The junk yard
projects & experiments in various states of completion
Sketches & concepts, exploration & variations, rough & fine. The birthing of success & failure.
Ever pre-book merchandise at a festival only to forget where the pickup table is? Me neither! But it was a good excuse to get my hands dirty with AR!
I guess my maths teacher was right all this time ... 😳
A high fidelity PoC / spike to determine feasibility of re-using UI / UX effort from an iPhone app to an inlined web widget / component.
In a celebration of simulacra, the widget successfully mimicked the blurred glass background & docked avatars of the individual stylist cards while maintaining fast loading speeds and snappy scrolling. All this and a TINY footprint over the wire!
It was a joy to use but was quietly shelved, I think the reasons will be obvious.
Yeah, we never did nail an elevator pitch...
Featuring a tremendously tiny & snappy UI, the real secret sauce was a request processing system nicknamed 'Project Magic'.
Magic automatically managed supplier discovery, enrollment and fulfillment based on often vague user requests, an achievement in 2014 / 15.
Sadly, this project was seminal, for all the wrong reasons...
But at the very least there's a decent book or documentary to be made about this one. You know, for when we have the time... 🧐🤔
Because there was just too many loading spinners out there in 2007!
MkiScript (typically pronounced Mark One Script in more formal circles) was conceived on a global project hosted on lethargic, bandwidth starved servers.
JS libraries of the time were too large and the creation of something smaller became a necessity to reduce bounce rates.
So we stuffed AJAX, animation, CSS selector based element gathering, a plugin system and a whole lot more into a tiny 3k package.
Time & browser advances have removed the need for this little fella, but it sure gave us some big advantages at the time!
A bespoke hypermedia API stress testing tool. Using an APIs own logs, this tool would assemble a series of EC2 instances and beat the living hell out of the target API with realistic requests.
Child instances (originally called 'the horde') maintained a real-time connection with the instigating instance, which in turn fed real time results to the user.
A lightweight blood glucose diary for my T1D buddies!
It was mobile first and featured a text-to-speech interface to improve ease / likelihood of use. In addition to recording just numbers, users were encouraged to share consumed food / drink and exercise between each reading.
Trends were presented on a line-chart built purely from DOM elements, which I assure you was pretty cool!
This project was my first shipped project on a MongoDB backend, but it introduced me to the potential of NLP which I apply to this day.
Ever get asked to build something you know is just a bad idea? Not just an idea that's hasn't been fully thought through, but one that's fundamentally flawed.
The problem: "try on" clothes over the 'net.
The prescribed solution: the '80s Barbie comic book style 'paper dresses' try on game, reimagined...
Minds were already made. Confidence in the solution unwavering. No amount of discussion would allow a pivot towards a better direction.
So, disagree but commit! The resulting demo was met with a deafening silence. It was never spoke of again.
Isaacson's Steve Jobs biography has cast a long dark shadow...
Even moar webXR!
A high fidelity PoC to explore the concept of a scriptable, quest based, augmented reality learning tool for students with special needs.
Building on the Ar merch finder code base, this PoC matured quickly. While proving the project and approach very feasible, and genuinely fun, it raised 'video gamification' concerns, and was subsequently sadly shelved.
I've been very tempted to dust this off and create a multi-player 'treasure hunt' style AR game 🤔
Wordpress before Wordpress! With a name that aged like milk! (who remembers Microsoft Front Page??)
BackpageCMS is still under active development, but the vaunted version 3.0 never saw the light of day. Its goal was to be a clean sheet project where A11y & UX were prioritised, building on top of a validating WYSIWYG xHTML CMS. (additional acronyms available on request)
BP3 was a bastion of technical purity. Compliant with all the fads du jour, structured around an extremely versatile xHTML engine. It was also very slow, awkward for template designers to use and brittle as f***.
Suffice to saw I learned my lesson and priorities have become much more user-centric since.