I gave a talk at EnthusiastiCon on the weekend and it was really fun. The format is a bunch of 10 minute talks and a 40 minute keynote, and this year, it was online-only because, y'know, Corona etc etc.
This post is about things I learned while prepping for / delivering this talk.
Real brief thoughts on the conference
EnthusiastiCon is the sort of conference that you go to for exposure to new ideas, not because you want to “level up your skills in a particular technology”.1
It's ostensibly “for the programming community” but with a very wide definition of ‘programming community’. Because the audience is programmers from lots of different backgrounds, talks can't assume much background knowledge. For example:
- This fantastic talk about AI Explainability requires that you know what AI is, but… not really much more than that!
- This wonderful talk about writing a how-to manual so that Grandpa can use open-source video calling software with confidence will probably be accessible to you if my blog is accessible to you :)
- The keynote talk was about the intersection of tech and politics and how the accumulation of decisions that tech workers make over time affect society, which is relevant to… well, everyone.
I would therefore suggest that this conference is also valuable to the broader technology community – you'd probably get just as much out of it as a technical writer, designer, or product manager.
Things I learned while writing the talk
My approach to writing this talk was to loosely follow Gary Bernhardt's approach in How to Prepare a Talk, which was recommended by the EnthusiastiCon organisers. By “loosely”, I mean, my affinity for procrastination meant that all my timelines were compressed compared to what the article recommends. This turned out ok, though, because writing a 10 minute talk requires less time than the 30-40 minute talks referenced in that article.
Pick one idea, and go deep
In my initial talk proposal, I had to give a breakdown of how I was planning on spending the 10 minutes. I'd initially planned to spend a few minutes talking about demosaicing and a few minutes talking about colorspace conversion, but as I wrote and practiced the talk, I realised that I'd only be able to give an extremely superficial treatment of both of them. In the end, I cut colorspace conversion and spent more time on demosaicing.
This meant I was able to do something special with the format:
- Show that something doesn't have to be scary (“I did this, you can do it too!")
- Show that it also has unfathomable depth and you could keep exploring it for a lifetime and people have literally built careers on top of this thing.
I'm really glad I took this approach! The format works really well for talks that are designed to be a springboard for people who want to learn more, and when someone asked me if there was code on the internet for my terrible demosaicing algorithm, I was chuffed 😁
Be real picky with your details
I'm very easily distractible and love to go down rabbit holes. Earlier versions of the talk had pieces about why there are twice as many green pixels as red or blue in most camera sensors, and about Subpixel Rendering, and both of these were definitely useful to my narrative, but there simply isn't enough time to get sidetracked in a talk like this.
There is an important tradeoff to watch out for here, though – in a practice run for the talk, Nasreen said they were wondering about this “extra green pixels” thing and the fact that my talk didn't answer that was distracting. There's a puzzle here about anticipating the questions your audience members will have so that they can stay maximally focussed on your main idea. My solution to this specific problem was to hand-wave over it more, and avoid drawing attention to the fact that there's so much green 🙃
At the end of the talk, I demonstrate the results of Lightroom's demosaicing algorithm, but because Lightroom does so much extra stuff by default, I had to address that in order to get back to demosaicing. That was a calculated decision – do I burn time talking about something interesting, but ultimately auxiliary? Do I hack TIFF fields out of my RAW files to disable certain Lightroom functionality, so there's less to explain? Do I switch to different software which also implements demosaicing, at the expense of then venturing into a realm I'm less familiar with? In the end, I stuck with Lightroom and explained the other things, because I felt like I could work them in to the central story.
Give yourself room to breathe
The reason why you're being so aggressive about cutting content, is because it's the difference between being able to take your time, and having to have all of your points, transitions, segues, and details incredibly tight. If you have too much content, there is no margin for error. I didn't internalise this until Veit's talk, which stood in contrast to mine with an incredible amount of zen, easygoingness and patience.
Practice, practice, practice (exactly three times!)
I knew this one beforehand, but I was surprised at how impactful practice was! Practice got my 11:30 talk down to 10 minutes – as phrases and ideas crystallized in my brain, they needed less time to explain, and the bad ones got cut. I found recording my talk and then playing it back was great for figuring out which parts I needed to think about more, too.
Throughout the talk development process, I found that I could run through the talk exactly three times before feeling like I'd stagnated and had started talking like a robot.
Towards the end, this becomes even more important: you want the content to stay the same, but you want to deliver it with your full personality, and that means you have to like your personality as you're delivering it, and that's not going to happen if you've heard yourself give the same joke again and again and again2.
Very grateful for support from organizers
So far, this post has been lessons won through experience, but there's a bunch I didn't have to learn because the organizers did such a good job. I want to explicitly call those things out! Do them again next time please :)
- When submitting a talk to EnthusiastiCon this year, I was required to submit an account of how I was planning on using the 10 minutes allocated to my talk. This made the submission process much harder, but once the submission was accepted, I'd already done a bunch of the hard thinking required to boil my idea down. Similarly, the submission process requires you to outline knowledge your audience should have to get the most out of the talk, and because I'd decided this ahead of time, it meant that I didn't have to worry about assumed knowledge as much when actually writing the talk.
- We had a “tech check” session a week or two before the conference which also had slots for people to seek feedback on their talks. This was incredibly helpful for me because:
- Checkpoints are good for preventing procrastination 😅
- I hadn't given a conference talk in like 5 years, so it was a great confidence-booster to hear I was on the right track.
Thanks, organizers! :)
Delivering the talk
Here's some super practical day-of considerations.
Pay attention to lighting, what's in the background, etc.
Lots of other smart people have already written about this, and it's not entirely different from meetings. How to look your best on a video call is a good place to start.
Streaming & Presentation Software for a minimum number of screens
We were using StreamYard for streaming, which was great. Simultaneously keeping my presentation, my speaker notes and the StreamYard speaker display on-screen, though, was extremely frustrating. I ended up using three screens:
- My Laptop Screen; which I wasn't actually looking at, but displayed the presentation
- External Display #1, with the Powerpoint Presenter View
- External Display #2, with the StreamYard controls.
The problem is that Presenter View and the actual slideshow take up an entire screen each, so I couldn't put either of them side-by-side with StreamYard. Keynote has the same problem, but with the added disadvantage that if you alt+tab to a different window, then it exits presentation mode. 🙃
To get it down to two screens, you could either:
- Print your notes and ditch presenter view
- Give up on seeing the StreamYard controls
- Use Google Slides, which, because of the limitations of the web platform works better (web browser windows don't have the ability to force exclusive fullscreen).
To get it down to one screen, you could:
- Use Powerpoint or Google Slides (not Keynote)
- Set up presentation options to present into a window
- Screenshare that window instead of your whole screen
- Live with the fact that you'll have the window title bar visible in your talk.
This seems like a use-case that hadn't been considered before COVID-19, and I wonder if newer versions of Powerpoint / Keynote will support this better.
Have a preflight checklist
There were a bunch of minor things that went wrong in my talk, and they could have been worse if not for dumb luck.
About 7 minutes into my talk, I heard Whatsapp's iconic “Badoong!” sound and realised I'd forgotten to turn off notifications:
This was both audible and distracting, but I recovered, and decided to mute my computer audio, in case I recieved another message. But then, at the end of the talk, I'd forgotten I'd muted my speakers. Moritz came back on screen, and started speaking, and I called out “I think your microphone is muted!” Fortunately, I'd been taken off screen immediately before I managed to get the first word out:
I'm cringing a little just thinking about it. The problem was that I didn't close all the other programs on my computer before starting, because I forgot, because I was nervous.
Having a preflight checklist solves this problem. Here's some things you could put on it:
- tell housemates to be quiet (please)
- close other programs on computer
- ensure ethernet cable connected
- set up lighting (close curtains, turn on lamp)
- check correct microphone selected
- start timer
This isn't a comprehensive list, nor will everything on it apply to you. The point is that you've thought through it all beforehand and are minimising the risk of forgetting something on the day; in the same way that you might make list of last minute things to pack so that you don't need to make decisions on the morning of your flight3.
Finally, practice following this checklist step-by-step once or twice and ensure that it actually leaves you in a good position to do the talk. ✨
Online conferences are great for first-time speakers!
If you've ever wanted to give a conference talk, but balked at the idea of standing in front of a crowd, there's never been a more perfect time to submit a talk! Many conferences are going online. I'm normally a pretty nervous person and I found this whole process much less stressful than I've found public speaking in the past.
Presenting on a stage? In front of hundreds of people?
Presenting at my desk? In front of a webcam?
Much less intimidating 😁
If you are speaking at an online conference, hopefully you've found something useful here! As always, let me know if there's something I've missed or if you want to talk about anything in this post.
This is, incidentally, my favourite type of conference! If I want to deep dive into something, I'd rather study it myself, but I love seeing other people get excited about things I've never heard of before! Beyond Tellerrand, also in Berlin, was my introduction to the genre. ↩︎
On the second last practice run, I called color filter mosaics a “shenanigan” and thought it was brilliant and hilarious. I think if I'd practiced just one more time before the actual talk, it would have been cut – it started to get old, even though it probably wasn't to my audience 🙃 ↩︎
You don't do that? It's just me? Huh, ok. ↩︎