Slideshow: Functional Presentations
Contents
1. define two fish sizes rbl superimpose vl append 5 in aq two fish 718 101bs 7 Ibs 10 Ibs 2 4 Picts and Identity Pict objects are purely functional A particular pict such as Little fish can be used in multiple independent contexts Functions that adjust a pict such as scale do not change the given pict but instead generate a new pict based on the given one For example we can combine two fish and a scaled version of two fish ina single pict hce append 10 two fish scale two fish 0 5 D D oe Picts nevertheless have an identity in the sense of Scheme s eq and each use of a Slideshow function generates a pict object that is distinct from all other pict objects 2 5 Finding Picts To add an arrow from 7 lbs to the little fish we could insert an arrow pict into the vl append sequence with negative space separating the stacked pict but then adding an arrow from 10 lbs to the big fish would be more difficult Slideshow provides a more general way to extend a pict which is based on finding the relative location of sub picts To locate a sub pict within an aggregate pict Slideshow pro vides a family of operations beginning with find These operations rely on the identity of picts to find one pict within another The suffix of a find operation indicates which corner or edge of the sub pict to find it is a combination of 1 c or r left center or right with t t1 c bl or b top top baseline2. 16 R B Findler and M Flatt lc superimpose large fish hc append dinner fish spacer lc superimpose medium fish medium aaah fish he append lunch fish spacer lc superimpose small aaah fish small fish Now the large fish and the medium fish overlap with the fish to the right but the storyboard shows that the overlapping fish will never appear together in a single frame Meanwhile this pict preserves the original spacing of the dinner lunch and small fish Once we have the basic layout designed the design recipe tells us to parameterize it so we can hide certain fish In this case we parameterize the scene with a list of the fish that we want to show and we define an on helper function to either show or hide each character in the scene 7 Main characters listof pict main char define main characters active lc superimpose on active large fish hc append gt pict frame on active dinner fish spacer lc superimpose on active medium fish on active medium aaah fish hc append on active lunch fish spacer lc superimpose on active small aaah fish on active small fish rr On listof pict main char define pict main char on set pict if gt pict memq pict set pict ghost pict To assemble the main characters for each frame we simply apply main characters to a list of fish to show define mainl main characters list
3. That is ghost makes big fish invisible to the eye but not to the find operations Thus big fish exists twice ineach all fish pict and add big label finds the wrong one in all fish 1 To hide a pict s identity Slideshow provides launder as a complement to ghost The launder function takes a pict and produces a pict with the same dimensions and drawing but without any findable sub picts To ensure that add big label finds the right big fish we can both ghost and launder the pict to hide define all fish 1 in aq cc superimpos both fish launder ghost bigger fish add big label all fish 1 101bs 10 Ibs define all fish 2 in aq cc superimpos launder ghost both fish bigger fish add big label all fish 2 171bs 17 lbs Alternately we might define bigger fishas launder scale big fish 1 7 so that bigger fish would never be confused with big fish In that case we must also adjust add big labe1 to accept a target fish either big fishorbigger fish 2 7 From Pictures to Slides Pict construction primitives are only half of Slideshow s library The other half defines pict operations that support common slide tasks and that cooperate with a slide display system Common tasks include creating a slide with a title creating text with a default font breaking lines of text bulletizing lists and staging bullet lines Cooperating with the display includes correlat
4. gt pict define connect main from to add arrow line 6 main from find ct to find cb Slideshow provides several libraries that are built in terms of finding picts For example the balloon ss library provides wrap balloon for wrapping a pict into a cartoon balloon plus place balloon for placing the balloon onto a pict define yikes wrap balloon text Yikes bold roman 9 s 0 10 spike direction and displacement 7 Panic pret gt piet define panic p place balloon yikes p little fish find ct o panic two fish D The place over operation preserves the bounding box of its first argument instead of extending it to include the placed pict This behavior is useful for adding arrows and balloons such that further compositions are unaffected by the addition For example we can still put the fish into an aquarium after adding the panic balloon in aq panic two fish Slideshow Functional Presentations 7 2 6 Ghosts and Laundry We eventually want a slide sequence with variants of the aquarium pict One variant should have the little and big fish together in the aquarium as before define both fish hc append 10 little fish big fish in aq both fish Another variant should have just the big fish now even bigger since it has eaten the little fish If we generate the pict as define bigger fish scale big fish 1 7 in aq bigger fish then our slides will not look right becaus
5. pudgy small fish 2 define medium aaah fish launder pudgy small aaah fish 2 define large fish launder pudgy small fish 3 fish scene list dinner fish lunch fish small aaah fish 10 Ibs 10 Ibs 10 Ibs fish scene list dinner fish medium aaah fish 10 Ibs 20 Ibs Slideshow Functional Presentations 21 fish scene list large fish 30 Ibs Finally imagine generalizing from three fish to an arbitrary number of fish Although this change is more complex the design recipe suggests a way to break down the changes and points to a clear set of modifications to the code The first step is to adapt our definitions of the main characters We can organize the them into three lists the fish that are eaten and the fish that do the eating with their mouths both open and closed define fc 9 define meal fish build list fc 1 A i launder lunch fish apply hc append meal fish define eating fish build list fc A i launder pudgy small fish i 1 apply hc append eating fish define eating aaah fish build list fc A i launder pudgy small aaah fish i 1 apply hc append eating aaah fish 22 R B Findler and M Flatt This modification also affects the second step of the design recipe the placement of the main characters We must re define main characters to generalize from just three fish to an arbitrary number of fish With the e
6. 13dec2004 Language Slideshow Fig 3 Mousing over expressions to see resulting picts a pict by drawing a black box at the start of the expression A programmer can move the mouse over a box to see the pict s generated by the expression In the screen dump the mouse is over the start of an hc append call The right hand side of the window shows eight picts one for each time that the expression was evaluated during the program run Specifically the first pict shows the result of the first evaluation where rest of scene contains only one eating fish including its aaah variant su perimposed so the result of hc append contains two fish As the program continued this result was further extended with an eating fish and then it became rest of scene for the evaluation of the hc append expression and so on Overall the expression under the mouse is evaluated eight times when the program is run so eight results are shown for the expression 6 3 WYSIWYG Picts The left hand side of figure 4 shows a program that uses WYSIWYG layout for a composite pict The box with a projector screen icon in its top right corner is a pict box Inside the pict box each box with a top right the comma suggestive of unquote is a Scheme box A programmer can create any number of Scheme boxes within a pict box and then drag the Scheme boxes to position them within the pict box A pict box is an expression in graphical syntax and its res
7. dinner fish lunch fish small fish D D D 3 3 Supporting Cast For the supporting cast we start with the fish s weights Typically a supporting cast mem ber s position depends on one or more main characters As it happens we can go even further for the weights by generating the label pict based on the area of a fish Slideshow Functional Presentations 17 7 weight pict main char gt num define weight fish inexact gt exact round pict width fish pict height fish 1 60 define small lbs lbs weight small fish 10 Ibs To place the weight of the fish directly below the fish we first compute the position of the fish Then we can use place over to combine a frame with the weight The second argument to place over is the horizontal position of the left edge of the weight The following code ensures that the weight is centered below the fish The third argument to place over is the vertical position and the following code places the weight just below the aquarium define values fx fy find cc mainl small fish define mainl weight place over mainl fx pict width small lbs 2 DD amp pict height mainl 6 small lbs 10 Ibs Combining the weight and the original scene in this manner helps us test our computation of the weight s coordinates but the design recipe tells us to build an overlay pict that only contains the weight Building the overlay as a separate pict enable
8. environment that we ad vocate An integrated viewer meanwhile can easily support executable code that is em bedded within slides Embedded code is particularly useful in a presentation about pro gramming or about a software system since an eval hyperlink can be inserted into any slide More generally Slideshow gives the presentation creator access to the complete PLT Scheme GUI toolbox so anything is possible 8 Conclusion In only the last few years laptop projected slides have become the standard vehicle for delivering talks and tools other than PowerPoint are still catching up We offer Slideshow as a remedy to PowerPoint s lack of abstraction HTML s lack of flexibility and BIEX s lack of maintainability We also offer a design recipe for slide sequences that is supported by Slideshow s set of primitives Programmatic construction of pictures and slides is probably not for everyone even with powerful programming environment tools For various reasons many people will prefer to create pictures and slides in PowerPoint and without significant abstraction no matter how nice the language of picture construction For the authors tastes and purposes however programmatic construction works well and we believe that it appeals to many programmers In our slides with proper code ab straction we can quickly experiment with different configurations of a picture add slide by slide animation and evolve ever more general
9. functional programming Felleisen Matthias Findler Robert Bruce Flatt Matthew amp Krishnamurthi Shriram 2001 How to design programs Cambridge Massachusetts The MIT Press http www htdp org Findler Robert Bruce Flanagan Cormac Flatt Matthew Krishnamurthi Shriram amp Felleisen Matthias 1997 Sept DrScheme A pedagogic programming environment for Scheme Pages 369 388 of Proc international symposium on programming languages Implementations logics and programs Finne Sigbjorn amp Peyton Jones Simon 1995 July Pictures A simple structured graphics model Proc Glasgow functional programming workshop Henderson Peter 1982 Functional geometry Pages 179 187 of Proc ACM conference on Lisp and functional programming Hobby John D 1992 A user s manual for MetaPost Computer science technical report AT amp T Bell Laboratories CSTR 162 Kamin Samual N amp Hyatt David 1997 Oct A special purpose language for picture drawing Pages 297 310 of Proc USENIX conference on domain specific languages Kernighan Brian W 1991 PIC a graphics language for typesetting user manual Computer science technical report AT amp T Bell Laboratories CSTR 116 PLT PLT Scheme www plt scheme org Seranno Manuel amp Gallesio Erick Skribe home page http www inria fr mimosa fp Skribe Seranno Manuel amp Gallesio Erick 2002 Oct This is Scribe Pages 31 40 of Proc workshop
10. of text on a slide For example when posing a question to students an answer may be revealed only after the students have a chance to think about the question To support this simple kind of staging the slide function and each of its variants treats the symbol next specially when it appears in the argument sequence All of the picts before next are first used to generate a slide and then the picts before next plus the arguments after next are used to generate more slides slide title Food Chain scale all fish 1 6 t How many fish in the aquarium next bt Two 12 R B Findler and M Flatt Food Chain Food Chain How many fish in the aquarium How many fish in the aquarium Two The next symbol simplifies linear staging of slide content The slide function also supports tree like staging of content through the alts symbol The argument following alts must be a list of lists instead of a single pict Each of the lists is appended indi vidually onto the preceding list of pict to generate a set of slides The final list is further appended with the remaining arguments after the list of lists The next and alts symbols can be mixed freely to generate sub steps and sub trees slide title Food Chain scale all fish 1 6 alts list list t What if the big fish eats the little one next t It will become 17 pounds list t What if the littl
11. part of the name cc superimpose indicates that the picts are centered hor izontally and vertically as they are stacked on top of each other Slideshow provides a superimpose function for each combination of 1 c or r left center or right with t c or b top center or bottom One additional mode of alignment is useful for text When combining text of different fonts into a single line of text then neither ht append nor hb append produces the right result in general because different fonts have different shapes Slideshow provides hb1 append for stacking picts so that their baselines match 7 lbs num gt pict define lbs amt hbl append text number gt string amt bold swiss 9 text lbs swiss 8 define 10lbs lbs 10 10 Ibs Slideshow Functional Presentations 5 define 7lbs lbs 7 7 Ibs define 171lbs lbs 17 17 Ibs For multi line text hol append matches baselines for the bottommost lines Slideshow provides ht 1 append to make baselines match for topmost lines Naturally Slideshow provides variants of superimpose with t 1 and b1 as well Finally Slideshow provides vl append vc append and vr append to stack picts vertically with left center and right alignment To add the labels to our aquarium pict we can use vl append to first stack the aquarium on the 10 lbs label and then use rb1 superimpose to add the 7 lbs label to the bottom right corner of the pict
12. support our design recipe in Slithy but more work is required to adapt the pict abstraction to fully support movie animation The Slithy authors have identified principles for slide animation Zongker amp Salesin 2003 that should guide the design of new animation abstractions Slideshow s pict abstraction is by no means the first language designed for generating pictures and Slideshow s picture constructors are similar to those of Henderson s func 34 R B Findler and M Flatt tional pictures Henderson 1982 MLgraph Chailloux et al 1997 Pictures Finne amp Peyton Jones 1995 Functional PostScript Shivers amp Sae Tan 2004 FPIC Kamin amp Hyatt 1997 pic Kernighan 1991 MetaPost Hobby 1992 and many other systems Unlike pic and MetaPost but like MLgraph etc Slideshow builds on a general purpose programming language so it can support modular development it allows programmers to write maintainable code libraries and tests and it is supported by a programming environ ment MLgraph Pictures Functional PostScript and FPIC provide a richer set of transfor mation operations mainly because they all build on PostScript while Slideshow provides a richer set of text formatting operations More fundamentally Slideshow s combination of find ghost and launder appears to be unique These operations reflect gener ally the way that Slideshow is tailored for slide creation and specifically how Slideshow s
13. through portable family symbols such as roman a serif font swiss a sans serif font usually Helvetica modern a fixed width font and symbol a font with Greek characters and other symbols PLT Scheme users control the family to font mapping so a Slideshow programmer can assume that the user has selected reasonable fonts Alternately a programmer can always name a specific font though at the risk of making the presentation unportable Since specific fonts vary across platforms displays and users the specific layout of picts in a Slideshow presentation can vary due to different bounding boxes for text picts Nevertheless as long as a programmer uses pict width and pict height instead of hard wiring text sizes slides display correctly despite font variations This portability depends on computing pict layouts at display time instead of computing layouts in advance and distributing pre constructed picts Text scaling leads to additional challenges For many displays a font effectively exists only at certain sizes if a pict is scaled such that its actual font size would fall between existing sizes the underlying display engine must substitute a slightly larger or smaller font Consequently a simple scaling of the bounding box in the 1024x768 space does not accurately reflect the size of the text as it is drawn leading to overlapping text or unattractive gaps To compensate for text scaling problems Slideshow
14. Under consideration for publication in J Functional Programming 1 Slideshow Functional Presentations ROBERT BRUCE FINDLER University of Chicago e mail robby cs uchicago edu MATTHEW FLATT University of Utah e mail mElatt cs utah edu Abstract Among slide presentation systems the dominant application offers essentially no abstraction capa bility Slideshow an extension of PLT Scheme represents our effort over the last several years to build an abstraction friendly slide system We show how functional programming is well suited to the task of slide creation we report on the programming abstractions that we have developed for slides and we describe our solutions to practical problems in rendering slides We also describe ex perimental extensions to DrScheme that support a mixture of programmatic and WYSIWYG slide creation 1 Abstraction Friendly Applications Strand a computer scientist at an airport and the poor soul would probably survive for days with only a network connected computer and five applications an e mail client a web browser a general purpose text editor a typesetting system and a slide presentation application More specifically while most any mail client or browser would satisfy the stranded scientist probably only Emacs or vi would do for editing IAT EX for typesetting and Microsoft PowerPoint for preparing slides The typical business traveler would more likely insist on Microsoft Word f
15. alk Slideshow provides a comment constructor that takes a commentary string and produces an object that can be sup plied to slide When the slide function finds a comment object it accumulates the comment into the slide s overall commentary instead of generating an image The Slideshow viewer displays a slide s commentary on demand in a separate win dow e Ifaspeaker s machine supports multiple displays the speaker might like to see com ments and a preview on a private display Thus in addition to the separate commen tary window Slideshow provides a preview window that shows the current slide and the next slide in the presentation e For presentations that involve demos the speaker might like hyperlinks on certain slides to start the demos Slideshow provides a clickback operator that takes a pict and a procedure of no arguments the result is a pict that displays like the given one but that also responds to mouse clicks by calling the procedure In this case we exploit the fact that slide generation and slide presentation execute on the same virtual machine e Although many animations can be implemented as multiple slides that the speaker advances manually other animations should be entirely automatic Currently Slideshow provides only minimal support for such animations though an imperative scroll transition function that registers a scroll animation over the previously registered slide This feature has been used mainl
16. bel upper base hb append 7 add baseline pict num pict num pict pict gt pict gt pict 7 Add one baseline dy from top of p combine given 7 label with arrow using xx append don t change 7 the bounding box define add baseline p dy lbl xx append place over P pict width p bbox sep pict height p dy refocus xx append 2 little arrow lbl little arrow 77 Labele Str aee apet Build a label define label strs apply vl append 2 map A str text str swiss 8 strs 77 practically num num gt bool define practically a b let epsilon 0 01 and lt a epsilon b gt a epsilon b 77 Constants define bbox sep 5 define bbox color gray define little arrow inset arrow line 2 bbox sep 0 bbox sep 2 bbox sep 0 0 0 38 R B Findler and M Flatt References Chailloux Emmanuel Cousineau Guy amp Su rez Ascander 1997 The MLgraph system Clements John Felleisen Matthias Findler Robert Bruce Flatt Matthew amp Krishnamurthi Shri ram 2004 Fostering little languages Dr Dobb s journal Mar 16 24 Clinger William D 1998 Proper tail recursion and space efficiency Pages 174 185 of Proc ACM conference on programming language design and implementation Elliott Conal amp Hudak Paul 1997 Functional reactive animation Pages 263 273 of Proc ACM international conference on
17. bels add bbox p 7 annotate all pict gt pict define annotate all p add baselines annotate p 7 apply colorized pict gt pict pict gt pict 7 colorizes the annotations of annote define apply colorized annote p cc superimpose P colorize annote ghost p bbox color 7 add bbox labels pict gt pict Add height and width bars and labels but 7 don t change the boundbing box defin add bbox labels p refocus vr append bbox sep hce append bbox sep hc append label height vc append hline 5 0 vline 0 pict height p hline 5 0 p vc append hc append vline 0 5 hline pict width p 0 vline 0 5 label width p 7 add baseline lines pict gt pict Add baseline lines within the bbox defin add baseline lines p let hl hline pict width p 0 place over place over p 0 pict height p pict descent p hl Slideshow Functional Presentations 37 0 pict ascent p hl 7 add baselines pict gt pict 7 Add baseline labels and arrows defin add baselines p let p add baseline lines p if practically pict descent p pict height p pict ascent p 7 Show baselines together add baseline p pict descent p label bases hc append 7 Show baselines separately add baselin add baseline p pict descent p label lower base ht append pict height p pict ascent p la
18. bination with language based abstraction Section 2 is a tour of Slideshow s most useful constructs Section 3 presents the Slideshow design recipe for picture sequences Section 4 explains Slideshow s core implementation Section 5 briefly addresses practical issues for rendering slides on different media and operating systems Section 6 describes our prototype extension of DrScheme 2 A Tour of Slideshow A pict is the basic building block for pictures in Slideshow Roughly a pict consists of a bounding box and a procedure for drawing relative to the box Ultimately a slide is represented as a single pict to be drawn on the screen This section demonstrates how to use Slideshow primitives to generate slides like the following which might appear in a presentation about how fish gain weight when they eat other fish Food Chain 7 lbs 10 Ibs e Current total weight 17 Ibs After big fish eats little one 17 Ibs Slideshow Functional Presentations 3 2 1 Pict Generators Slideshow provides several functions for creating picts of simple shapes For example the filled rectangle function takes a height and a width and produces a pict filled rectangle 20 10 a Similarly the text function takes a string a font class and a font size and it produces a pict for the text text 10 lbs bold swiss 9 10 Ibs The standard fish function takes a height width direction symbol color string eye color st
19. center bottom baseline or bottom The results of a find operation are the coordinates of the found corner edge relative to the aggregate pict A find operation is often combined with place over which takes a pict hori zontal and vertical offsets and a pict to place on top of the first pict For example we can create a connecting arrow with arrow line which takes horizontal and vertical displacements plus the size of the arrowhead and place it onto two fish sizes 6 R B Findler and M Flatt define values ax ay find ct two fish sizes 71bs define values bx by find cbh two fish sizes little fish place over two fish sizes ax pict height two fish sizes ay arrow line bx ax by ay 6 7 Ibs 10 Ibs Since we need to multiple arrows we abstract this pattern into a function 77 connect pict pict pict gt pict define connect main from to define values ax ay find ct main from define values bx by find cb main to place over main ax pict height main ay arrow line bx ax by ay 6 define labeled two fish connect connect two fish sizes 7lbs little fish 10lbs big aaah fish 7 Ibs 10 Ibs Slideshow provides a function that is like connect called add arrow line In addi tion to the arguments of connect add arrow line accepts the find functions for each sub pict Thus connect can be implemented more simply as 77 connect pict pict pict
20. common Slideshow defines a single slide to be a pict that is 1024x768 units The default border leaves a 984x728 region for slide content These units do not necessarily correspond to pixels however Depending on the display at presen tation time the pict is scaled e g by a factor of 3 for an 800x600 display If the display aspect is not 4 3 then scaling is limited by either the display s width or height to preserve the pict s original 4 3 aspect ratio and unused screen space is painted black Typically unused space appears when viewing slides directly on a desktop machine such as during slide development Slideshow does not use a special file format for slide presentations Instead a Slideshow presentation is a program and pict layouts are computed every time the presentation is started This computation is rarely so expensive that it inteferes with interactive develop ment Consequently the target screen resolution is known at the time when slides are built This information can be used for example to scale bitmap images to match the display s pixels instead of units in the virtual 1024x768 space Information about the display is also useful for font selection 5 2 Fonts Fonts are not consistently available across operating systems or even consistently named To avoid platform dependencies Slideshow presentations typically rely on PLT Scheme s Slideshow Functional Presentations 29 mapping of platform specific fonts
21. crements the slide number As a result a condense slide s number is actually a range indicating the range of normal slides covered by the condensed slide Programmers can use the condense and printing predicates to further cus tomize slide rendering for condense mode and printing A skip slides function al lows the programmer to increment the slide count directly 30 R B Findler and M Flatt 5 4 Pre rendering Slides To avoid screen flicker when advancing slides in an interactive presentation Slideshow renders each slide in an offscreen bitmap and then copies the bitmap to the screen The time required to render a slide is rarely noticeable but since a programmer can cre ate arbitrary complex picts or write arbitrarily complex code that uses the drawing context directly the rendering delay for some slides can be noticeable To ensure instantaneous response in the common case Slideshow pre renders the next slide in the presentation se quence while the speaker dwells on the current slide If the speaker requests a slide change within half a second the slide is not pre rendered because the speaker may be stepping backward through slides 5 5 Display Interaction In addition to creating pictures for the screen slide presenters must sometimes interact more directly with the display system e A slide author might wish to attach a commentary to slides for the benefit of the speaker or for those viewing the slides after the t
22. ct width lunch fish gt 30 pict height lunch fish gt 20 In addition to a bounding box every pict has a top and bottom baseline that can be used for aligning text When a pict contains a single line of text the baselines coincide When a pict contains no text typically the baselines match the bottom edge of the bounding box 4 See the appendix for definitions of annotate and annotate all Slideshow Functional Presentations 25 10 Ibs annotate all 101bs 7 Ibs 10 Ibs 17 Ibs annotate all vc append 3 7lbs 10lbs 171bs annotate all lunch fish 10 Ibs annotate all vc append 3 lunch fish 101bs The pict ascent primitive accepts a pict and returns the distance from the pict s top edge to the top baseline and the pict descent primitive produces the distance from the pict s bottom edge to the bottom baseline pict height 10lbs 11 pict ascent 10lbs gt 9 pict descent 10lbs gt 2 pict ascent lunch fish gt 30 pict descent lunch fish gt 0 A pict s bounding box does not necessarily enclose all of the pict s image but usually it does In any case the pict s image is drawn relative to its bounding box so the bounding box and the baselines provide points relative to the image for alignment with other picts The only primitive pict combining function is place over l place over lunch fish 25 2 101lbs Tr The place over function does not combine the boundin
23. determines the expected scaling of slides based on the current display size before generating picts It then uses the expected scale to generate a bounding box for text picts that will be accurate after scaling Occa sionally the predicted scale is incorrect because the programmer uses the scale operation in addition to the implicit scale for the target display but this problem is rare When neces sary the programmer can correct the scale prediction by using the scale improve new text form and creating text picts within the dynamic extent of this form 5 3 Printing Slides A drawing context in PLT Scheme is either a bitmap display possibly offscreen or a PostScript stream Thus printing a Slideshow presentation is as simple as rendering the slides to a PostScript drawing context instead of a bitmap context Slideshow provides a condense mode for collapsing staged slides Collapse mode au tomatically ignores next annotations a programmer can use next instead of next to force separate slides in condense mode In contrast alts annotations cannot be ig nored because each alternative can show different information A Slideshow programmer can indicate that intermediate alternatives should be skipped in condense mode by using alts instead of alts Slideshow synchronizes page numbering in condensed slides with slide numbering in a normal presentation In other words when slide skips a next annotation it also in
24. e fish eats the big one next t It will become 17 pounds next t But the little fish can t eat the big one Food Chain Food Chain Food Chain What if the big fish eats the little one What if the big fish eats the little one What if the little fish eats the big one It will become 17 pounds Food Chain Food Chain What if the little fish eats the big one What if the little fish eats the big one It will become 17 pounds It will become 17 pounds But the little fish can t eat the big one With slide page item next alts etc a programmer can build a text oriented presentation in Slideshow almost as easily as in PowerPoint The purpose of Slideshow however is not to encourage bullet point presentations but to simplify the creation of more interesting graphical presentations which is why we began with standard fish in stead of page item Having concluded a tour of Slideshow features we now turn our attention to the design of slide presentations Slideshow Functional Presentations 13 3 How to Design Slide Presentations The strength of functional programming is that it supports and encourages good program design Felleisen et al 2001 We believe that Slideshow supports and encourages good presentation design not only because it is based on functional programming but because it directly supports a design recipe for image sequences This section demon
25. e the aquarium changes shape from the first pict to the second We can avoid this problem by constructing a large enough aquarium for the first pict Conceptually we d like to stack the large fish pict on top of the pict with the two fish together and then put the combined pict in the aquarium so that it is large enough to fit both picts in both dimensions and then hide the single fish The ghost function takes a picture and generates a picture with the same bounding box as the given one but with no drawing Thus we can create the right initial pict as follows define all fish 1 in aq cc superimpos both fish ghost bigger fish We can create the last slide by ghosting both fish instead of bigger fish define all fish 2 in aq cc superimpos ghost both fish bigger fish Since both all fish 1 and all fish 2 contain all three fish they are guaranteed to be the same size If we try to add a label and arrow for the big fish however something goes wrong 7 add big label pict pict gt pict define add big label all fish wt let labeled vr append 5 all fish wt connect labeled wt big fish o fom a add big label all fish 1 101bs 8 R B Findler and M Flatt add big label all fish 2 171bs 17 lbs The problem is that bigger fish is a scaled version of big fish and even though bigger fish is ghosted in all fish 1 it can still be found as a sub pict
26. ent w built in slide title center Fish Story ae page item One Fish One Fish scale big fish 6 D page item Two Fish ee scale hc append 10 a little fish lt gt big fish 6 Note that given a bullet pict item is easily implemented in terms of para Slideshow Functional Presentations 11 7 item num pict gt pict define item w picts htl append gap size 2 bullet apply para w pict width bullet gap size 2 picts Just as Slideshow provides many slide variants it also provides many para and item variants including variants for right justified or centered paragraphs and bulleted sub lists The para function for example typesets a paragraph like para but it allows the result to be more narrow than the given width in case no individual line fills exactly the given width 2 9 Staging Slides One way to stage a sequence of slides is to put one pict for each slide in a list and then map a slide generating function over the list of picts 77 agq slide pict gt void defin aq slide all fish slide title center Food Chain scale all fish 6 map aq slide list all fish 1 all fish 2 Food Chain Food Chain Interesting presentations usually build on this idea and we discuss it more in section 3 Much like text and hbl append for typesetting paragraphs however this strategy is awkward for merely staging bullets or lines
27. f interactive When the programmer moves the mouse pointer over a pict DrScheme shows sub pict bounding boxes that enclose the pointer In the figure the narrow bounding box corresponds to the blue fish under the pointer and the wide bounding box corresponds to an intermediate pict from hc append By default DrScheme does not show bounding boxes that contain other bounding boxes around the pointer so the visible bounding boxes are minimal but pressing the Alt or Meta key exposes all of the bounding boxes that enclose the pointer Pressing the Control key shows all bounding boxes in the entire pict Pressing the Shift key toggles the color of each bounding box between black and white 6 2 Tracing Picts Figure 3 shows a screen dump for another tool This tool records all picts that are generated during the run of a Slideshow program It then highlights each expression that produced 32 R B Findler and M Flatt 009e fish scm DrScheme oO fish scm v Gafine gt tal Save eg Step Q Check Syntax Run Stop define main characters active foldl lambda eal fish eating fish eating aaah fish rest of scene impose fon ac ting fish flon active Jating aaah fish Ree flon active Jeal fish 0 cc superimpo eating fish r eating aaah fish meal fish cdr eating fish cdr eating aaah fish main characters append meal fish eating fish eating aaah fish Welcome to DrScheme version 299 24 cvs
28. f providing short names for particularly useful abstractions Slideshow pro vides the function t for creating a text pict with a standard font and size which defaults to sans serif 32 units high Thus the label for the earlier example could have been imple mented as t Mmmm fish instead of text Mmmm fish swiss 32 The bt function is similar to t except that it makes the text bold and it makes text italic For typesetting an entire sentence which might be too long to fit on a single line and might require multiple fonts Slideshow provides a para function The para function takes a width and a sequence of strings and picts and it arranges the text and picts as a paragraph that is bounded by the given width In the process para may break strings on word boundaries The width argument to para is often based on the client w constant which is defined as the width of the entire slide minus a small margin Just One Fish slide title center Just One Fish lt p scale big fish 10 Give a man a fish and he ll put it para 1 2 client w pee Give a man a fish and he ll put it on a slide The page para function is like para but with client w as the implicit first width Slideshow accomodates bulleted lists with the item function It is similar to para except that it adds a bullet to the left of the paragraph In parallel to page para and para the page item function is like item but with cli
29. fc 1 Slideshow Functional Presentations 23 ith 0 10 Ibs 10 Ibs 10 Ibs 10 Ibs 10 Ibs 10 Ibs 10 Ibs 10 Ibs 10 Ibs ith 1 10 Ibs 10 Ibs 10 Ibs 10 Ibs 10 Ibs 10 Ibs 10 Ibs 10 Ibs 10 Ibs ith 6 10 Ibs 10 Ibs 10 Ibs 10 Ibs 10 Ibs 40 Ibs ith 7 10 Ibs 10 Ibs 10 Ibs 10 Ibs 10 Ibs 40 Ibs 24 R B Findler and M Flatt ith 16 90 Ibs Overall generalizing to an arbitrary number of fish required us to abstract over the fish construction and to write some list processing functions In fact we can use this technique to demonstrate the behavior of an algorithm in a slide show By first designing a scene that represents some aspect of the behavior of an algorithm animating the algorithm is merely a matter of adding a calls in the body of the algorithm to generate picts that capture snapshots of the algorithm s state 4 Anatomy of a Pict Internally a pict consists of three parts e A bounding box and baselines e A drawing procedure which produces an image relative to the bounding box e The identity and location of sub picts A few core pict operations manipulate these parts directly and all other operations can be derived from the core operations 4 1 Bounding Boxes A pict s height and width define a bounding box for the pict annotate lunch fish The primitive pict width and pict height functions accept a pict and return the width and height of the pict s bounding box respectively pi
30. g boxes and baselines of the two picts Instead it preserves the bounding box and baselines of the first pict T lbs annotate all place over lunch fish 25 2 101bs One way to enlarge the bounding box is to place a pict on top of a blank pict The blank function creates a pict with an empty image and with a given width height ascent and de scent We can package the necessary computations into a general dxdy superimpose function 26 R B Findler and M Flatt 7 Axdy superimpose pict num num pict gt pict define dxdy superimpose p dx dy q let w max pict width p pict width q dx h max pict height p pict height q dy a min pict ascent p pict ascent q dy d min place over h pict height p pict descent p h pict height q dy pict descent q blank w h a d 0 0 place over p dx dy q AT bs annotate all lt dxdy superimpose lunch fish 25 2 101bs Slideshow s append superimpose and inset functions can all be implemented similarly in terms of place over blank pict width pict height pict ascent and pict descent 4 2 Drawing Procedure Most picts can be implemented in terms of simple shapes with stacking and placing op erations Simple shapes in turn are implemented with Slideshow s dc primitive where dc stands for drawing context which provides direct access to the underlying drawi
31. icular character may change shape over multiple frames so design one expression per view of each character In this case we have three characters which are the three fish in the first frame The first two fish have only a single view each The last fish has five small with its mouth open and closed medium sized with its mouth open and closed and large with its mouth closed 2 Design a pict expression that includes all of the views of the main characters from all frames in the scene This pict may include a character in multiple places if the character appears in different places in different frames To generate a particular 14 R B Findler and M Flatt frame we parameterize the pict expression to hide characters that do not appear in the frame In this case we must generate a pict that has the three fish in a line plus different views of the last fish on top of those fish For example the largest fish will be on top of the first fish with their left edges touching 3 Design pict expressions for the supporting cast Unlike the main cast each support ing cast pict should have the same dimensions as the entire frame so it can be su perimposed on its corresponding frame Each supporting cast member is placed into an appropriately sized blank pict at a position that is determined by the its location in the complete frame In this case we design overlays for the arrows and for the weights Each weight will be centered below its corre
32. ideshow s current trade off flexible place over versus inflexible fonts has worked well in practice If a pict needs to be parameterized by its font we simply use functional abstraction or parameterize Countless packages exist for describing slides with an HTML like notation Such pack ages typically concentrate on arranging text and pictures are imported from other sources Countless additional packages exists for creating slides with ATEX including foiltex and Prosper Van Zandt n d With these packages IATEX remains well adapted for presenting math formulae and blocks of text but not for constructing pictures and not for implement ing and maintaining abstractions Like Slideshow Skribe Seranno amp Gallesio n d Seranno amp Gallesio 2002 builds on Scheme to support programmatic document creation in a general purpose functional lan guage Skribe s architecture targets mainly the creation of articles books and web pages Since Skribe includes a I4TRX output engine appropriate bindings could added to Skribe to create slides through I4TRxX based packages Unlike Slideshow most slide presentation systems including all I4TRX based PostScript based and PDF based systems treat the slide viewer as an external tool Separating the viewer from the slide generation language makes display specific customization more dif Slideshow Functional Presentations 35 ficult and it inhibits the sort of integration with a programming
33. ilable until after the program has been run once In this way a programmer can arrange the content of a pict box using actual images instead of just Scheme expressions Pict boxes illustrate how a programming environment can provide WYSIWYG tools that complement the language s abstraction capabilities A pict box can be placed in any expression position and Scheme boxes are lexically scoped within the pict box expression For example a pict box can be placed into a function and then a Scheme box within the pict box can access arguments of the function In this way a programmer can use both WYSIWYG editing and programmatic construction of picts instead choosing one or the other 7 Slideshow Design and Related Work By far the most closely related work to Slideshow is Slithy Zongker 2003 which is a Python based tool for building slide presentations Slithy is designed specifically to sup port movie like animation whereas Slideshow mainly supports slide stepping animation Like Slideshow however Slithy provides a programmatic interface for slide construction manages the display of slides directly and supports an explicit set of design principles for presentation authors Slithy s slide construction combinators seem more primitive than Slideshow s most of the work to implement a Slithy presentation is in building animations directly much like building pictures with Slideshow s dc primitive A pict like layer of abstraction would
34. ing titles with a slide selection dialog and enabling clickable elements within interactive slides Abstractly a slide presentation is a sequence of picts Thus a presentation could be represented as a list of picts and a Slideshow program could be any program that gener ates such a list We have opted instead for a more imperative design at the slide level a Slideshow program calls a slide function or variants of slide to register each indi vidual slide s content 1 We illustrate the effect of s1ide by showing a framed scaled version of the resulting slide s pict Slideshow Functional Presentations 9 slide scale big fish 10 We choose imperative registration through s 1 ide because a slide presentation is most eas ily written as a sequence of interleaved definitions and expressions much like the examples in section 2 A programmer could thread a list through the sequence but threading is awk ward to read and maintain The picts that are registered for slides remain purely functional i e they cannot be mutated so a small amount of imperative programming causes little problem in practice Furthermore we usually write slide at the top level interspersed with definitions so each use of slide feels more declarative than imperative The slide title function is similar to slide except that it takes an extra string argument The string is used as the slide s name and it is also used to generate a title pict
35. libraries to use in constructing talks Many tasks can be automated entirely such as typesetting code and animating reduction sequences All of the figures in this paper are generated by Slideshow s pict library using exactly the code as shown In fact like many other picture languages Slideshow began as a sys tem for generating figures for papers and the core pict language works equally well on paper and on slides A picture language alone is not enough however most of our ef fort behind Slideshow was in finding appropriate constructs for describing staging and rendering slides For further information on using Slideshow and for sample slide sets including slides for conference talks and two courses see the following web page http www plt scheme org software slideshow 6 We used Slideshow version 299 24 and we re defined the slide operations to produce boxed picts 36 R B Findler and M Flatt Appendix Annotate implementation 77 add bbox pict gt pict define add bbox p apply colorized add bbox p 7 annotate pict gt pict define annotate p apply colorized annotate p 7 annotate all pict gt pict define annotate all p apply colorized annotate all p 7 Non colorized versions 7 add bbox pict gt pict define add bbox p refocus cc superimpose p rectangle pict width p pict height p P 7 annotate pict gt pict define annotate p add bbox la
36. ned the two overlays we can write a single function that com bines the them Since place over does not adjust the bounding box the weight label is placed below its input pict So when we combine the overlays we must be careful to use ct superimpose 7 overlay pict frame pict main char gt pict define overlay pict fish let weight lbs weight fish w weight weight overlay pict fish weight ct superimpose w weight arrow overlay w weight fish weight A A add bbox overlay mainl small fish 10 Ibs 3 4 Put It All Together All that remains is to define a function that puts together the aquarium the main characters and the overlays and then call the function once for each frame in the scene fish scene listof pict main char gt pict defin fish scene set let frame in aq main characters set apply ct superimpose frame map A fish overlay frame fish set fish scene list dinner fish lunch fish small fish 10 Ibs 10 Ibs 10 Ibs Slideshow Functional Presentations 19 fish scene list dinner fish lunch fish small aaah fish 10 Ibs 10 Ibs 10 Ibs fish scene list dinner fish medium fish 10 Ibs 20 Ibs fish scene list dinner fish medium aaah fish 10 Ibs 20 Ibs fish scene list large fish 30 Ibs 3 5 Exploiting the Design Recipe By following the design recipe we have established a clear connection between the scene s storyboa
37. ng toolbox The dc function takes an arbitrary drawing procedure and a height width ascent and descent for the pict When the pict is to be rendered the drawing procedure is called with a target drawing context and an offset For example the four argument blank function is implemented as 7 blank num num num num gt pict define blank w h a d dc A dest x y void wha d More usefully we can define a hook function that is implemented using the draw spline and draw line methods of a primitive drawing context Slideshow Functional Presentations 27 7 hook num num gt pict define hook w h dc A dest x y let mid x x w 2 mid y y h 2 far x x w far y y h d w 3 send dest draw spline x y x far y mid x far y send dest draw spline mid x far y far x far y far x mid y send dest draw line far x mid y far x d mid y d wh 0 h define go fish lL ht append 5 hook 5 12 lunch fish D The primitive drawing context is highly stateful with attributes such as the current drawing color and drawing scale Not surprisingly slides that are implemented by directly manag ing this state are prone to error which is why we have constructed the pict abstraction Nevertheless the state components show up in the pict abstraction in terms of attribute de faults such as the drawing color or width of a drawn line In particular the linewidth colo
38. on Scheme and functional programming Shivers Olin amp Sae Tan Wendy 2004 Functional PostScript Industrial strength 2D functional imaging In preparation http www scsh net resources fps html Steele Guy Lewis 1977 Debunking the expensive procedure call myth or procedure call im plementations considered harmful or LAMBDA the ultimate GOTO Pages 153 162 of Proc ACM conference Van Wyk Christopher J 1981 IDEAL user s manual Computer science technical report AT amp T Bell Laboratories CSTR 103 Van Zandt Timothy Prosper prosper sourceforge net Zongker Douglas 2003 Creating animatation for presentations Ph D thesis University of Wash ington Zongker Douglas amp Salesin David 2003 On creating animated presentations Eurograph ics SIGGRAPH symposium on computer animation
39. or both text editing and typesetting Computer scientists may prefer Emacs and I4TRxX because text editing has little to do with typesetting and these different tasks are best handled by different specialized applications More importantly tools such as Emacs vi and ATEX are programmable Through the power of programming abstractions a skilled user of these tools becomes even more efficient and effective Shockingly many computer scientists give up the power of abstraction when faced with the task of preparing slides for a talk PowerPoint is famously easy to learn and use it produces results that are aesthetically pleasing to most audience members and it enables users to produce generic slides in minutes Like most GUI WYSIWYG oriented applica tions however PowerPoint does not lend itself easily to extension and abstraction Pow erPoint provides certain pre defined abstractions the background the default font and color etc but no ability to create new abstractions Among those who refuse to work without abstraction many retreat to a web browser because HTML is easy to generate programmatically or the various extension of TEX 2 R B Findler and M Flatt plus a DVI PostScript PDF viewer Usually the results are not as aesthetically pleasing as PowerPoint slides and not as finely tuned to the problems of projecting images onto a screen Moreover novice users of TEX based systems tend to produce slides with far too many word
40. rd and its source code If we imagine a change to the storyboard the design recipe helps us determine how to modify the implementation For example suppose we decide to color the arrows that connect the weights to the fish The corresponding code is the the arrow overlay function Simply adding a colorize expression to its body produces colored arrows define arrow overlay frame fish lbs colorize add arrow line 8 ghost frame lbs find ct fish find cb purple fish scene list dinner fish lunch fish small aaah fish 10 Ibs 10 Ibs 10 Ibs 20 R B Findler and M Flatt fish scene list large fish 30 Ibs This modification relies on implementing the supporting cast as overlays If we had not followed the design recipe and placed the arrow directly on the scene colorize would have colored the weights as well as the arrows 7 Not what we wanted colorize add arrow line 8 mainl weight small lbs find ct D D small fish find cb purple 10 Ibs As a second change suppose that we want to make the fish grow pudgier as it eats instead of growing equally in both dimensions This change to the storyboard affects only the main characters and correspondingly we need to modify only the definitions of the main characters The aquarium and arrows adapt automatically to the new fish dimensions 7 pudgy pict num gt pict define pudgy p n scale p expt n 1 4 expt n 3 4 define medium fish launder
41. ring and boolean to indicate whether the fish s mouth is open standard fish 30 20 left blue black t D The standard fish function s many arguments make it flexible If we need multiple fish but do not need all of this flexibility we can define our own fish function that accepts only the color and whether the fish s mouth is open 7 fish strf color bool gt pict define fish color open standard fish 30 20 left color black open define big fish fish blue f D define big aaah fish fish blue t 2 2 Adjusting Picts If we need fish of different sizes after all instead of adding arguments to fish we can simply use Slideshow s scale function define little fish scale fish orange f 0 7 TD We certainly want to place our fish into an aquarium which we can draw as a light blue rectangle behind the fish The rectangle function does not accept a color argument instead it generates a rectangle that uses the default color and the colorize function lets us adjust the default color for a pict Thus we can create a light blue rectangle as follows colorize filled rectangle 20 10 sky blue 2 3 Combining Picts To create a pict with two fish we can use ht append ht append 10 little fish big aaah fish D 4 R B Findler and M Flatt The first argument to ht append is an amount of space to put between the picts It is optional and it defaults to 0 The ht pa
42. rize and scale primitives change the line width color and scale of a pict pro duced by hook scale 4 colorize linewidth 1 go fish red 2 Although the underlying drawing context includes a current font as part of its state a pict s font cannot be changed externally unlike the pict s scale color or line width Changing a pict s font would mean changing the font of sub picts which would in turn would cause the bounding box of each sub pict to change in a complex way thus invalidating computations based on the sub pict boxes We discuss this design constraint further in section 5 2 4 3 Sub Picts When a pict is created with place over the resulting pict stores the location of the two input picts relative to the result bounding box Since all pict combination operators are defined in terms of place over all combination operators reliably track sub pict locations The find 1t primitive searches a pict s tree of sub picts All other find variants are easily implemented with find lt pict width pict height pict ascent and pict descent The panorama primitive also searches a pict s tree of sub picts to determine the enclosing bounding box Finally three other primitives manipulate sub pict information The ghost and launder primitives provide independent control of a pict s image and sub pict records The scale function which is primitive is terms of adjusting pict images also explicitly manages the effect of
43. rt of the name ht append indicates that the picts are horizontally stacked and top aligned Analogously the hb append function bottom aligns picts If we want to center align picts we can use hc append define two fish hc append 10 little fish big aaah fish D D Now we are ready to place the fish into an aquarium Our old aquarium colorize filled rectangle 20 10 sky blue Sj is not large enough to hold the two fish We could choose a fixed size for the aquarium but if we later change the size constants in the f i sh function then the aquarium might not be the right size A better strategy is to create a function aq that takes a pict and makes an aquarium for the pict 77 aq pict gt pict define aq p colorize filled rectangle pict width p pict height p sky blue aq two ish Sa The pict width and pict height functions take a pict and produce its width and height respectively This aquarium is large enough to hold the fish but it s tight We can give the fish more room by adding space around two fish with Slideshow s inset function and then generate an aquarium pict Finally we put the fish and aquarium together using Slideshow s cc superimpose function ip in ag pict gt pict define in aq p cc superimpose aq inset p 10 p in aq two fish The leftmost argument is placed bottommost in the stack so the fish end up on top of the aquarium rectangle The cc
44. s and far too few pictures due to the text bias of their tool Meanwhile as a programming language TEX leaves much to be desired Slideshow a part of the PLT Scheme application suite PLT n d fills the gap left by abstraction poor slide presentation systems First and foremost Slideshow is an embed ded DSL for picture generation but it also provides direct support for step wise anima tion bullet style text slide navigation image scaling to fit different display and projector types cross platform consistency Windows Mac OS and Unix X and PostScript output for ease of distribution Functional programming naturally supports the definition of picture combinators and it enables slide creators to create new abstractions that meet their specific needs Even better the Slideshow programming language supports a design recipe to help slide creators build and maintain animation sequences Our design recipe guides the programmer from a storyboard sketch to an organized implementation and it also suggests how changes in the sketch translate into changes in the implementation Even with the best of programming languages some slide sequences can benefit from a dose of WYSIWYG construction WYSIWYG tools should be part of the slide language s programming environment analogous to GUI builders for desktop applications We have implemented extensions of the DrScheme programming environment that support interac tive slide construction in com
45. s us to add or remove it without disturbing the rest of the scene We can obtain a separate pict for the weight overlay by simply ghosting the input frame 7 weight overlay pict frame pict main char pict weight gt pict define weight overlay frame fish weight define values fx fy find cc frame fish place over ghost frame fx pict width weight 2 pict height frame 6 weight add bbox weight overlay main1 small fish lbs weight small fish 10 Ibs The other supporting cast members are the arrows To draw an arrow from the small fish s weight to the small fish we can use add arrow line add arrow line 8 mainl weight D D small lbs find ct small fish find cb 10 Ibs 3 See the appendix for the implementation of add bbox which adds a gray box to the given pict We use add bbox to show the location of the bounding box in a mostly blank image 18 R B Findler and M Flatt Again to satisfy the design recipe for this step we must build a pict that draws only the arrow add bbox add arrow line 8 ghost mainl weight small lbs find ct small fish find cb A Once we have designed the arrow overlay for a single frame we can abstract over its construction as with the weight overlay arrow overlay pict frame pict main char pict weight gt pict define arrow overlay frame fish lbs add arrow line 8 ghost frame lbs find ct fish find cb Now that we have defi
46. scaling on sub picts locations 28 R B Findler and M Flatt 4 4 Primitives To summarize the following are the essential primitives for Slideshow picts dc the main constructor of picts pict width pict width pict descent and pict ascent dimen sion accessors place over combination operator find lt launder ghost panorama operations to manage sub picts e scale linewidth and colorize image adjusting operations All other operations can be implemented in terms of the above operations For historical reasons the primitives in the implementation are less tidy than this set but Slideshow continues to evolve toward a more principled implementation 5 Rendering Slides Slideshow is implemented as a PLT Scheme application PLT n d PLT Scheme provides the primitive drawing contexts accessed by dc as well as the widget toolbox for imple menting the Slideshow presentation window and other interface elements Slideshow is designed to produce consistent results with any projector resolution as well as when generating PostScript versions of slides The main challenges for consistency concern pict scaling and font selection as discussed in the following sections We also comment on Slideshow s ability to condense staged slides for printing to pre render slides to minimize delays during a presentation and to trigger interactions during a presentation 5 1 Scaling Since 1024x768 displays are most
47. sponding fish and the arrows point from the top of the weight to the bottom of the fish 4 Design an expression that for each frame combines the main character pict and the supporting cast picts Each of the following sections implements one step in the recipe 3 1 Main Characters Our main character expressions can re use the fish function from section 2 define lunch fish fish orange f define dinner fish launder lunch fish define small fish fish blue f define small aaah fish fish blue t define medium fish launder scale small fish sqrt 2 define medium aaah fish launder scale small aaah fish sqrt 2 define large fish launder scale small fish sqrt 3 In general each of the main characters in our scene must have a separate identity from each of the others and must not contain any of the others as sub picts Otherwise the overlays we build in step 3 might find the wrong main character 2 Technically this particular scene does not need the the small medium or large fish to be laundered but for consistency we launder them anyway The dinner fish on the other hand must be laundered or the scene will look wrong As you continue reading see if you can guess how Slideshow Functional Presentations 15 3 2 The Scene s Frames The first frame in our scene has three fish lined up in a row so we begin by combining our fish with hc append We add space be
48. strates the design recipe for a single scene i e a sequence of related picts in a presentation The example scene animates our earlier fish example to illustrate conservation of mass in an aquarium when one fish eats another the total mass of the aquarium content does not change The first step in designing any scene is to create a storyboard for the scene A storyboard sketches the sequence of individual frames each of which corresponds to a slide Our example storyboard is shown in figure 1 AAA pores ae ot o Olbs wibs OLS moth aihe oper Fig 1 Storyboard for the example talk Once we have a storyboard we must identify the main characters that is the elements of the picture that define the overall layout across frames We call the rest of the frame elements the supporting cast they will be layered on the basic frame as determined by the main characters In this case the fish are the main characters because the layout of the fish determines the overall shape of each frame in the sequence In contrast the water in the aquarium and the weights and arrows are the supporting cast We can draw the aquarium under the frame shaped by the fish and we can draw the arrows and weight labels on top The programming task is to convert a scene and the characters into expressions that generate picts for the frames There are four steps in our recipe 1 Design an expression that produces the pict for each main character A part
49. that is placed above the supplied content picts The title pict uses a standard configurable font and is separated from the slide content by a standard amount Big Fish slide title Big Fish scale big fish 10 The slide and slide title functions do not merely register a slide If they did pro grammers would prefer to use more elaborate abstractions and part of Slideshow s job is to provide the most useful of such abstractions Thus Slideshow allocates the relatively short names slide slide titl1e etc to functions that provide additional functionality The simplest such addition is that each s1ide function takes any number of picts and it concatenates them with vc append using a separation of gap size which is 24 The slide function then ct superimposes the appended picts with a blank pict represent ing the screen minus a small border The slide center function is like slide ex cept that it centers the slide content with respect to the screen The slide title center function accepts a title and also centers the slide Big Fish Mmmm fish slide title center Big Fish scale big fish 10 text Mmmm fish swiss 32 10 R B Findler and M Flatt The set of pre defined s1ide layouts includes only the layouts that we have found to be most useful Programmers can easily create other layouts by implementing functions that call slide 2 8 Managing Text In the spirit o
50. tween the fish using a blank pict instead of an initial number to hc append to make the space explicit define spacer blank 10 0 hc append lunch fish spacer dinner fish spacer small fish D D D Next we add the eating fish small aaah fish through large fish The posi tion of each eating fish is determined by one of the three initial fish Since the nose of each eating fish such as medium fish should line of with the nose of an exist ing fish such as Lunch fish we might try to add the eating fish into the scene with lc superimpose hc append lc superimpos large fish dinner fish spacer lc superimpose medium fish medium aaah fish lunch fish spacer lc superimpose small aaah fish D D D small fish On close examination of the spacing however we can see that the medium and large fish have stretched the space between the original picts In fact the space between the dinner and lunch fish is no longer the same as the space between the lunch and small fish To see the spacing problem clearly we can ghost out the medium and large fish and compare it with the original pict that shows just the first frame hc append lc superimpos ghost large fish dinner fish spacer lc superimpose ghost medium fish ghost medium aaah fish lunch fish spacer lc superimpose small aaah fish small fish D D D To fix the problem we push the calls to hc append inside the calls to 1c superimpose
51. ult is always a pict value When the pict box is evaluated its Scheme box expressions are evaluated to obtain sub picts and these sub picts are combined using the relative positions of Scheme boxes in the 5 The 1c superimpose call in the third line has no black box because it is in tail position To preserve tail calls Steele 1977 Clinger 1998 our prototype tool only records the values of expressions that are syn tactically in tail position Slideshow Functional Presentations 33 e008 fish scm DrScheme 00e fish scm DrScheme O fish e siecle asf Step Q Check Syntax A Run Stop taSave esf Step Q Check Syntax Z Run Stop define fish scm v define v define freeform swimmers define freeform swimmers A AJ x A lunch fish lunch fish D D lunch fish D lunch fish P a lunch fish lunch fish T 1unch fish D D l lunch fish 1unch fish o Fig 4 A pict box containing Scheme boxes source pict box In this example the program defines freeform swimmers as a pict that contains a school of manually positioned fish The right hand side of figure 4 shows the same program but after the programmer has toggled the box to preview mode In preview mode each Scheme box is replaced with the image of the pict that resulted from the most recent evaluation of the Scheme box Preview mode is not ava
52. upports a design recipe for slide sequences In the IDEAL Van Wyk 1981 picture language programmers define pictures by de scribing constraints such as arrow X s endpoint is attached to box Y s right edge In Slideshow the programmer effectively writes a constraint solver manually using functions like pict width ands find 1t We have opted for the more direct functional style instead of a constraint based language because we find that many patterns of constraints are easily covered by basic combinators such as vl append while other patterns of constraints like adding lines to connect nodes are easily abstracted into new functions Our choice of direct computation instead of using constraints affects the kinds of prop erties that can be adjusted from outside a pict As discussed in section 5 2 a pict s font cannot be changed externally because changing the font would invalidate computations based on the pict s bounding box In a constraint based system or where all primitive pict combination operations are encapsulated in operations like vc append then invalidated computations can be re executed With more general combinations using place over however the offsets are computed by arbitrary Scheme code so that automatic re calculation is not generally possible Functional reactive programming Elliott amp Hudak 1997 might be the right solution to this problem and we hope to explore this possibility Meanwhile Sl
53. xception of the first eating fish each eating fish is lc superimposed on one of the food fish and the rest of the scene is placed just behind with spacer in between The procedure in the body of main characters constructs this portion of the scene Using fold1 we iterate over the three lists of fish The second argument to fold1 is the initial pict which contains only the first eating fish in this case The last three arguments of fold1 correspond to lists of arguments to the folding procedure define main characters active foldl A meal fish eating fish eating aaah fish rest of scene lc superimpose on active eating fish on activ ating aaah fish hc append on active meal fish spacer rest of scene cc superimpose on active car eating fish on active car eating aaah fish meal fish cdr eating fish cdr eating aaah fish main characters append meal fish eating fish eating aaah fish The weight and arrow overlays remain the same per fish in the storyboard so the code from the third step of the design recipe does not change To follow the recipe s fourth step and complete the modification we define ith to accept a frame number and produce the corresponding pict using our earlier definition of fish scene 7 ith num gt pict define ith i fish scene cons list ref if even i eating fish eating aaah fish quotient i 2 sublist meal fish quotient i 2
54. y to animate an algebraic reduction making the expres sion movements easier to follow In the future the pict abstraction might be enriched to support more interesting kinds of animation 6 Environment Support Slideshow programs can be implemented using the DrScheme programming environment Findler et al 1997 since Slideshow is an extension of PLT Scheme All of DrScheme s Slideshow Functional Presentations 31 e098 fish scm DrScheme fish v pe esf Step Q Check Syntax Run BStop Welcome to DrScheme version 299 24 cvs 13dec2004 Language Graphical MrEd includes MzScheme gt main characters append meal fish eating fish eating aaah fish Fig 2 Picts in DrScheme s REPL programming tools work as usual including the on the fly syntax colorer the syntax checker the debugger and the static analyzer Non textual syntax can be used in a Slideshow pro gram such as a test case boxes comment boxes or XML boxes which support literal XML without concern for escape characters Clements et al 2004 More importantly we can use DrScheme s extension interface to add new tools to DrScheme that specifically support slide creation 6 1 REPL Picts To better support interactive development DrScheme s read eval print loop REPL prints picts as they are drawn in a slide Figure 2 shows a call to main characters and the resulting pict Each pict result in DrScheme s REPL is itsel
Download Pdf Manuals
Related Search