Accu 2014 Conference Notes
I had the chance to go to ACCU 2014 the other week (full conference schedule is here) and I have to say it was one of the best conferences I’ve had the pleasure to attend. And while it did confirm my idea that C++ is getting the point of saturation and ridiculous excess (C++11 was needed, as a result so was C++14, but C++17… Just use a more suitable language if these are the features you need), the talks I went to, on the whole, we’re great.
So I thought I may as well post up the notes I made from each talk — and while they might be a bit of a brain dump of the time, if there’s something that sparks your interest, I’m sure the full presentations will be posted up at some point soon.
Get Archaeology
Charles Bailey, Bloomberg LP
Interesting talk looking at some of the more esoteric ways of presenting and searching for information within an existing Git repository. Unfortunately, time was short so the speaker had to rush through the last part, which was, for me, the most relevant and interesting part.
Designing C++ Headers Best Practices
Alan Griffiths
For most experienced C++ developers the content of this talk is probably quite basic, as you’d have a large portion of this covered already through sheer trial and error over the years. But clarifying some good practices and interesting side effects was enjoyable.
Version Control — Patterns and Practices
Chris Oldwood, chrisoldwood.blogspot.com
Highlevel overview of the various patterns we use when using version control (focused mostly on DVC), some useful examples and some interesting discussion about trust issues…
Performance Choices
Dietmar Kuhl, Bloomberg LP
Proof that if you want to know about optimisation and performance, stick to asking people in the games industry.
I’m not posting these notes.
Crafting More Effective Technical Presentation
Dirk Haun, www.themobilepresenter.com
Really good presentation on how to craft good presentations — some interesting discussions on the make up of the human brain, why certain techniques work and why the vast majority of technical talks (or just talks in general to be honest) do what they do.
The Evolution of Good Code
Arjan van Leeuwen, Opera Software
Great talk, not telling us what good code is, but examining a few in-vougue books over the last decade to see where they sit on various contentious topics. Note that when the notes say “no-one argued for/against” it’s just referencing the books being discussed!
Software Quality Dashboard for Agile Teams
Alexander Bogush
Brilliant talk about the various metrics Alexander uses to measure the quality of their code base. If you’re sick of agile, don’t let the title put you off, this is valuable reading regardless of your development methodology.
Automated Test Hell (or There and Back Again)
Wojciech Seiga, JIRA product manager
Another great talk, this time discussing how the JIRA team took a (very) legacy project with a (very) large and fragile test structure into something much more suitable to quick iteration and testing. When someone says some of their tests took 8 days to complete, you have to wonder how they didn’t just throw it all away!
Why Agile Doesn’t Scale — and what you can do about it
Dan North, http://dannorth.net
Interesting talk, arguing that Agile is simply not designed to, nor was it ever imaged to, be a scalable development methodology (where scale is defined as bigger teams and wider problems). Excellently covered why and how agile adoption fails and how this can be avoided to the point where agile principles can be used on much larger and less flexible teams.
Biggest Mistakes in C++11
Nicolai M Josuttis, IT-communications.com
Entertaining talk where Nicolai, a member of the C++ Standard Committee library working group, covers various features of C++11 that simply didn’t work or shouldn’t have been included in the standard. When it gets to the point that Standard Committee members in the audience are arguing about how something should or does work, you know they’ve taken it to far.
Everything You Ever Wanted to Know About Move Semantics (and then some…)
Howard Hinnant, Ripple Labs
Detailed talk on move semantics which are never as complicated as they (and even the speaker was) are made out to seem. There’s some dumb issues, which seem to be increasing as the C++ standards committee don’t seem to understand the scale of the changes being made, but never-the-less it was a good overview of what is a key improvement to C++.