tl;dr … well, honestly, go read something else if you don’t like long form. This is 3600 words of navel-gazing detail, and I’m not about to apologize for it.
A companion piece to my earlier post on the process of designing a poster, this post deals with the talk on the same material for a different conference (vastly different audiences, so I don’t mind overlapping). As I said for the post on designing the poster, this is a snapshot, or series of snapshots, of my process for doing science and preparing talks. It’s not the whole picture, and I’m deliberately exposing the warts and bumps that go with doing science; I don’t get to control the image you form of me as well as I otherwise might, but I feel that the resulting material is more honest and informative.
In any case, I hope you enjoy it. Please leave feel free to leave comments or questions, and I’ll do my best to answer them.
- Monday, July 23, 2012: The ISBE 2012 conference is a couple of weeks away, so it’s time to start thinking about the talk. The initial steps will be a little slow, but today I’ve created the presentation file as a symbolic step. I haven’t yet conceived of the overall visual theme of the talk, so for now I’m adopting a simple black on white approach.
- Thursday, July 26 4:30 p.m. I’ve got about a half an hour before I need to leave the lab to go meet my long suffering wife for dinner. Time to outline some content! I’m working quickly, creating new slides and just typing main ideas of the story I’m telling into them.
- 4:51 p.m. 20 minutes later, I’m done a really quick outline.
A couple of things to note. First, considering that this is a 12 minute talk, you may be wondering if 22 slides is too much. Yes, and no. For most people, 22 slides is too many for this length of talk; a good rule of thumb is – depending on the density of your slides – allow for at least a minute for any slide you’ll be saying more than ‘hello’ over. This is a mistake that I see people make time and time again: they make hugely dense slides with dozens of graphs, and then leave themselves about 15 seconds per slide. This won’t work. They either end up blasting through slide after slide of results, or they go way over time . Aim for simplicity, and remember that simplicity is hard. Simplicity doesn’t mean dumbing down your message, it means presenting your message in as straightforward and audience-appropriate a fashion as possible. On the other hand, I deliberately present more slides with fewer ideas on each one; this is a conscious strategy aimed at controlling what the audience is seeing and thinking about on a more fine-grained level. However, this is a more difficult approach, and you should be careful about adopting it. Long story short, if you have more than about 1 slide for every 30 seconds to a minute, you should have a good reason why. Also, the outline is hardly set in stone. As when I did the poster, it’s an iterative process which will lead to me adding and subtracting material as I get into the content and the design. I’ve already got some ideas that may add in a few slides, so I’ll probably need to subtract some elsewhere.
- Monday, July 30, 4:36 p.m. Squeezing in a few minutes to work on the slides before I head for home. I don’t have a cohesive plan for the design of the slides yet, so I’m going to iterate the content a little and see what suggests itself.
- Tuesday, July 31, 11:30 a.m. Only got a few minutes in on the talk yesterday before I got distracted by an ‘emergency’ (read: time-suck). I just realised this morning that I really need to create two versions of this talk, because I’m going to be giving it at a couple places I’m visiting in Europe after the conference. This means that I need a 12-minute version for ISBE, and a 45-minute-ish version for the seminars I’ll be giving. This isn’t as bad as it looks, because creating the 12-minute version requires cutting out a lot of material that I would otherwise put in; while it makes for more work creating slides for the longer version, it’s more relaxing because I can afford to go into details that I would have to otherwise avoid in the shorter version. This post, however, will focus on the 12-minute version which I will create first.
- 4:20 p.m. It’s been a bit of a slow day, but some of the pieces are starting to come together. I’ve got a few of the visual ideas worked out, and though there is a massive amount of work left to do, at least I’ve got a direction.
You may notice a few things. First, I’ve littered the slides with notes to myself explaining where I want to go with that slide, reminders about content to add or delete, and even notations on which notes might be suitable to cut from the final version. Second, if you look closely, some of these images are decidedly low-res. That’s because they’re “comps” of stock photos (from iStockPhoto), which are super-low-res versions that are watermarked so as to be unusable in a production document. They are, however, useful for trying things out and deciding what image works best before you lay down money for the final image file. This lets me play with the slide deck before committing (an example is the image of the dog and the bat; I’ll only use one when I discuss rabies, but I’m trying them out to see which I like better), and it might even be possible to find free alternatives to the images I’ve used. The final thing of note is that I haven’t addressed the typography of the presentation yet; the font used in the slides so far is Keynote’s default Gill Sans, but my next step is to choose some appropriate fonts now that I have a bit of content in place.
- 2:36 a.m. I’ve been working for the last three hours transcribing every common name and genus-level-or-above taxonomic name from the index of Odling-Smee et al’s monograph on niche construction in an attempt to set the stage for why I’m giving this talk; namely, that viruses are under-represented here. To make this point visual, I’m turning it in a word cloud (you can see the placeholder I whipped up in the slides above). I’ve reached the T’s and I have to stop now because otherwise I’ll be doing this all damn night.
- Wednesday, 12:26 p.m Back to it, and I’ve finally finished the index. Now to throw it into R (using the “wordcloud” package), pretty it up, and insert it into the talk! (And yes, I *will* go way too far for a detail no-one will care about).
- 1:06 p.m. Here’s the new placeholder that I’ve created in R. It’s still a placeholder because I’m going to try to match the fonts and colors to the rest of the slides; making those decisions is the next step.
- 4:25 p.m. I’m ‘auditioning’ some font and colour scheme choices. To do this, I’ve duplicated my presentation and slides with in it, and I’m applying various styles to see how they work. I’m looking for a bold, attention-grabbing combination, because I want this to stand out from a sea of similar-looking talks; since I’m not adopting any sort of high-concept approach for this talk (mostly due to a lack of time!), I’m focusing on using typography and colour in a more aggressive way than is usual. With that said, I could really use my wife’s designer eye on this, because I’m having anxiety attacks over what combinations might work. I like the use of Bebas Neue and a script font for the headers, but I’m having trouble with a body font (because neither of those choices work well as body fonts). I’m in a bit of a grey area because the presentation really only has a couple of blocks of text that need to be set, so I need to balance readability with mood.
Incidentally – and this is important – I’ve also been ducking into an unused conference room with a project to try this out on the bigger screens. Always try your talk slides out on a setup that is as close to the final venue as possible. You want to make sure that the colour combination that looks great on screen actually works when you project it!
- Thursday, 4:36 p.m. I’ve been working on the slides throughout the day, in and amongst other things on my todo list. Today I’ve been focusing on the results section, which has seen some progress.
I’ve made some subtle modifications, including breaking the green color of the palette into a brighter green for text on black slides (like the title slide), and a softer green for backgrounds. If you compare this snapshot to the previous one, you should be able to see what I mean. Also, I’ve started redoing my figures to use the fonts that I selected for the talk. It’s a small thing, and perhaps no-one would consciously notice, but I believe in minimising friction for the viewer; different fonts and designs between parts of the talk can be jarring even if the audience can’t figure out why, and I want to avoid that as much possible. It may not be entirely doable (I still have to figure out a better way to present that tree, for instance, and I’m not sure if I’ll be able to find a way to change the font on that), but I’ll go as far as I can to homogenise the design.
- 4:52 p.m. I’ve ducked into the conference room to check on how the slides are showing up on the screen. I’m generally happy with it so far, but projecting it makes it clear which version of the word cloud I’m going to keep; the script version is painful at large sizes.
- 11:04 p.m. I’m continuing to work on the slides. I’ve been going back and forth between the bat picture and the dog picture for rabies (another potential example of viral niche construction, methinks), but now it finally occurs to me that the dog picture just doesn’t read well to anyone but me. So, it has to go.
- 11:55 p.m. I’m working on a slide that suggests a speculative link between viral niche construction and sociality; this is based off of work on a cat virus, so I’m using a picture of kittens to illustrate the point. My first version, though, illustrates a design issue: if you use a picture that has eyeballs in them, the rest of the slide has to relate to the eyeline (somewhat similar to the concept of eyeline matching in film editing) or else the viewer gets uncomfortable.
As you can see, the kittens are looking down and I have text above them; this creates a visual tension that has no reason for being there. Putting the text below the kittens, besides looking bad because of the shading at the bottom of the photo, also fails because the kittens are all looking in different directions. Once I’ve identified this problem, I have to find a new photo; thankfully, the internet seems to be big on cats (who knew?).
- 12:09 p.m. I’m wrapping up for the night. I’ve made reasonable progress today: aside from a set of slides in the middle that I’ve engaged my wife to do drawings for, the last thing that I need to do for this first, rough version is to redraw the phylogenetic tree and find a way to present it.
If you’re paying attention, you’ll notice that I’ve still got too many slides. I’m going to be practicing this talk (including a lab practice talk next week), but it’s almost certain that I’m going to need to cut some material. Like any other content editing, there’s going to come a point where I have to kill my darlings. This doesn’t bother me as much as it normally would, because most – if not all – of what I cut will end up going into the longer seminar version of this talk, where I’ll be making the same case in greater depth. You can see that I’ve already started doing this, as I’ve moved some slides after the acknowledgements at the end; these will be included in the longer version unless I cut them entirely.
- August 6, 2:26 p.m. I’ve been fiddling with the slides over the last few days, just trying a few things out and moving things around. I’ve decided on one of the cat photos, the middle one, as it’s the most engaging; my wife pointed out that this is because of the way they’re looking, including the one staring straight at you. I’ve got her working on producing a diagram for me to explain the way baculovirus manipulates its hosts, which goes in the blank spot in the middle, and I’ve placed images in there to help get me over the hump. Today, I need to fix the phylogenetic tree and place it in; whether I use it in the short or long version, I’ll need it at some point. And I want to get the short version done tonight if I can, because I plan on practicing it tomorrow before I present it to the lab on Thursday. So here’s the current state of affairs:
- August 7, 2:15 a.m. Small refinements now. Unfortunately, even in consultation with my talented wife I couldn’t come up with a good illustration for the slide I’ve been holding on the various genotypes; thus, I’ve decided to break down and use (gasp) text. I know, I know. In the mean time, I’ve also managed to refine the tree diagram (which requires further refinement, but the pieces are there now).
- 1:02 p.m. I’m searching for images to illustrate the hypothetical genotypes (zombie, non-gooey; non-zombie, but gooey). I’m having trouble meeting the criterion of Creative Commons or stock that I can purchase as well as being the right image for the idea.
- 3:02 p.m. I’ve found images and replaced the phylogenetic trees. I’ve also replaced a slide that I apparently deleted at some point along the way without noticing; you’ll notice that the second slide in the talk is missing if you compare the last two snapshots above. Using OS X’s Versions, I was able to graphically browse to an old version from a couple of days ago, find the slide, and drag it and drop it directly into the current version of the talk. It may not be git, but it’s still cool. And it’s also a good lesson: keep old versions! Keep backups!
I think that the short version of the talk is in good enough shape now that I can practice it, so I’m going to go see if I can find a room with a projector to play in. If you can, it’s best to practice talks under conditions that are as close to the real thing as possible; that means standing up in front of a room, even if it’s empty, and playing your slides behind you as you address the room. Muttering under your breath as you stare at the slides may seem like a good way to practice, but you’ll never find the timing problems and flow issues unless you force yourself to stand up and actually talk.
- 4:52 p.m. I just finished practicing my talk for the first time. As I expressed on Twitter:
Seriously, people. Practice your talks before you give them. Then, practice them again. And then three more times. What I’ve learned is that I need to do some rearranging, because the flow of ideas in the talk didn’t quite work; I’m going to jettison a few slides and use them in the longer version, and I’m going to see if I can add a few elements to the text that I abruptly noticed were missing.
- August 8, 12:24 a.m. I’ve spent some time rearranging slides and writing down what I want to say on each slide. I like to have my material memorized to the point where I can present it without notes, but I sometimes find that writing down key points of each slide when I’m practicing helps me to achieve that goal. Here’s the current state of the short version, with changes incorporated.
I’m still struggling with some aspects of the design. In particular the genotype slide (slide 17) is bugging me; I had to add the model diagram because it was too difficult to explain the genotypes by referring to the parameters alone. Now that I think about it, though, I may try playing with text instead spelling out the assumptions. But that can wait until tomorrow, because I need some bloody sleep.
- 11:49 a.m. Back to the conference room to practice again!
- 12:42 p.m. I tried it three times, but I’m still coming in too long. The talk is supposed to be 12 minutes with 3 minutes for questions, and I’m clocking in at 18-19 minutes. It looks like I’ll need to pare some things down to put into the longer version. It breaks my heart, but I think that I’ll have to put the word cloud into the longer version; it’s a great image, but under time constraints it’s not pulling its weight. When that happens, you need to kill your darlings.
- 1:07 p.m. I’m cutting it to the bone, but I’ve got things down to 20 slides (simplicity is hard). The room I was using is booked right now, I’m going to have lunch and do some work until it’s open and I practice again.
- 4:35 p.m. I’ve practiced this thing backwards and forwards, but I can’t get the time down! From 19m 28s to 14m 12s, I’m still two minutes over. I may have to remove the phylogenetic results, though it kills me to do so. I know that they’ll be in the longer version where I’ll have plenty of time to go over them, but it still pains me.
- August 9, 12:49 a.m. I’ve spent the last couple of hours finalising the design, including replacing all of the comp images with the full versions that I’ve purchased. It’s pricy ($86 AUD for 50 credits on iStockPhoto), but worth it. If you can’t afford to pay for good images, then find them under a Creative Commons license on Flickr, or take them yourself. But always use high-resolution images! And don’t steal them.
- 11:58 a.m. Okay, further practicing yields no advances. I’m going to have to cut the phylogenetic results in favor of asking people to talk to me if they’re interested.
- 12:45 p.m. 11 minutes, 58 seconds! Finally, we’re ready. Here’s the state of the talk before I give it to the lab this afternoon. Don’t forget that I’ve got extra slides tacked on (after the slide with the big Thanks! on it). I’ve also added a slide with photo credits; again, acknowledge your sources and don’t steal other people’s work.
- 4:45 p.m. Well, I gave it to the lab (and a distinguished visitor!), and things went pretty well. It’s clear that the work I put into the design and practicing the talk has paid off, because I received multiple comments that it was a very polished talk. There were some good questions, and a couple of good suggestions for minor improvements, but otherwise it’s done and dusted!
- August 11, 12:39 a.m. I leave for the conference tomorrow afternoon, and I’ve just thrown my talk files onto my USB drive – and I’ve got them in my Dropbox, on my iPad, and in my email. You only have them in one place? You’re begging for a disaster. But, I digress. At this point, it’s worth reviewing the lessons I learned while designing this talk. First and foremost, as I wrote above, simplicity is hard, and you have to be prepared to kill your darlings. I had more content than I could present, so I had to cut it down and make it as simple as possible. Practice is king. I practiced this talk no fewer than eight times to an empty room, and it paid off; the people I finally gave it to were impressed at how fluent I was. What they didn’t see was the hours I spent stumbling and swearing and fumbling my words. If you suck in private, you’ll be great in public. And finally, iterate, iterate, iterate! To make good posters and good talks, you need to advance and revise, create and critique. If you scan back through this post and look at nothing but the slide pictures I’ve included, I hope that you’ll get a feeling for this.
So, if you’re still reading after all of that, thanks for sticking with me! I hope you learned a little something, and I welcome your thoughts. But for now, I’m off to Sweden!