Earlier this month I attended the 2015 edition of the North American Yet Another Perl Conference in Salt Lake City. This was the second YAPC I had ever attended (and, happily, the first I attended without dividing my attention to deal with emergencies unfolding elsewhere). I arrived early, and I have already written about the couple of days I spent eating and sightseeing before the conference started. Now I will write about the conference!
Actually, I will write about the venue first. I loved the venue, the Litte America Hotel, especially compared to that of 2013’s YAPC::NA in Austin. That year’s YAPC enjoyed an remarkably low attendance fee — I think it was only $75 — but attendees ended up paying in other ways, most notably in the lengthy hike through the harsh Texan sun between the Doubletree Hotel and the clutch of UT buildings that held the conference’s events. (And then a further hike, or a lengthy drive, to food or drink or anything else interesting around the city.) I skipped last year’s YAPC in Orlando (I just couldn’t afford it after 2013’s financially draining family matters), but I’m told that it resembled this year’s in both price and in venue quality.
Simply put, everything happened in the same hotel that most attendees slept in. This introvert found especially great value in the ability to fade to my room whenever I wanted at the cost of a three-minute walk. (And then jump back into the game just as quickly, once my people-coping battery had sufficiently recharged.) That said, I found lingering on the conference floor easier due to the event’s thoughtful provisioning: hotel staff refreshed a shifting range of drinks and snacks in the conference’s common area throughout all three days of talks. It felt like a professional conference that cared about its attendees’ comfort, both mental and physical, and teamed up with a venue quite happy to help accomplish this. Even though this meant a higher attendance fee, I — a professional, far from home — appreciated this tremendously.
In my first draft of this post I dug into writing about every talk I attended, but that threatened to turn into an ever-more fiddly post I’d never finish. Thus, I shall instead take a post-mortem-style tack, calling out three things I loved from the talks, and three things that I think could have gone better. Naturally YAPC::NA’s talks held far more than a mere trio of high notes for me, and of course I’ve still only seen a fraction of the talks — so I hope very much you’ll join me in watching them all on the YAPC::NA youtube channel, immediately and freely available to the public thanks to the efforts of the conference’s heroic A/V team.
The good parts of the lightning talks were good. Favorite beats:
SawyerX talking about the negative growth of a key Perl project in terms of codebase size, and receiving applause.
David Golden’s unexpected and convincing argument to syntax-highlight code comments as brighter than actual code, to help encourage minimal-comment hygiene.
A brief explanation of plans to more actively provide stewardship to CPAN, Perl’s ancient and enormous shared-code repository that grows increasingly prone to dizzying levels of unchecked internal cross-dependencies.
One organizational detail about the lightning talks impressed me especially: all the conference’s lighting-talk blocks featured a separate queue for those with promotional announcements (largely want-ads for Perl talent or local user groups), and they’d control the mic between talks while the upcoming speaker went through their unavoidable minute of A/V fumbling.
Less impressively, many lightning talks involved little more than a speaker flipbooking through slides of code and explaining what that code was doing. I don’t remember what any of these were about. Lightning talks work much better for show-and-tell, not high-speed deep dives. They should make a single point swiftly, with energy and humor. Too many of these instead felt like listening to a tech podcast at 2.5x speed. I don’t know if there’s anything to be done about this.
Daisuke Maki’s talk about organizing YAPC::Asia was so good. I took pages of notes during this excellent description of the YAPC::Asia team’s reinterpretation of what “YAPC” means, recasting it as less a Perl conference and more a conference of all kind of topics interesting to the technically apt, which just happens to be hosted by the Perl community. (They do use the presence of Perlish subject matter as a tiebreaker when choosing between two equally valid content-contestants, but otherwise don’t seek out Perl-specific stuff.) In this way and others, they attract thousands of attendees who would have not felt otherwise interested.
Daisuke broke down the financial, organizational, and logistical challenges of running a larger-than-small conference, including the reasons to seek sponsors — and how to land them, once you know you need them — and what what should do about the fact that a venue’s free WiFi is always terrible. (Spoiler for the latter: YAPC::Asia has a 22-member network-maintenance staff.)
This talk really inspired me with ideas to improve IFComp’s reach and visbility, despite all the differences between the venueless, weeks-long event I manage and a focused and physical technical conference. I strongly recommend anyone who runs (or is considering running) an non-trivial event to give this one a watch.
Nick Patch gave us a good talking-to about storing and displaying names properly. Everyone who manages an application that collects the most basic of peoples’ public-facing personal information — their names — ought to have a look at this talk. It went deeper into the technical affordances and requirements of Unicode than I expected (as I had expected it to focus mainly on social awareness), and I welcomed this.
Another take-away from this talk I really enjoyed, regarding the thorny issue of collecting gender information: your organization’s first question to itself should be, Do you really need to know that in the first place?
(I also just realized that “Nick Patch” sounds a bit like a possible commit message regarding this very topic. But, enough of that.)
I can no longer hide from the reality that smart hackers automate everything. (Yes, this is the fourth thing I liked. Let’s count the lightning talks as Thing Zero.) Over and over, from that CPAN lightning talk through MST’s talk on build management and then Florian’s talk about Ansible and beyond, YAPC talks assumed that a modern professional software engineer operates from within an automation-rich environment, comfortable with software such as Vagrant, Puppet, and Ansible. One so equipped can make entire fleets of remote machines engage in beautiful synchronized dances by writing short scripts on their laptop, or instance a new, perfectly project-configured VPS with a few keystrokes and a coffee break. I… cannot do this, myself, quite yet.
I have the good fortune to work with a colleague well-versed in all this stuff, and who has been for years — long before, I believe, this style of work started arcing towards ubiquity. So, I had some advance warning. But its tacitly assumed presence in the YAPC audience’s toolkit really struck me this year, then sat on my chest, and stared into my eyes until I felt ready to admit that I can’t ignore this technology any more. At said colleague’s recommendation, I begin my journey right now with Ansible. It’s too late to roll it into my current freelance project, but by year’s end I hope to have several “playbooks” tied between my fingers and the various remote servers entrusted to me, both personal and professional.
And now, some aspects of YAPC::NA 2015’s talks I think could have gone better:
The heckling needs to stop. A few talks felt marred by people in the audience shouting “clever” talkback to the speaker. This seemed to happen more often in cases where the speaker had a reputation for humorous talks, as with Matt S Trout and Larry Wall. The latter case became especially egregious, with at least one audience member feeling free to shout out both questions and answers during the Q&A between Larry and host Ricardo Signes, disrespectfully ignoring how Ricardo had spent care and effort constructing a list of community-collected questions for Larry in the days leading up to the conference.
I expect that hecklers see their shout-outs as welcome and appropriate contributions, offered in the same humorous spirit as the talks they “enhance”. This view, unfortunately, does not reflect reality. Heckling selfishly grabs attention away from the speaker, the person the audience has gathered to hear, and wastes the time of everyone else in the room as the speaker decides how to respond to the unprompted and unwelcome derailing. Ironically, it can become most disruptive when the speaker takes the super-polite tack of treating shouted outbursts as invited questions worthy of immediate consideration, leeching time away from the actual planned talk that everyone came to see.
I don’t like after-talk Q&A sessions, which I consider the comments sections of conferences, and I really, really don’t like heckling. Please, just don’t do it.
I found one of the keynotes weird and embarrassing. “What Perl Taught Me about Life” fell flat for me, in part due to what I saw as the speaker making some poor assumptions about the level of culture shared between himself and the audience. I don’t even remember what the talk was about, frankly; all I remember is the presence of several minutes-long film and TV clips that the speaker repeatedly ceded control to, turning up the audio and letting the room have a little YouTube break. I got the impression that we were supposed to recognize the clips’ provenance and enjoy seeing something familiar in a new context. But, except for a couple of bits near the end, they were all new to me, and thus simply perplexing.
In one case, perplexing turned to downright embarrassing. We watch a white guy walking with his ladyfriend run into an old flame, and tensions rise — so he directs another guy wearing a dime-store Viking costume to bash both women over the head with a club, sending them sprawling lifelessly to the sidewalk. I feel quite willing to accept that this scene worked as absurd comedy in its original context, whatever that may have been, but here it seemed nothing other than bizarre non sequitur at best, and completely inappropriate for an inclusive tech conference at worst.
The male-to-female speaker ratio was not notably better than typical. A perennial difficulty with technical conferences, and one that I know YAPC::NA’s organizers feel very interested in. While the 2015 conference did indeed have several excellent talks presented by women, they remained vastly outnumbered by male speakers. (Which implies the very true silver lining: boy, did YAPC have a lot of great speakers this year! But, yeah, still.)
Larry Wall himself took some of the edge off this fact’s disappointment during his Tuesday Q&A, stating his goal to see Perl 6 adopted by the next generation of programmers — “that’s why I designed its logo to appeal to seven-year-old girls.” This won applause, but I do hope that YAPC can continue improving its gender imbalance on a bit more aggressive timetable than Perl 6’s.