-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathconclusions.tex
131 lines (111 loc) · 7.65 KB
/
conclusions.tex
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
% vim:ts=1:et:nospell:spelllang=en_gb:ft=tex
\chapter{Conclusions and Future Work}
\ppt* remains the most popular presentation tool worldwide. We believe this
is mostly because people are generally afraid of change, and would rather
stick to their habits. Since all of their existing work is stored in the \ppt
format, they keep using \ppt to access that content as well as create new
presentations. On top of that, the process of creating a presentation is
still heavily burdened by the layout of the content we want to present.
Layout is an important and vastly underrated aspect of presentations in
general, which uses up an astonishing amount of time during the creation of a
presentation.
\section{Contribution}
We proposed an approach for converting existing \ppt presentations into \mxp
presentations, along with a way for the author to relinquish control over
layout to the computer and improve upon flawed human design by
programmatically calculating ideal content placement and size. We delivered
a proof of concept implementation that puts this approach into practice,
first letting us show a \ppt presentation's content in \mxp, then showing us
the possibility of applying an automated layout algorithm to that an any
other \mxp content at will.
Considering the first part of this thesis, which consists of the conversion
between \ppt and \mxp, we can conclude that conversion from any other
presentation format into \mxp is a feasible concept. Closed-source formats
will obviously be more of a challenge than their open-sourced cousins,
especially if no API has been created for them as we had the fortune with
\ppt and Apache's POI/HSLF implementation. That said, open-source formats
may be more easy to take apart but if no API exists for them it would still
require a substantial amount of effort. Having an existing API readily
available has definitely helped us a great deal in our efforts.
Converting these other formats into \mxp remains an important goal in the
endeavour to raise awareness and increase popularity of \mxp's features and
possibilities. Our implementation provides a way to convert \ppt\ .ppt
presentations, but \ppt* has switched to using the Office Open XML-based
\ .pptx format in recent years, so newer \ppt presentations cannot currently
be imported into the \mxp system.
Our implementation is written in a way that should make it straightforward
to adapt for other formats, provided there is a way to get the separate
components out of those formats. If slide-based conversion is desired, then
obviously a way to extract the components on a per-slide basis is also
required. Additionally, our approach using an intermediary form during the
conversion process allows for adaptation of the tool to generate other
output formats as well. The current implementation generates HTML5 which
includes the \mxp standard library and a set of plugins, but the original
goal of generating \mxp XML files should be easily attainable; the only
reason we did not implement this was the unavailability of the \mxp
compiler, which made it impossible for us to test the generated XML files.
As for the second part of this thesis, concerning the automated generation
of presentation layouts, we have discussed why this is necessary. When
creating traditional slideware, as well as using more advanced and modern
presentation tools, layout remains a problem that for many presenters
becomes the biggest timesink in their work. On top of that, the layout they
create is not always a good one, and bad layout has been proven to have
negative impact on the effectiveness of a presentation. As such, having a
way to automatically generate a layout would save a lot of time while also
improving the information transfer effectiveness of presentations.
We have demonstrated such an automated layout mechanism based on theory and
research found in related works, which we adapted and improved upon for our
purposes. Our constraint-based approach considers every component
separately, to combine all components into a layout where no overlapping
exists, components can be grouped together, clear margins are put in between
content and surrounding limits in the form of slides and other fixed-size
containers are respected.
In the future, the implementation of this layout mechanism may still be
improved upon in several ways, which will be discussed in section
\ref{future}. However, it currently does provide the most basic form of automated
layout, which may not always succeed in generating an aesthetically pleasing
layout but at least attempts to combine content in a way that makes the
content easy to focus on, thus increasing effectiveness of the presentation.
It also succeeds in letting the presenter focus on the content rather than
the layout. It thus reduces stress and arguably increases quality of
presentations, especially when we look at time spent creating the
presentation versus its effectiveness.
\section{Future Work}
\label{future}
In this thesis we have presented a proof-of-concept implementation of both a
tool to convert \ppt presentations into \mxp, and an algorithm for
generating an objectively effective layout. Due to the limited time
available for this thesis, we were not able to go into the finer details of
these tools, and the results of our tools could still be improved. However, within this
limited timeframe we did deliver a solid core containing the most important
features, in a way that allows future research to improve upon it and easily
add any missing details.
\subsection{Other formats}
Our conversion tool currently allows to convert \ppt\ .ppt files into \mxp
presentations, bypassing the \mxp compiler. The tool internally uses an
intermediary structure to store the presentation's content, and this
facilitates the implementation of conversion tools for other formats. As
such, it might be a good idea to extend the tool to convert other popular
formats like \ppt\ .pptx files, Apple Keynote presentations and many others.
It would also be a good idea to change the output of the conversion tool to
generate \mxp XML files to compile further using the \mxp compiler. In and
of itself this would not seem advantageous, but with \mxp IDEs and other
editing tools in mind it would be better to have XML files which would be
editable using those tools, rather than raw HTML5 which presumably would
not be readily available in any IDE.
\subsection{Integration}
While on the subject of IDEs, it would be interesting to integrate the conversion
tool into a \mxp editor. This would allow \mxp users to just open their
\ppt files in the \mxp editor, immediately providing access to its contents
and letting the user edit the presentation as if it had always been a \mxp
presentation. This would greatly improve usability of the conversion tool
as well, since it currently does not have a graphical user interface and
thus needs to be invoked from the command line.
\subsection{Improving the automated layout}
There are many ways in which the automated layout algorithm may yet be
improved. Jock Mackinlay's work \citep{mackinlay-1} includes significant
research on how to use artificial intelligence to create effective
graphical visualisations. Combining his work with ours could potentially
improve the results of our algorithm. An interesting angle here might be
the use of a learning AI, which can be trained on sets of good and bad
layouts, or observe the user's actions and try to mimic their behaviour.