Is this your first MonkiGras?
(Everyone I spoke to)
Yes - it was. And it was a real eye-opening, thought-provoking and exhausting couple of days.
Every. Single. Speaker was excellent, and what’s more the connections and synergies between their diverse topics and backgrounds was fascinating. Context matters. Trust matters - how can we build it? Skill and practise matters.
The hallway track was excellent - I had so many wonderful conversations. The location was too noisy. There was nowhere quiet except the cold cold outdoors.
My notes and thoughts follow. Please know that I didn’t attend every session, probably misheard or misunderstood plenty. Interesting that talks were mostly untitled.
Intro (James Govenor)
Sustaining the craft is about sustaining relationships
Aneel spoke about value-chains and supply-chains. About information-streams (ie supply-chains for information). Every layer in the chain takes a cut of: power, money, influence.
Consumers have to have trust in the supply chain because it is completely opaque, but this trust is abused (example: peanut butter which used to be only 75% peanuts). In being forced to trust the chain, consumers are disenfranchised and disempowered. Brutal phrase:
Consumers are the product, industrially farmed for their cash.
Wherefore sustainability in this? Removing layers allows increased trust - eg craft fairs, farmers markets. Open source?
For me, this set a bold tone for what is nominally a tech conference. Hardly about tech at all, but a wide-ranging view on an industry which produces.
Dormain talked about her interest in letterpress printing, paralleled with her husband’s interest in restoring vintage cars, and further paralleled with sustainability in software.
Why is this still a thing? It makes no sense. Letterpress was surpassed by lettersetting in the 1950s and by digital publishing in the 90s. But despite seeming obsolete it has not stopped. You can buy a cheaper, safer, more reliable car than a 1950s Porsche but that doesn’t stop someone having 30k followers in Insta for restoring them.
Her thesis (tongue-in-cheek “completely unscientific”) is that there is an intersection of: Community, Passion, and Commercial Value which sustains a craft.
The metaphor applied to software became: OSS, developers, and vendors.
Throughout the day I kept coming back to this, thinking about how it could apply to the hobbyist communities I have been part of. I’m delighted when I learn something new which can help me see something I know well in a new way. She did not mention that Pivotal is hiring.
Catherine talked about typography, a subject about which I know almost nothing, so it was a pleasure to learn something new. About the modern invisibility of typography, and about the importance of context. Example was the sign on the Royal Homeopathic Hospital where the word Royal has been inserted with adequate type, but without considering context and site-specificity or applying the skill which is evident in the lettering of the other words.
It’s cheaper to teach generalisation than specialism.
Indeed Catherine is a teacher so this comes with some authority as an explanation of lack of appreciation of the deeply skilled.
Caution: retro-fetishism of craft for the production of status objects. Counter-example: letterpressing (again) to produce material for political activism.
Chas talked about 100-year patterns in software. And PDF, yes that PDF. Binary data on-screen showing some internal structure of a PDF doc. He talked through the inception of PDF, how it improved upon its predecessors.
He showed part of a paper in which the designers described their intention: the fundamental way people work will change. Scope of ambition that “makes my socks go up and down”. (what is the goal of your current project?)
Releasing the PDF as a “standardized” data-format provides a future-proofing which can survive any individual vendor.
I know Chas’ work for some years in the Clojure community and wondered about the parallels there. Hickey has been very vocal about the philosophy of Clojure and how it is designed to solve enduring problems in software development.
Lars: AEM - Adobe’s CMS, which by some imaginitive application of filters can in some way be considered the 3rd most popular in the world. It is also decades-old software.
Several tips for longevity:
- Trust people with long experience
- Process and tooling matter
- Manage technical debt and feature creep carefully
- Deprecate slowly (“like any other hostage situation: no sudden moves”)
- Don’t get too attached
Tension leads to innovation, eg static/dynamic or compiled/interpreted. Even if there are no absolute answers to these questions, the conversations around them are valuable.
Architecture is destiny
Cultural aspects: Empathy (send developers to work with customers), Curiosity (risk as experiment), Transmission of culture (stories, shared goals).
Jess (I can call her Jess: we’re friends, I’d lend her money) talked about burnout and self-care. You’re not crafting anything if you’re burned out.
Metaphor for mental capacity: Working memory as a beaker, cognitive load (ie anything else at all that you’re doing) as liquid in it - the space left is your available capacity. Important: The size of the container changes! (eg tiredness)
Symptoms of burnout similar to depression - can take months to recover. Recovery: time off, do less, outsource, say no, “selective emotional investment”, ask for help, forgiveness (behold my field of hecks to give is barren). Also: don’t be a jerk.
I guess it takes time to get used to managing this. And it can be hard to judge your workload when the nature of your work changes. I feel at the moment that I am likely overstretching myself, so this was a wakeup call to be proactive about it. Thanks Jess.
Charity knows how to make being on-call awesome. Ops has a history of masochism. But managed right on-call rotation is something people can love doing!
Lots of war-stories and practical advice - can’t wait for this one to go online for a rewatch.
Load-balance and distribute knowledge and experience. Ask people’s significant other what on-call is like (“she flushed my pager”, “he should stop writing shit software” - the audience).
If you haven’t used your SLA-allowed downtime, you’re not taking enough risks
Care more about less.
TIL O’Reilly don’t allow mythical creatures on their book covers, but Charity had stickers to turn the horse into a unicorn.
Luis on sustainability and open-source. Using software used to have a lot of friction, ie negotiating enterprise licenses from a vendor. Now we are used to using OSS, is there no friction any more? (spoiler: there is)
Friction in the form of OSS licensing. A tool exists for one-click raising an issue on someone’s github to request that the software be licensed correctly, but this just makes it easier to piss someone off.
Complexity of licenses - so many thousands of dependencies with different licenses that how the hell can you even?
Frictional people. Saboteurs and luddites.
Theo Ethics in software. Piss and vinegar. “Ethical debt”. Software is a young profession. Ethical standards change over time, which is good. Ethical professionals (ethics is what defines a professional) can be trusted. Links back to Aneel’s theme of trusting a supply-chain.
History of professional ethics: Clergy, medics, business, professionals …. wherefore software?
VW emissions - there is 1 person in jail but
it takes a community to fuck up this bad
Uber. Uber again. Many people had to have been involved in unlicensed autonomous cars running red lights in SF. Non-whistleblowers should be excommunicated from the profession.
Hand-soap dispenser which didn’t recognize black people’s hands. This was unintentional but the onus is not to ask only “who will this help?”, but also “who could this hurt?” or “who could this disenfranchise?”. Isn’t it negligence?
Strava global heat-map. “army bases can look after themselves” but this data also shows home locations of people with expensive bikes. Who could this hurt?
Ref. ACM’s ethical policy (1992, due an update in 2018)
Pia. Runing for political office in Argentina but found they had to obey the rules of the status quo in order to effectively challenge the status quo. Ugh.
Legal frameworks for finance, politics, contracts expect heirarchical organizations which is super out of sync with network-like communities.
With many small collectives there is duplication of effort (eg in accounting, legal) & no shared benefits. Need a new transparent, trusted, distributed way of operation. Fascinating observation: transparency doesn’t foster trust but makes it easier to operate in situations where trust is difficult. 3rd speaker to talk about trust in distributed systems of people.
Enter: Open Collective. Starting a collective should be as easy as creating a github repo.
End of day 1
I was totally exhausted by the end of these talks. Just enough time to visit the cheese mountain, meet Pid, and then go home to crash out.
Mazz. Fashion industry slow-moving production model is broken. 10% of new clothes go straight to landfill as long lead-time means much waste. Zara has a 2-week lead time. Customized trainers are 25% of Nike’s online sales.
Mazz works for unmade which, if I understood correctly, is knitted-products on demand, with excellent web tools for people to design their own unique clothes.
You think software is hard you should spend a day in a factory
Interesting constraints when your code has to interact with unreliable systems (DHL) and pysical systems (knitting machine floors don’t have computers. Take time to load raw materials. Can be expensive if a software bug shears off all the needles). Make your code easy to delete :) Most unsustainable tech practise: hiring only senior devs. Optimize for learning.
Hiring bias: Men recruited on potential, women on past results.
Amanda gave a really interesting talk on Marie Curie. Parallels between Curie and OSS. Curie developed a method to refine (tiny amounts of) radium from (huge amounts of) pitchblende. But women in 1900s France were not allowed to own property (I was shocked!?) including Intellectual Property (ie patents) so the refining method was published (ie open-sourced). Curie sold contracting services and tools to people. Sound familiar business model?
Radium was ludicrously expensive so Curie didn’t actually own any until a proto-kickstarter collected enough (and more) for her to own some. People donated even single dollars.
Ricardo. Craftsmanship - WTF does that even mean. Plastic food in restaurant windows in Japan looks just like the food you will be served. Someone in a no-name curry restaurant cares enough to do that - this is craftsmanship. Jiro (dreams of sushi) is less interesting because he’s an example of survivor bias. Perfection is an end-state but craftsmanship is a process.
Consistently avoiding shortcuts.
Find allies who: help, challenge, pick up for you, you pick up for them, provide diversity, share your aesthetic.
Frameworks are training wheels. Take them off when you know what you’re doing. Be sceptical of authority.
Joni works on IBM’s corporate design program. IBM employs 1600 designers. What kills or sustains culture? Culture sustains craft. People + practises + places = culture. So you can change people, practises and places in order to affect culture which leads to high quality craft.
Best practises evolve, craft is constant. A loop: Plan->reflect->do->reflect.
“Love your work”. Compare with “don’t get too attached” from Lars. I think the difference is to love doing your work, but don’t get too attached to any one specific part of what you have made.
Sustainable Fashion Panel
Well, “fashion” is overloaded with meanings so fo this panel we’ll
An ancient craft. Has been through multiple phases of craft/mass production (eg luddites, again)/outsourcing. “Fast fashion” changes weekly.
Unintended consequence: overconsumption of clothing leads to donations to charity leads to saturation of charity-shops leads to export of old clothes leads to deskilling of clothes-makers in poor countries. Cf Theo’s talk “who could this hurt?”
Fashion is emphemeral, by definition. 3 sustainable clothing manufacturers on the panel. Tech helps by allowing direct access to customers & manufacturers globalls, new production techniques, transparency within supply-chain (again!) Tech also enables massive overconsumption/production “who could this hurt”
Waste is a design flaw
Design either for long lifespan or easy deconstruction.