Hacker News new | ask | show | jobs
by codeflo 1562 days ago
Linked from Tufte’s article, I found this interesting comment from someone’s experience at Microsoft: [1]

> Attempting to have slides serve both as projected visuals and as stand-alone handouts makes for bad visuals and bad documentation. Yet, this is a typical, acceptable approach. PowerPoint (or Keynote) is a tool for displaying visual information, information that helps you tell your story, make your case, or prove your point. PowerPoint is a terrible tool for making written documents, that's what word processors are for.

I think that’s on point for many companies. A lot of the terrible slides you see in meetings are actually intended as documentation after the fact, and few people recognize (or care) that this makes for a terrible presentation.

Ironically, I think Powerpoint isn’t such a bad tool for creating handouts. If the intended reader reads the document on their screen instead of printing it, a nice PDF with screen-shaped pages might actually be close to optimal.

You just have to be 100% clear whether you’re creating a document or a presentation.

[1] http://mamamusings.net/archives/2005/11/19/the_culture_of_th...

8 comments

I don't know. Over time, I've learned to appreciate the kind of slides physicists make, which follow all the "worst practices". They are very dense, with lots of plots and text. They are like, I imagine in the olden days, when people would draw things on blackboards or print out their plots, and sit around them and interpret them. They are the material, and what the presenter does is walking the audience through the material. Compared with a scientific paper, they lack all the prose and weird formal phrases that make papers hard to read. You basically just show what your colleagues need to work with, not less or more. Most importantly, you are not trying to persuade anybody.

My experience outside of academia is that people use PowerPoint exactly for persuasion. Even to the degree that when I wanted to prepare a presentation internally, one of my bosses got angry because he thought I was going to bullshit him. (PowerPoint is something you do to a customer or to the board, not to your colleagues.)

Your boss must have been traumatized by some salesman.

PPTs are great, when they have been written by someone who knows what to communicate and how to do it, which involves knowing your audience. For boffins, slides that are denser than average but lighter than what they would otherwise have to read, are just fine. For the layman, one better err on the side of simplicity and direct impact.

> PowerPoint is something you do to a customer or to the board, not to your colleagues

That is bullshit borne of insecurity. If one can be persuaded so radically by a presentation, one's convictions are pretty weak to begin with.

I've been taught that if you are selling high tech products to C-levels at the biggest companies you want to have highly information dense slides (e.g. condense 2 slides into 1, then condense 2 of those slides in 1) but also have them well organized so you can make a presentation that drives home a few key points.

Personally I am a big fan of powerpoint for making "boxes and lines" diagrams about how a product works, what process we're using, things like that. I love making stuff like

https://www.slideshare.net/paulahoule/making-the-semantic-we...

> at the biggest companies you want to have highly information dense slides

One of my bosses got that memo - the whole thing is one single giant slide with 12 indicators or so. Their sessions are hell, they make people roll eyes and disengage out of boredom. The message I get is that they either don't really care to interact with us hoi polloi, or are just bypassing their own inadequacy at managing a deck at speed - which is understandable, but at that point why are you even there?

IMHO, if one has 6-9 segregated boxes on the same slide and is going to describe each one, they could have been 6-9 slides with exactly the same content but bigger and more impactful. But people have grown to fear the whitespace (can't help themselves overfilling each slide) and transitions (the dreaded pause and meaningless "let's read the title" which is horribly endemic in the States), so they go for this compromise. The impatient reads the whole slide and tunes out, but is not annoyed at you for keeping him from doing that, so it's considered a win. Still, he's not listening to anything you're saying - you could have as well just sent an email.

(It might well be just a fad. C-levels love fads like your next person. In fact, they often love creating fads, just to watch the monkey dance.)

IMHO, a slide should have the elements you're linking together to make a single point. If the point is "sales are good but margin is bad", you'll need a few graphs in the same slide, for sure. But you need transitions at some point to keep people awake and focused on the train of thought you want them to ride. Even in a TED talk, where you really just want people to look at you and only you, the occasional flash of light from transitions is necessary to keep the senses engaged.

>My experience outside of academia is that people use PowerPoint exactly for persuasion.

It happens in my corner of academia too, and I'm all for it. When I give a talk, I'm trying to persuade people to read my paper and ask me interesting questions about the work. The paper has all the gory details. The point of my talk isn't to duplcate the written text, but to make it accessible enough that people want to learn more.

100%. One of the things I'm drilling into my team (we do a lot of slides!), is you have to decide - are you making slides for:

1. presentation - should be sparse, key, anchoring data that enables people to ground themselves while they listen and pay attention to YOU

or

2. Reference - dashboard and background and details and density

If a deck may be used for both purposes, intentionally or not (e.g. we expect it to be shared and referred to by people not attending), resist the urge to do a hybrid slide, and if at all possible make presentation slides at front, reference/supplementary slides at back.

Worst situation is when somebody uses what are effectively reference slides for a live presentation. People's attention is split between reading and making sense of dense information on screen, and the key important points you are trying to verbalize.

"resist the urge to do a hybrid slide, and if at all possible make presentation slides at front, reference/supplementary slides at back."

This has been my SOP for years, but its amazing how resistant some folks are to it (like $NUMBER_OF_SLIDES is the sole measure of quality...with "more slides are bad, mmmkay" being the default). I fought with my last boss on this endlessly, as I would rather have many slides with small digestible chunks of info that I could scaffold my presentations through at a brisk clip, and they preferred as few slides as possible, with dense text and charts, that they would elaborate on in long exposition.

I always got better positive feedback, to no avail. My decks would be 'edited' into a compacted slop-fest. Drove me batty.

One effective way of juggling the two types of slides I've found is to stick the reference slides into an appendix that isn't part of the actual presentation - but can be quickly pulled up during ensuing discussions, and for the benefit of people reviewing the slides later.

High-level stats during the actual presentation, meaty details available upon request.

Most slide decks I ever see I see long after they are presented. As a result I want slide decks to stand alone -- that or else video to be published instead of slide decks.

As a presenter I have to worry about whether my slides will be seen w/ or without video of my presentation. Therefore I try to make my slide decks able to stand alone, though I admit this is not great.

The common pattern now instead is to make colorful and near-content-free slide decks where the presenter simply speaks with colorful backgrounds. I am terrible at crafting such slide decks -- well, I've not tried. This pattern works while presenting though, and I should probably adopt it. Or perhaps I should adopt the Jeff Bezos approach.

What presenters and audiences need is a commitment to publish video of presentations, including Q&A segments. And video has to be playable at 2x speed.

(Yes, almost every video I watch I watch at 2x speed. Where platforms allow faster playback speeds I've even gone faster. Speech is extremely low-bandwidth. This does mean I've to backtrack more often than I'd like, but it's still better this way. I pay attention more when speech is faster -- up to the point where I can't follow, of course, and which I shy away from.)

>Most slide decks I ever see I see long after they are presented. As a result I want slide decks to stand alone -- that or else video to be published instead of slide decks.

I use presenter notes precisely for what you're describing. My actual slides are sparse, but I distribute slides with full presenter notes so all the main points get across in a written form after the fact, without distracting from the talk.

In Powerpoint, I use the "print to PDF" function and tell it to generate pages containing slides with notes.

Yeah, this is a good idea.

I'd add a tip: start writing your notes in the speaker notes and leave the slides blank, instead of the other way around (which is kind of what PowerPoint's UI pushes you towards). Only create the actual slide content after refining your speaker notes once or twice.

For me, at least, I find it's much easier to add content to the blank slide than it is to refine (aka: erase words from) the sentences I initially type out. I tend to like slide content that is graphics (photos, screenshots, graphs, diagrams), titles, or lists of short headline-style text (1~3 words).

> start writing your notes in the speaker notes and leave the slides blank, instead of the other way around (which is kind of what PowerPoint's UI pushes you towards). Only create the actual slide content after refining your speaker notes once or twice.

At that point, you might as well extract those notes into a proper technical report—which is the _actual_ conclusion that Tufte came to:

> Attempting to have slides serve both as projected visuals and as stand-alone handouts makes for bad visuals and bad documentation. Yet, this is a typical, acceptable approach. PowerPoint (or Keynote) is a tool for displaying visual information, information that helps you tell your story, make your case, or prove your point. PowerPoint is a terrible tool for making written documents, that's what word processors are for.

(Aside: Did anyone actually click through and read Tufte's original "PowerPoint Does Rocket Science" (that this piece borrows liberally from—and then perverts)? This piece amounts to middling-quality reblog spam, except worse—because it comes with the added downside that it injects some synthesis about the slide in question (containing "a huge amount of text") being described as too long. Anyone following up and checking sources will see that not only is this not supported by Tufte's work, but it wouldn't be a stretch to say that what is presented here as a summary of the work of Tuft et al runs counter to what they had to say.

It gets truly bizarre when you start reviewing the samples that Tufte provided. It turns out that the slide pictured in this article is not even the slide that was "in play" during the relevant incident. If this blog post were a serious publication (instead of a grey area, slightly scummy, content-marketing afterthought that it is), it would constitute academic fraud...)

Nice, I like that!
There is a great example of this from the fine folks at ITA Software (now Google Flights). Their founder wrote basically a paper on the challenges of writing airline ticket optimization software…in PowerPoint. The slides are thin on information but the speaker notes explain everything.

Since seeing that, I double-down on making my speaker notes tell the story for those that couldn’t be in the room.

Presenter notes though are intended to help the presenter, and their form is not great for users who missed the presentation. As well, the presenter often does need presenter notes that are terse and not suitable for the public. We need slides that are shown at the presentation, slides not shown at the presentation but which are public, and presenter notes.
I think that's the presenter's responsibility. I use the presenter notes to organize my thoughts. By the time I give a talk, I already have most things memorized and only need to grab a few phrases here and there from my (rather detailed full-sentence) notes. The notes easily stand on their own.

It's certainly an improvement over the usual ineffective way people make slides and give talks: by reading bullet points off the slides themselves. I keep over 90% of the text for my talks in the presenter notes. It's equally useful for prepration and for future reference.

Video is of course not necessary, just sufficient.

What you want are the talk notes, something you can read to understand what the shape of the talk was, what context each image came out in. The video of the actual talk substitutes quite nicely for these notes.

It would be preferable if more presentation media had a “front of card/back of card” approach where the card itself specifies how long it thinks it should be on the screen (evaluable over “timing runs”), bullet points for the speaker and readers of the printed form, as well as the image that will be shown.

> Video is of course not necessary, just sufficient.

It should be so, but that means writing different materials, one for the actual presentation, and one for everyone who couldn't be there. It would be better to have the video.

Mind you, for accessibility and general user-friendliness, you want both. But audio/video is a great help because there will be things that the off-line materials miss.

> It would be preferable if more presentation media had a “front of card/back of card”

Yes. PowerPoint and the like generally have a public / presenter distinction, but we want three distinctions: online / off-line / presenter.

I work in the medical field and my brain cells have been constantly assassinated for years by bad presentations. Every once in a while I find unicorns - people who tell a story, teach and present using Powerpoint/Keynote slides as support - conveying information that is hard to put into words -> a photograph, an animation, some simple flowcharts to refer back to whilst talking about what is really the subject of the matter.

I've seen Powerpoint since medical school evolve from trying to squeeze as much information as possible on slides to utter nonsense and hundreds of words per displayed page. Heck, just stay home and automate your presentation to display the slides and mail them to your students instead, don't do this.

I've also been guilty when having to turbo half ass wing a teaching session of stuffing a lot of text in my slides. But I was aware this was not appropriate and I always tried to not be "that person". However, when you have 30 million presentations per day around the world, this is almost impossible to do.

This. I remember the day when I casually asked my cofounder, an amazing storyteller, how his day was.

He answered: „Awesome. I had 8 slides this morning. Now I have three.“

That day, something fundamentally clicked in me.

I learned that day that condensing and sharpening information towards a punchline is real, hard and meaningful work.

Or, as Mark Twain once said, “I didn’t have time to write you a short letter, so I wrote you a long one.”

It was Blaise Pascal in his Lettres Provinciales.
For an in depth review of the phrase: https://quoteinvestigator.com/2012/04/28/shorter-letter/

I am partial to Woodrow Wilsons' variation on the theme; something resonates when I'm asked to for a development/programming estimate.

Powerpoint even has a special mode for creating handouts, by printing a non-displayed notes section. There's no need to put the outline on the slides... other than being less effort.
What's are "screen-shaped pages"? My screens are all kinds of ratios and orientations.

I wish pdf would die. It's a left over vestage of a paper world. It's also left over from a non-connected, non-international world A4 vs US Letter.

PDF really has no place in this smartphone first world

Pdfs are terrible for letter style docs on phones (reflowable text is better, say in emails), but they are great as slides. Certainly easier to distribute or display than proprietary PPT editors.
Reflowable text is fine for conveying information on a tiny screen, but that’s about it.

When your screen is bigger than a page, reflowing the text is almost never what you want because long lines are hard to read. Thus even though computer screens aren’t limited to the size of a sheet of paper, you want to wrap your lines at paper-width (or better, book width) anyway.

Note that programmers typically define an effective page width for source code by enforcing an 80-100 column limit.

> PDF really has no place in this smartphone first world

:shivers:

Makes me think of being forced to dig a hole with tweezers.

"a nice PDF with screen-shaped pages"

Right, but you can create such a thing with a word processor. Just change the page size or orientation.

Sure, the page breaks might not default to where you want them, but you can add page breaks wherever you want, and don't need to create a text box for each paragraph.

More importantly (to me anyway), copying large swaths of text is so much more reliable in a word doc. Copying text in a pdf (derived from PowerPoint no less) is a special hell.
Unbelievable, I've never copied text from a PDF without some error. It's baffling how broken that system is compared to say, text selection in web browsers.

If you start doing fancy CSS tricks you can break that too, of course.

In a PDF, every single glyph could be independently positioned. Most aren’t quite that pathological, but frequently every text run is independent.
A PDF can contain metadata that correctly sequences the characters which is great for cut-n-paste, accessibility and reflowing the text. That metadata is frequently absent and the mechanism for incorporating it is by no means simple.