All posts made in Aug. 2017:

Gratitude and a possibly inappropriate technological intervention

I was reading Melanie Klein's Envy and Gratitude and Other Works (which I still haven't finished) and there's something about Kleinian gratitude which is crucial in developing the primal relationship between mother (the good object) and child. It is also the basis for the child perceiving goodness in others and herself.

Conscious gratitude seems to be more focused on the other, rather than a self-centred idea of being the cause of goodness or its reverse. Developing gratitude might allow for greater capacity for appreciation, acceptance, and the sharing of love.

Gratitude is inherently outwards looking. And surprisingly hard! It touches all kinds of other feelings like deservedness, and is easily corrupted with responses like entitlement.

So I was thinking: a habit of gratitude would be an interesting thing to foster. Gratitude being a component of prayer, I know, but I don't pray. So. I need to get it somewhere else.


We can fix this with technology. I know, I know. Forgive me.

What I do is I have a folder in Ulysses, which is a writing app I have on my iPhone (and I use for everything). The folder is called: What I Am Grateful For.

Please also forgive the ugly dangling preposition. It upsets me too.

In that folder are tons of notes. Each note has a date, and a line of text: the thing I am grateful for that day. Sometimes big, mostly small. Sometimes easy to observe, sometimes really, really difficult. Always interesting to note when I’m going through a phase in which gratitude is a challenge to attain, and with what that correlates.

Back to the tech.

Once a day, at midday, I get a notification which says "What are you grateful for today?" I tap the notification, and a text box opens up on my phone. I type into the text box and it gets saved into the folder.

Here's how that bit of automation works:

  • I use an app called Workflow which links together different apps and lets you program them in a flowchart sort of way
  • I've written a particular workflow called "Grateful Daily" that does all the work of opening the text input box and saving it to Ulysses. You can get the workflow here. If you copy the workflow, you'll have to update the special Ulysses code bit to make sure it saves to the right folder
  • Another app called Launch Center Pro is able to trigger workflows on a timer. I have it set to run Grateful Daily at midday

Cross-app automation is a nascent but interesting area. I'm finding myself able to do pretty complex workflows from my phone now (I also have a process to edit and deploy code, using multiple different apps). It's got a way to go as a pattern of user behaviour, but I'd like to see iOS or Android take automation more seriously. To see where it could go. It has a different nature to automation on PCs, and I think there's the opportunity for these automation scripts to unbind from the smartphone and move into the cloud (somehow). Maybe use a bit more intelligence too. Centaur automation.

Yeah but so: gratitude.

To receive - and to be open to receiving! - something which is good, and to take in that goodness and to internalise it, but to also appreciate the goodness itself, and its source and the source’s reasons. A tricky business.

I don't even pretend to have even half a handhold on Klein, or Kleinian gratitude, or hell even gratitude, but her words opened something in me. (Thanks!)

A few technical words about Upsideclown, and some thoughts about audiences and the web

I also write stories over at Upsideclown, which is both a website and a small writing group. Well I should be careful about the present tense: I wrote there between 2000 and 2003, and I shortly will again. We recently resumed after a 14 year hiatus.

First time round we published twice a week. There are a ton of stories! We had a party for our readers! We self-published a book! Which back then was hard. I had to do layout in Quark and install a special printer driver that would send pages to a printing firm in Tennessee which could do short-run binding. Taking payments online to sell books was, well, I think we ended up mostly doing cash.

This time round it's lower-key: we're publishing once every two weeks. It's the same group of seven, so each of us comes round every three months and some. I'm enjoying seeing how our writing has changed and also how it hasn't.

Upsideclown is old-fashioned I suppose. I started it as an excuse to keep in touch with friends from uni. It’s me, plus Dan, George, Jamie, James, Neil, and Vic. And it's mostly about that plus of sprinkling of some useful pressure to keep my hand in writing fiction. It's not the central purpose but it's always nice to have readers so I do a little light promotion too. (Subscribe to the newsletter!)

It feels transgressive to have a website in 2017. Something about having a domain name and about coding HTML which is against the grain now. It's something big companies do, not small groups. We're supposed to put our content on Facebook or Medium, or keep our publishing to an email newsletter. But a website?

Technical details

Resuming 17 years after starting gives you a glimpse of the long now of the web.

There are pages I made - the first stories - that I haven't touched in almost two decades and they still work. Web-world that's pretty rare. They’ve outlasted most places that encourage you to host their content with them, and even the popularity curve of many programming languages and web frameworks. Database technologies have come and gone.

And then there's my own attention and ability... I no longer program for a living as I did when Upsideclown started. I keep in touch and still make the odd thing, but I've forgotten a ton. So I have a philosophy around choosing what tech to use when I’m building this stuff: will I be able to fix it, half drunk, ten years after I've lost all the tooling.

On the design side I’m pleased that when the design changed, I made sure the old stories kept the old design when the newer ones picked up the new one. Design changes meaning. A story would mean something else if I retroactively put it in a classy frame, or a punk frame, or added highlights.

So it's all about longevity and data. I rewrote everything for the reboot and here’s how it works now:

New Upsideclown stories are stored in Markdown because it's a simple format and, in another decade, when I’ve forgotten everything I know, I’ll be able to tell what I meant just by looking at it. Here's Neil's recent story as published. Here it is in the Markdown "source" format. No databases, it's all files.

Rendering is in PHP. Old school I know, but it's a language which has stood the test of time, and I can go from a standing start (seriously I hadn't coded with it for ten years) to a functional and decent looking site in a leisurely weekend. So if I need to rewrite I know I can. I'm not planning to: unless some burning desire strikes, not until 2034 at the earliest. It’s served with Apache, no app engine. I figure web servers will be around for the duration.

(I have the same approach with my blog: posts are text files and go back to February 2000. The publishing engine I've rewritten a few times.)

The reason the stories aren't kept in HTML is so I can publish out to other formats: there's an RSS version that is picked up by Mailchimp so readers can subscribe to the newsletter. The archive is dynamically generated.

I’m looking at whether to also generate Facebook Instant Articles and Google AMP, perhaps plug into Apple News too.

People don’t visit or link to websites like they used to.

You gotta fish where the fish are.

But you also want to be able to look back on what you’ve created once you’ve retired. So it’s a balance.

How web publishing has changed

I knew that publishing to the web was out of fashion. What I hadn’t realised was how much the tools had eroded. Or rather, two things: if you want to know how many people have read your stories, and where they came from, then (a) the analytics tools haven’t kept up with how and where people read; and, (b) the analytics tools are made for big companies optimising the flow of audiences down funnels to achieve particular goals. Not for small, independent publishers.

Here’s an example. There’s no simple online tool that lets me add up how many people have read a particular story on Upsideclown via the website, the RSS feed, and the email newsletter. Why not? If I add syndication to Facebook, Google, and Apple, I’m even more at sea.

This isn’t because I want to optimise an audience; this isn’t because I want to sell ads. This is because it’s nice to know that 17 people read the website and 21 people opened the newsletter, and 36 people read the same story on Facebook, and 6 in an RSS reader -- and gosh that’s like the whole top level of a double decker bus, all those people read my story! When companies deal with millions and billions, I think perhaps they forget how the intimate feels. How sometimes it’s not about a thousand retweets but instead about an audience of readers who come back. With whom you have a relationship. Who appreciate you, and you appreciate them. Yes it’s a pleasure to write, and yes I will do it without needing to get 1,000 likes on each and every story, but also let’s not forget that it’s more pleasant with company.

These likes, faves, and claps are cold. No wonder we need so many of them to feel sated. Instead I’d like to look at the depth and duration of meaningful relationships, and to share that with my fellow authors. I know analytics feels like a dirty word in this context, tarred as it is with A/B testing and e-commerce flows, but there’s a joy to be had in being on stage and seeing the faces of your audience — rapt. The erosion of tools for modern online publishing has, bizarrely, made the intimate audience invisible. What can we count so that we’re over the moon when there’s twenty of it? And simultaneously I’d like to make it easy for readers to read wherever they are, whether that’s the web, Facebook, email, or whatever. I can handle that last bit but Google Analytics doesn’t help me with the former. Nor does Medium.

Not without budging on my desire to make pages which I can still read in 2034, anyway. It seems to me that, sometime in the last 17 years, the web forgot the simple pleasure of making, and appreciating what’s made, together.

Two positive signals for the Smart Home

I'm bullish on the Smart Home, and as someone with a professional interest in the Internet of Things who was consumer-IoT-shy this time last year, I’ve been thinking about what changed my mind.

This isn’t a whitepaper, or a even a properly considered analysis: just some notes about where my head’s at and what I’m looking at. I’d appreciate feedback — both supporting points (especially pointers to UK startups who are taking advantage of these trends) and counter-examples. If I’m off-base I’d like to know!

Over the last couple years, the Smart Home has been getting a bad rap. Connected consumer products suffer for a bunch of reasons, including but not inclusively:

  • the sane products aren’t worth the money. Features for this generation of products are generally iterative not transformative: is it useful to remote-check your slower cooker. Sure. But at 5x the price, with the cognitive overhead of apps for everyone in the family, etc? Not so much. Many of the benefits become clear when there's a critical mass of products that can orchestrate and learn from one another
  • the category-busting products aren't acceptable in the market. Consumer behaviour is hard to shift, both for features and also for the business models required for connected hardware. Look at the outrage when a software application goes subscription-only; now think of how a subscription washing machine would go down
  • the barriers to entry are too high when you combine hardware and software. Shifting consumer behaviour is possible... but it takes a lot of experimentation on both ends. Hard to do when hardware is so expensive in small batch production, especially for a startup

All of that said, this year I'm getting excited about consumer Internet of Things again. There are a few trends that make it easier for the Smart Home to get out its slump, such as the ever-increasing acceptability of e-commerce and direct sales, which reclaims the retailer margin.

But two signals in particular.

The smart home platforms have finally given up their fight to own it all

The first signal is smart lighting from Ikea which both fulfils the promise of a low-cost modular system, and also has sane interaction design (that is: it includes physical controls and works when the internet is absent).

More importantly it works with a gamut of Smart Home controls: Apple HomeKit, Google Assistant, and Amazon Alexa. This tells me that the GAFA stacks (aside: where is Facebook in the Smart Home?) have given up on their unrealistic desire to treat the home as a monolithic own-able platform. The layers are emerging: it will soon be possible for a startup to innovate on a new type of bulb without having to also break into the service layer (and yes, I've met companies with internet-connected bulbs showing a 10x life at comparable cost. Being able to plug-and-play HomeKit, Assistant, and Alexa would be a godsend for them).

At the service layer, it should also become possible to innovate on software and orchestration between devices: I look forward to services that are able to plug in to a smart home from a mix of manufacturers, providing highly specific and differentiated functionality. I think voice has a part to play here: it’s the excuse we’ve been looking for to put our phones down at home.

Interop is good news: more ways in for startups, more places to innovate, and better value for consumers.

Reference designs lower the barrier to innovating on features

The second signal is also a healthy emerging fracture point in the connected hardware stack: Qualcomm's reference designs, including this speaker platform. Reference designs allow for some interesting manufacturer efficiencies. A small company can go and ask for a customised version of this product, benefiting from a supply chain shared with other small companies, and with engineering costs amortised across the same.

The reference design linked above is for a smart speaker. A speaker is no longer just a speaker: it's a speaker with directional microphones, a wi-fi connection to a cloud somewhere, and enough on-board GPU to run a voice assistant, whether that's from Apple or Google. I'm interested to see what the equivalent reference designs are for a smart screen, smart doorlock, smart book, and so on.

If these appear, it will show that the consumer categories for smart products are stabilising. Categories are useful because they allow the rest of the industry to align: retail buyers can set up aisles; marketing educates the consumer; it becomes worthwhile for distributors to do their thing. With a baseline of many products in the same category, it becomes possible to experiment.

Critically reference designs provide an entry point to startups that lets them mimic Apple's business model: hardware differentiated by software. To date this has been inaccessible to startups because hardware development is a huge barrier to overcome before service innovation can begin (not to mention the challenge of distribution). The table stakes are, happily, coming down.

Overall the Internet of Things is going to see an interesting few years. The digital world has seen rapid change: Blockchain, A.I. (in a thousand forms), and next generation interfaces too: voice and augmented reality. The recently stabilised IoT tech stack is pretty solid: digital dividends should be coming into the real world much faster than before.

And then, maybe, finally, we’ll start seeing those category-busting transformative products that we’ve been waiting for.