Close Reading PowerPoint
This text draws on our full-length study of PowerPoint as Occasion for Speech (under review) and on a collaboration with Johanna Drucker on intellectual middleware.
This is a draft version/work in progress document. Comments and feedback are very welcome! Please use the commenting function in Medium (the publishing platform used) or send an email to firstname.lastname@example.org. We would like to thank Lauren Klein for comments on an earlier version.
This paper is an attempt to explore some of the organizing principles and ideational underpinnings of PowerPoint through looking carefully at the level of the interface. We argue that this material attention — which is not restricted to the software, but extends to situations in which the software is being used — is needed for a fuller understanding of the role played by the software and the way it conditions presentations and presentation situations.
In particular we focus on two key aspects of PowerPoint:
- The integrity and role of the slide. If the slide is the main organizing principle (of presentation situations in the world), how is this enacted through the software?
- The separation between the production and presentation modes. The success of PowerPoint can partly be attributed to the standardization of presentation situations and narrative strategies. How does the separation of production and presentation modes in the software support this standardization?
Many of the design considerations of PowerPoint (and Presenter) have been documented and archived by Robert Gaskins. A useful basic document is “Presenter Specification, May 22, 1986”.
It was very clear from the beginning that the slide was the organizing principle:
Presenter is a program designed primarily for producing overhead slides (in color and in black and white), and can be used also for producing 35mm slides and video presentations
“Presenter Specification, May 22, 1986”, Austin, Rudkin and Gaskins
The idea of mapping the slide (as presented in Presenter or PowerPoint) directly onto the screen in a 1:1 relation is central to penetration of the software. This presupposes that the slide (as overhead slide or as projected slide) normally does not contain any of the machinery of the software. The slide does not show the interface. In the Presenter specification it is said that the slide show mode displays slides one at a time and that: “No window borders, scroll bars, menu bars, etc. are shown, only each individual slide”. This removal of the operating system had not been done by any other Macintosh software, and Gaskins describes how Apple had to be persuaded to allow this for Presenter (Gaskins 2012).
There can be no stronger assertion of the slide as the organizing principle than letting the slide take up all the presentation space. The frame of the presentation recedes and the display becomes the slide. In terms of narrative structure, the single-slide-at-a-time display — without any material sense of previous or future slides — has a strong conditioning power. And because the production and presentation modes are kept separate, there is little risk of the production tools influencing the narrative strategy.
The Slide as a Building Block
Slides form the smallest unit across PowerPoint as a software platform. They make up the language of the platform. This focus is a considerable strength as slides are appropriately sized chunks of rhetoric that are mapped onto the exact size of a presentation screen. The slides create a one-to-one connection between what is shown on the local screen and on the display screen (whether run by a computer, slide projector or an overhead projector). They also make certain operations more difficult. For instance, it is difficult to make arguments outside the conceptual space of the frame. Furthermore, the slide as building block also influences underlying data structures. There is no simple way to get at PowerPoint-based content outside the individual slides — for instance seeing all the included images at the same (Drucker and Svensson). The content sits inside the slides and while slides can be copied and individual content (e.g. an image in a slide) can be duplicated through copying it from one slide to another, there is little control of the database. PowerPoint has no proper repository function and it cannot handle cloud-based content in any convincing way.
Illustrating the Idea of Organizing Principles
Franco Moretti argues that the paragraph “is the forgotten unit of the articulation of our thought”. Moretti and his colleagues argue that the paragraph falls somewhere in between the very large and the small.
Their interest is mainly analytical (using topic modeling to explore the meaning of the paragraph). Let us now illustrate different organizing principles and the importance of their material implementations through imagining a word processor program organized around the paragraph as the building block.
In the main production view of this tool, one paragraph at a time would be visible. There would not be any indication of page size, but rather the document would consist of a series of clearly demarcated (and enumerated) paragraphs that could easily be moved around and listed in a slide-sorter type view. The software would keep count of the paragraphs. Any operations between paragraphs would have do be done through the paragraph view — e.g. clicking on a paragraph in the paragraph sorter, copying a sentence, activating another paragraph sorter, inserting the cursor and paste the content (much like moving content between slides in PowerPoint).
Under presentation view (when reading content on a tablet for instance), the paragraph would be shown one a time (with possible transitions between them). The paragraphs would still be the organizing principle although they would have receded to the background in the presentation view (with only one paragraph at a time visible, does the idea of several paragraphs making up a text get weaker?).
One question is whether text produced in a paragraph-centric word processor would be any different from text produced in traditional word processors? And whether the reading experience would be any different?
Trapped Within the Slide
This slide as an organizing principle in PowerPoint is strongly manifested in the interface, where the content creation and presentation processes are centered around clearly defined slides.
In this example the two (empty) slides in the slide deck overview (on the left) are represented as outlined rectangles. The main work area consists of a large rectangular space for the slide. Slides are unsurprisingly the building blocks of the application (unless otherwise specified the screenshots and discussion relate to PowerPoint 2013). It is easy to copy and paste slides, to order and reorder them in the slide sorter and to incorporate content on the slide level from another presentation file.
The slide as a view port
There is no way to simply extend content (pictures, text, film) across several slides in PowerPoint. The logic of the software does not allow such operations and this predisposition is manifested in the material interface. For instance, when content is made larger than the slide space, it will expand to the local work area (for that particular slide). Content outside the slide space will not show in the presentation view and the slide then functions as a “view port”:
If this presentation was run in presentation/slideshow mode, it would show what can be seen on top slide in the left-most list of slides (i.e. three full letters — ry l — and two partial ones: e and a).
While PowerPoint does not support zooming to create narratives (which a tool such as Prezi does to some extent), it would be possible to place the same large complex visualization on all the slides (and work areas) and use the slide view port to show different parts of the large image. The first slide could show the very large visualization as a miniature (showing all the content) and the subsequent slides various bits. Given the above example, the first slide would show “Very large text” and the subsequent slides might each focus on one letter. This would be a clunky way of simulating a zooming presentation principle without any way of giving a sense of movement between nodes. This lack of interactive affordance is very significant — PowerPoint allows animations and transitions within its universe of slides, but does not provide interactive modalities that could challenge the sequence of slides as a narrative engine. In the above example, there would not be any natural way of showing the whole at the same time as the parts unless a copy of the large image was sized into a miniature and manually placed in the slide view port on each slide.
This strategy is not very useful in practice, but pushing the software in this way allows us to discuss the conditioning built into the program. We will soon return to the question of the separation between production and production modes, but it is worth noting that the above example would work much better if the production interface (as in the image above) was used for the presentation. The problem is that presentation view does not include production features and that very few presenters use the production view when presenting.
The slide as a high-resolution canvas
Another strategy (for the zoom/scale-oriented story-teller) would be to fit a great deal of content into one slide and using a modern large very high-resolution screen or display wall to show that content. There is some tension and constraints here though. The software was built for low-resolution screens and does a great job of mapping the screen resolution between the setup used for producing content and the presentation setup. When the discrepancy between screens becomes larger, this mapping gets more problematic (see below for a discussion). The content created for a large 4K screen may not work for a standard presentation room with a fairly small low-resolution projector. There are also constraints within the software. The screenshot below shows an example of detailed content within one slide.
What is striking here is that the maximum magnification possible in PowerPoint (PowerPoint 2013, Windows) is 400% (used above). This means that it is difficult to manage precise management of detailed content (it is not possible to get “closer” to the small text than above). This is not an arbitrary constraint (although it could probably be easily changed), but rather speaks of the history and use of the tool. The style of PowerPoint presentations and the infrastructural setups typically used have not required such fine-grained manipulation of content. Here the physical setup of presentation spaces play an important role. Normally the audience is placed at a distance from the screen and participants are not expected to walk up to the screen and engage with it at close distance.
This example demonstrates how the slide perspective organizes narrative structure. A paradigm based on a series of slides does not lend itself easily to alternative narrative principles such as zooming, layering, panoramic or time-based story-telling. While it might be possible to manually tweak the software to simulate zooming and other narrative/navigational principles, the integrity of the slide and the lack of other narrative modalities in software like PowerPoint makes alternative narrative conceptions difficult or even impossible.
Further Engagement with the Production Interface
The rest of the default layout of the interface is also organized around slides. For instance, the status bar includes information on what number the current slide has in the sequence of slides.
This particular piece of information can be important to an audience in a situation where they get to see the interface before presentation mode is activated (it is normally worrying to realize that someone about to do a 20-minute talk has a presentation file with 80 slides). This measurement (number of slides) is only relevant in the context of a slide-based piece of presentation software. Importantly, it reminds us of the temporal aspect of PowerPoint. Presentations obviously have duration, but unless pre-timed (in which case each slide has a specific amount of time), time is not really an organizing principle. There is no timeline in the interface and time is either managed by manual sequencing in relation to the slides (next or previous slide) or through the pre-timing function (rarely used if not for pecha-kucha-style presentations or kiosk presentations), which is very simple and does not support complex timing such as time-based layering of different kinds of content.
All the view modes accessible from the status bar are organized around slides (normal view, slide sorter, reading view and slide show). Comments and notes are attached to specific slides (there is no way of attaching such content to multiple slides or to no slide at all). The zooming and “fit slide to current window” functions (on the right) relate to how slides are viewed in the production mode of PowerPoint.
There are a number of operations (e.g. inserting content, formatting content etc.) that can be used on content within individual slides, but that do not relate to more than one slide. Exceptions are style templates/themes/backgrounds (that operate across the slide deck) and transitions. Style templates/themes/backgrounds simply determine what a slide looks like and make it easy to design a uniform slide-based visual experience. Transitions work across slides, but these do not threaten the integrity of the slides, but rather create a connective space between two slides (arguably in fact concealing or smoothing over the repetitious and sharp slide-per-slide “cut”). They were not part of the original setup of PowerPoint (nor surprising since the primary focus was on producing overhead transparencies), although transitions were mentioned in some of the early documents.
Production and Presentation
There is an interesting difference and tension between the production and presentation modes in PowerPoint. Essentially these modes or views are distinct in the software. This is not surprising given the history of the tool and the fact that it was mainly used to manufacture overhead transparencies and 35 mm slides in the early stages. In such processes (described in some detail in our article), the production of materials and the manifestation of them are connected (after all the production process renders the slide), but disconnected in the sense that the presentation technology is largely divorced from production technology. There could be no last-second change of a slide just before the presentation using the tool used to create the content (although one could of course use a marker to draw on a transparency etc.).
The presentation mode is the sleek, single-slide manifestation, where one normally does not see what is beneath the hood. As noted above, in the “Presenter Specification, May 22, 1986” document, it is said that the slide show mode displays slides one at a time and that “No window borders, scroll bars, menu bars, etc. are shown, only each individual slide”. Again, we are reminded of the temporal aspect here. Slides do not only organize the narrative in the sense of framing the argumentation synchronically, but also over time. The interface is made fully transparent, there are no internal windows or buttons on display and the content takes up all available screen space. In this way, the physical screen becomes the frame of the slide, and it can be argued that this overlap actually contributes to taking emphasis away from the framing and outlined slides (although the slide template may include an internal framing). The production environment, although organized around the slide, is much more multiplex. One example is the slide sorter, which fulfills an important function. It allows the user to easily organize a narrative based on a collection of slides (much like a light table, which is what the equivalent function is called in Apple’s Keynote).
Here a number of slides can be shown at the same time and while the integrity of the individual slide is not threatened (slides remain the building block), this view gives a very different perspective on the narrative of the deck. Because the presentation and production modes are so distinct it is very rare to see the slide sorter being used in presentations. However, it is more common to get a glimpse what is “behind the scenes” before someone starts a presentation.
This example is from a 2015 talk at CUNY is somewhat rare because it also shows the notes (in this case the speaker’s planned “thank you” script).
Some speakers are not so concerned about revealing the production view of the software while speaking. In a seminar in HUMlab, Umeå University from 2003, Sven Strömqvist shifts between several slide decks and positions within those slide decks.
This is more likely to happen if the speaker makes use of several pieces of software, which typically requires displaying and using the operating system and the desktop space. It is also an issue of speaker style. Normally, however, the presentation screen and the production view of the software are quite distinct.
The Speaker’s View
There is one aspect of the presentation mode — the speaker’s view (optional)— that shows more than one slide, but this view is exclusively for the benefit of the presenter.
This view functions much as a teleprompter. Normally the presentation view is only visible to the presenter (showing on her or his laptop). It contains three main elements: the current slide (what is being shown on the presentation screen), the subsequent slide (on the right) in the sequence and any notes (script) associated with the slide. There is also a timer, information about the number of the current slide and the total number of slides, and arrows (to move one slide forward or backward). The slide-by-slide structure is very clear here and so is the assumed forward movement. The presenter sees the next slide, but not the preceding one. This means that moving backward is not as well supported by this interface as moving forward (which is the expected direction). In the unlikely event that someone would like to run a presentation from the final slide to the first slide (backwards), the slide preview function in the presenter’s view would be fairly useless. However, it is possible to pull up miniatures of all slides in the deck and pick a specific slide. This view adds some flexibility, but it cannot be shown to others (it exists only in the presenter’s view).
An interesting and very rarely used function (at least in our experience) is the zooming function (part of the presenter’s view), which allows the presenter to select a rectangular part of a slide to be displayed on the presentation screen and on the screen running the presentation view. The size of the rectangle cannot be changed and there is only one zoom level.Significantly the zooming rectangle present in the presentation view is not shown on the presentation screen.
The selection process is thus not visible to the audience, although it could provide material and narrative support; not least giving the audience a sense of the process. Again, the presentation screen is made into a transparent surface, rather than a work space or a multi-layered presentation space. The zooming-in function is triggered through clicking, which results in the zoomed in picture showing on both screens.
There is no indication of the whole slide here (although the viewport can be moved around to show different parts of the slide). It seems clear that the zooming function is underdeveloped and underutilized, and more akin to the pen and laser pointer tool (also available in the presentation view) than to a tool that supports exploration and argumentative structure.
An important observation is that the distinct separation between production and presentation views in PowerPoint — a consequence of the strong slide emphasis in the software — discourages narrative structures beyond the slide-per-slide model. The zooming function just discussed could actually be a tool for engaging participants through working with focus, scale and zooming, but this is not really possible. The slide sorter is actually a very useful presentation tool (although it could be developed as such), but through clearly defining this as a production feature, it is very rarely used as a presentation view. Imagine a presenter with all her or his slides in one deck picking and choosing segments of slides within that large group. This is not impossible, but the software is rarely used this way and there are also limited search functionality. Essentially the data structure (as accessible to the users of the software) is flat and cannot be recruited to support something like a query-based light table used to create a story with the audience following the process.
Sticking to the Slide
One reason for PowerPoint’s dominance is that it is based on a simple concept (story telling through a series of slides, one slide at a time) that is not compromised in the material implementation. Including the slide sorter as a presentation tool or having a complex zooming function would break the unity of the concept behind PowerPoint. Not only would such functions compromise the one-slide-at-a-time model, but they would also impose a sense of time and process not supported by the software. The argumentative structures made available by the software tend not to be processual beyond the cut between slides (the sharpness of which may be alleviated by a transition). Unlike software based on a more cinematic model, PowerPoint has no timeline apart from the simple automatic timing that can be imposed on a slideshow (each slide is given a certain time in a prerecorded sequence).
Furthermore, changing or amending the concept may also require changing the infrastructure and, conversely, if the infrastructure changes, the concept would have to be instantiated differently, which might change the experience and narrative possibilities. Such change and variation make standardization challenging, and PowerPoint works so well because both the concept and the infrastructure are stable and are held stable. Essentially we are concerned with single slides projected in a sequence on a single screen. As pointed out by the inventors (Gaskins 2012:93), the boundedness of the slide created a strong connection between what was on the screen (inside the slide) with the display. There was a real-world size to the slide and content created within the software. This does not mean that the production screen is the same size as the presentation screen, but rather that the slide will be intact and look good regardless of the exact configuration of the presentation screen. Frequent users of PowerPoint will know that the software manages different resolutions well, which has the consequence that users do not really have to be concerned about the details of the presentation situation.
A Model under Pressure?
There are two developments that exemplify a tension in relation to the 1:1 mapping when the technology and use patterns change (in this case within the single-screen paradigm). The first example is the large-scale adoption of wide-screen displays — in particular when both 4.3 screens and wide-screen displays were both fairly frequent. It is not the case that a 4:3 presentation cannot be run a wide-screen projection screen, but it is clear that the mapping is not 1–to-1, and careful presenters have had to think about the configuration of their production machine and the intended presentation environment. The second example concerns the increase in resolution in mobile, laptop and desktop screens. A small-screen laptop may well have a resolution several times the resolution of a standard presentation screen. As already noted, PowerPoint can manage the shift in resolution, but if users start to produce content that makes better use of high-resolution content, the restrictions of the slide model become more pronounced (where is the zooming function that I can use as a presentational and narrative tool?) Furthermore, anyone who has tried to make content in PowerPoint for very high-resolution presentation screens will have noticed that the software does not manage very high resolutions well — it gets very sluggish (it was not built for super-high resolutions).
The model is put under even more pressure if we consider more far-reaching infrastructural changes or alternative narrative needs not accommodated by the software. Think of all the mobile screens around us today for instance and how these could be used as conceptual and material lenses or interaction tools in relation to an ongoing “presentation”. An example from our own practice when PowerPoint-style software is clearly not sufficient is the multiple-screen infrastructure in HUMlab, Umeå University. Software was built to handle this infrastructure (still partly slide-based, but for multiple-screen infrastructure) and as part of ongoing experimentation with knowledge production, there was a conference in December 2014 (Genres of Scholarly Knowledge Production), when slideware such as PowerPoint was essentially disallowed (see Svensson 2015 for an account of the event).
There is no risk that PowerPoint will go out of business, but there is certainly pressure from new information ecologies (multiple screens including mobile displays, large online repositories) and conceptual-argumentative needs (the need to navigate in large data sets for instance, where zooming might be a very useful navigational feature).
Austin, Dennis, Rudkin, Tom and Robert Gaskins. 1986. “Presenter Specification May 22, 1986”, Forethought, Inc. http://www.robertgaskins.com/pages/presenter-%28powerpoint%29-spec-austin-rudkin-gaskins-may-1986.pdf.
Gaskins, Robert. 2012. Sweating Bullets: Notes about Inventing PowerPoint. Vinland Books.
Svensson, Patrik. 2015. “’Genres of Scholarly Knowledge Production’ Revisited Post-event curatorial statement for the GOSKP conference at HUMlab, Umeå University, December 10–12, 2014". https://medium.com/genres-of-scholarly-knowledge-production/revisiting-genres-of-scholarly-knowledge-production-fd240e9631ec.