with going to so many lately I’ve had the increasing suspicion that I’m missing the point.

I love conferences

The talks at dConstruct were great:

  • I missed Amber Case due to a train-based screw up which was a great pity.
  • Luke Wroblewski gave a lecture on how we’ve gone from mouse and keyboard to about two dozen potential inputs forms in a scarily short time.
  • Nicole Sullivan gave advice and reflection on the phenomenon of trolls and trolling. Particularly interesting was how to manage your inner troll, the hangups you wish you didn’t have.
  • Simone Rebaudengo asked what might happen if products (toasters, specifically) started communicating with each other instead of exclusively with humans.
  • Sarah Angliss gave a fascinating and left field talk about unnerving sounds and awe, followed by a sweet performance.
  • Keren Elazari’s talk was basically a loveletter to hackers. Quite uplifting if oddly idealistic.
  • Maciej Ceglowski was maybe my pick of the day, on slash fiction writers and managing communities on the web.
  • Dan Williams spoke on some unexpected results of the technology that surrounds us, from unlikely gadgets to stalker bins. It’s the future. TAKE IT.
  • Adam Buxton was hilarious as ever but phoned it in a little tbh.

My favourite conferences this year are still easily Ampersand and Responsive Day Out – both at the same venue by the same people – but all the talks were nerdy and thought-provoking and a joy; that’s not my problem.

But I can’t networking for my life.

I’ve often heard it said that the main point of conferences are the networking opportunities, but as an introvert going alone I’m unlikely going to impose myself onto other people’s afternoons. My reluctance was nicely highlighted when web-luminaries Laura and Aral randomly sat next to me, both of whom I’ve followed online for a year or two. I was amused and delighted to discover the weekend before that Laura used to live with a friend of mine, but when she began an exchange with me about whether the seats were taken I kept it as short as possible. Bizarre. I guess I’m too used to London where people are white noise and conversations are a terrible faux pas? It’s probably not that strange, but it does make me feel I might be missing the point.

Normally I spend the elongated lunchbreaks wandering around Brighton with a camera which is lovely, but this year I’ve been unlucky with the weather, and an hour+ sitting alone in the café is pretty sobering. It’s something I need to solve if I want to keep getting my fix of web nerdery:

  • Going with friends would be amazing, but I’m a bit mental taking time off work and paying a couple of hundred quid to hear some people hold forth on these subjects, so it’s a difficult sell to friends and colleagues. It would be awesome if I worked somewhere that had a culture of attending these things.
  • Get a laptop. A bit sad but I could trawl twitter or stay productive and not worry about not having anything to do? I can’t really afford a laptop at the moment though, what with all the conferences…
  • Stop going to so many. Yeah, probably not likely. The next one I’m going to is Full Frontal which I’m really looking forward to, and GeekyConf after that which I even managed to rope my boyfriend into. It’s gonna be good.

I think I have a problem.

Parsing for fun


Creatively I’m an editor, a curator. I take things, move them around, tweak, reorder, refine, tell a story. In terms of photography I mostly document what’s there. I think I’m getting better at setting up the shot and playing with light, but as for telling people how to arrange themselves, forget it. Clueless! The real work begins once I get the pics back to my computer, but I’m starting to think I’m overdoing it slightly.

For every photo I have to work out if it’s any good:

  • Is the focus sharp?
  • Is there a clear object?
  • Could it be improved with a better crop, light balance or levels?
  • Is it shiny?

Then there’s the part it plays in the narrative:

  • Does it look busy?
  • Does it tell the story of the night?
  • Are there more smiling faces than glum faces?
  • Is there too much derpface?
  • Will it show a good mix of people?
  • Would it make a decent GIF?

Given I have four to seven hundred photos in a set, it’s a lot of questions, and it takes hours to whittle that down to the few dozen that I put up on Facebook and the handful of favourites I put on my Flickr.

Click to see GIFs and read the rest of the post

Fools rush in


As a front end web dev it’s standard for a producer to give me a site to build, telling me to start with a given page. I often have no knowledge of the project but like a good little worker I open up the right PhotoShop comp and get straight to work styling it up. Of course this is a terrible idea.

I’ve been building sites for ten years now, large-scale international sites for seven: I know very well that they’re not built page by page. That didn’t stop me diving straight in and starting out my current project with the homepage header. Seems legit, right? But a cursory glance over the designs won’t uncover the layers to the onion or the pitfalls scattered around dozens of PSDs. Just that lowly header has 3 configurations for desktop, tablet, and mobile view, 3 more for the home-page-specific version, 2 more for scrolled-down view in tablet and desktop and the main nav has 2 compacted toggleable states: hover for scrolled down and on-click for mobile. Responsive gone awry.

TLDR: the header has about ten views.

For non markup monkeys, imagine a transformer with 10 forms. Each one may be simple but all coexisting in the same space it’s a nightmare. If I’d taken a step back at the beginning of the project to grok the site I’d have been able to minimise the impact; rationalise the design, keep things modular in places and siloed in others, and gotten through it in half the time. Instead in my haste I’ve fallen foul of every pitfall, rewritten everything three times, gained a few white hairs and frustrated poor PMs wondering what’s taking so long. DERP.

Note to self: next time go all Ackbar on their asses.