I had not been to YOW before myself. In my previous job I had sent my team there though; the timing works out perfectly for a business that tends to be quiet over Christmas and the new year.
I had been to TechEd before, and that was the closest benchmark I had.
YOW definitely wins on a broader scope of topics, and yet not so scattered that it was hard to find a relevant topic in each session timeslot. I have been to TechEd with timeslots reserved for walking the floor, because despite 5-6 sessions running in parallel there wasn’t one that I actually wanted to see.
TechEd wins on the catering front, hands-down. But then, in all the years I went, it never ceased to amaze me how well-oiled the Microsoft conference machine is, and the amount of effort they put into a great show.
Having said that, YOW wins hands-down on cost… it’s kinda the way they pitch their conference – focused on the technology and affordability at the expense of spectacle. I don’t think that’s a bad call, because TechEd was definitely never a trivial sell to management.
And yet… the one point I have a hard time qualifying specifically; YOW doesn’t feel as amenable to making new connections on the floor. During sessions the focus is on the speaker, and during breaks it is on getting some food and finding a place to eat it. Maybe it’s the TechEd lunch set-up around large tables that force you to sit down with strangers that helps? Maybe it’s the social evening of fun they put together (one year: go-kart-racing in the parking garage… yes… seriously) that makes friendly conversations easier to come by. I don’t have a clear answer, but YOW does feel stiffer than TechEd.
Having said that… if you’re just there to learn, the set-up is ideal.
To my shame, I barely remember the opening keynote.
I hadn’t had coffee yet.
I was trying to work out when I was going to be where.
And then I had 11 sessions and 3 more keynotes with memorable elements.
I think it just displaced (almost) all my memories.
The one thing I recall with clarity is the question how it can be that the most complex piece of technology in existence today is usable by kids as young as 2 years old in a meaningful way; the mobile phone. Something about it’s design is the pixie dust that hides the million things you don’t want to do right now.
Session 1 – Rethinking MVC with React Native & ReactiveCocoa
Ben Teese and Sam Ritchie
I was only superficially familiar with React before this talk, so it probably wasn’t aimed at me. But I still learned a lot about the way React uses it’s Virtual DOM to do delta-updates on-screen, and how that can be made to translate to native apps on mobile devices as well.
I believe Angular 2 is going down a similar path of supporting some kind of Native variant, although I may have just dreamed that. Either way; not quite relevant to me at the moment.
Great talk though, good speakers, especially the groan-worthy punny pictures.
Session 2 – 40 Agile Methods in 40 Minutes
Spoiler: it’s actually 50+ minutes; the session ran a little too slow to make good on its titular promise. Having said that, it was a very enjoyable whirlwind through a lot of Agile Development approaches (some of which didn’t look anywhere near as Agile as they purported to be).
There were a few slides that specifically piqued my interest, but the pace prevented me from taking notes, so I look forward to the slides getting published. Especially the slide with a book that Craig recommended as providing a good underpinning of why Agile works.
Great talk, and more useful than the frivolous title might lead you to believe.
Keynote 2 – Using Formal Methods to Eliminate Exploitable Bugs
It’s been only 17 years or-so since I graduated from University, and at long last the central pillar of the Computing Science program of Eindhoven University of Technology is actually becoming useful.
See ma, my degree has practical applications in the real world!
And apparently, Australia is one of the fore-runners in this field. I don’t want to say it’s because of me, but hey… I’ve been here 17 years now. Coincidence? I think not!
In all seriousness, it is great to see Formal Methods taking their rightful place as a central tool for the development of provably correct software.
Session 3 – Adaptive Security
The key tenet of this talk was: exploit your logs; exploit them for all it’s worth.
Know what your messages mean, count them, and then look for patterns. And then act on those patterns. And start simple, because the business knowledge that produces will lead to requests for more of the same, and before you know it you’ll be tracking and measuring everything.
I couldn’t think of better real-world advice.
Even beyond just security matters.
Session 4 – Production Haskell
This was by far the greatest disappointment of the conference to me. Based on the excerpt I had hoped to see some samples of practical use of Haskell in a real-world production scenario.
In the end I walked out before the session was over, because I just couldn’t muster the will to look at further tool-chain scripts to build Haskell. That was so not the point I was coming to see.
I’m sure I could figure it out myself, but I wanted to come and be sold on the idea.
Session 5 – Pragmatic Microservices: Whether, When and How to Migrate
I slumped through this; not a bad talk, but after a full week of Udi Dahan, there wasn’t really a lot more anyone could tell me about big balls of mud and how to take bites out of them.
I had hoped those nice big open questions in the title would lead to new practical insights, but I think I just kinda zoned out and let my afternoon snack digest.
Not a bad talk, just nothing much in it for me.
Session 6 – Property Based Testing: Shrinking Risk in your Code
This talk felt like a “missing link” between Formal Methods, Unit Testing and .NET Code Contracts. After listening to it all, I like the idea of higher-order tests, and I see how you could leverage them in procedural languages.
But it feels like perhaps it’d be easier to just go Functional, or use Theorem Provers instead of wasting this approach in its pure form on C#. Still, I like the ideas underpinning this way of testing, because as I’ve blogged previously, I’m very unsatisfied with the cost/benefit balance of most of the automated testing I have been exposed to in my working life.
Keynote 3 – Engineering and Exploring the Red Planet
Anita Sengupta and Kamal Oudrhiri
Anita is a great speaker. Kamal was clearly nervous.
Having said that, it’s hard to botch a topic as inherently interesting as trying to land complex and fragile technology on another planet within a very small target area. It’s hard to appreciate how awesome the stuff is that JPL and NASA do without someone explaining all the details that you’d never have thought of.
I secretly suspect there are still a lot of Waterfalls in these places to deal with the careful design required for a release that you don’t get a second chance at. Once it leaves the planet, it better work.
Also, it made me want to go and see The Martian again.
Session 7 – Building Microservice Architectures in Go
This was actually a very fun Microservices talk. Not as much hands-on Go as I might have hoped, but some very salient points were made. It hadn’t occurred to me before that the combination of a language that statically links everything together with a very lightweight container is immensely powerful for Microservices / Service Oriented Architecture.
I guess he just made a very compelling case for dotnetcore without even realising it.
Also, he was an excellent and engaging speaker. He felt by far the most polished out of the speakers I saw. Having said that,… the thongs were visually incredibly distracting.
And just as I thought thongs would be about as distracting as it could possibly get. Bare feet on stage.
After this talk I feel very self-conscious about my use of the term “Agile” as a short-hand for Agile Development Practices. A very well-put rant against the commercialisation and co-opting of common-sense to extremes where it just stops making sense altogether.
I suspect the fifth agile tenet that he can no longer remember might have been “Pizza over Steak”; that sounds like something a programmer would declare.
I guess the biggest lesson from this session is a cautionary tale; to keep an eye on the practices you follow and to make sure you don’t fall in the trap of trying to buy Silver Bullets. We all love them so much, and we know they don’t exist, but we still buy ’em.
Keynote 4 – Thriving in a Stochastic World
And this one takes the title for “dullest yet most worth taking note of”.
The speaker reminded me of a professor with a slow drone. I’m glad I managed to barely keep my eyes from shutting, because the key thesis on how to exploit the asymmetry of the upside and downside of experimentation in an unpredictable environment is a great lesson for all start-ups. Heck, all technology businesses.
I would have hated to miss that nugget.
The lead-up to it, I could have done without I think. Maybe start closer to that point and then spend more time on concrete examples, and it’d be a much more relatable talk.
Session 9 – Making Hacking Child’s Play
This one was a lot of fun. Also terrifying.
For opening gambit: a dumb YouTube video showing a terminal with a clueless teenage voice-over explaining how to DDoS someone with a “ping -t” command, and how it’ll only work if “your network is powerful enough”.
A brilliant feint.
Later in the session, the most terrifying thing ever, an early-teen girl, in her early-teen bedroom, speaking into a laptop webcam selling DDoS services to knock your gaming competitors off the net. And completely serious and real.
We live in a world where the tools to disrupt services on the internet can be wielded by the completely clueless. It’s like a phaser; you just point and hit the button and stuff happens.
Very effective presentation.
And also, we’re all doomed.
Another talk that didn’t quite make good on its title, but nevertheless a talk with some interesting points.
Basically, Wotif ended up crawling out of the pit of despair by creating a better deployment story, but rather than using the hard-sell, they developed it alongside their existing deployment path and then let market economy take care of the rest.
“Do you want to go in the release queue, wait weeks, then have your code hit production; all safe and secure, or would you like to use this faster SLIPWay which can turn around your deployment in an hour, but you’ll have to change a few of your assumptions and processes?”
These were the only paired speakers that had put their talk together so that their perspectives complemented each other well. Not flawless, but definitely seamless.
The biggest under-sell of YOW.
The title doesn’t do the content of this session justice by a long stretch.
For warm-up, we walked trough the history of C# (Mads leads the language team at Microsoft), with some miscellaneous barbs and snipes aimed at Georges Saab who did a Java talk before this session.
“C# 1.0, back in 2000, where we introduced value types”, /significant long silent glance to Georges/.
Poking fun was only the secondary purpose of this quick retrospective, because his real purpose was to show the language evolve to where it gained Roslyn. And then he went into a live demo.
He starts up Visual Studio.
He starts a language rule project.
He starts a nested Visual Studio within which the language rule he is developing lives.
He edits the language rule with code-and-resume.
And as he adds this new language rule and it incrementally applies squigglies in the nested VS, and then adds automatic correction options to apply fixes to the code; I want to play with Roslyn so desperately now. It is FxCop on steroids. It is magical. And also a little meta-insanity. But the good kind.
And then to finish he runs through some of the new language feature options on the table for C# 7. Note that hyperlink right there. That’s the GitHub repo where Microsoft keeps the live list of language feature discussions going on for the next version of C#. Microsoft are now not only open-sourcing their framework, but they have also opened up the design process. So have a look around and be part of the discussion!
It sounds like Pattern Matching by types, strong Tuples and non-nullability are strong contenders for features that might be in. But no promises just yet.
I could not have wished for a better closing session, because it sent me into the weekend very energised. I then proceeded not to play with Roslyn for lack of time after my other chores, but I think that flame will burn through a while longer.
My next goal: devise a talk worthy of YOW and get onto the speaker roster.
It is easy to criticise, but much harder to step up and do it.