This document collects a number of scenario's which relate to the viscous triangle. They all describe a situation where a choice needs to be made between multiple alternatives of which there is no obvious best one. The "best" solution is dependent of contextual information and consequently changes when the context changes.
The scenario's are used in the first place to get a better understanding of what the vicious triangle exactly is. A secondary function is to learn about, and compare problems and trade-offs in other systems. When writing papers you might use them to illustrate your problem in a more convincing way.
| Author | Joost |
| System | Cuypers |
| Description | Scenario which describes conflicting strategies for author, designer and visitor. |
The objective of this simple scenario is to identify some of the choices which a presentation engine needs to make. The choice made is mostly arbitrary and is dependent of the goal of your presentation.
One of the goals in the presentation is to compare 4 concepts. The ideal case, according to the rhetoric would be that the four concepts can be compared simultanously. If the concepts can be represented by images, this means 4 images are presented at once on the screen.
In a presentation about Rembrandt's work his use of chiaroscuro is compared to work of other chiaroscuro artists. The objective of this comparison is to get the viewer acquainted with the chiaroscuro technique.
Suppose the situation is not optimal and the 4 selected images can not be presented together at once because of insufficient screen space. To cope with this situation there are alternative way the presentation can adapt.
Badness: smaller media items lack detail for comparing, images should be of similar quality.
Badness: smaller media items lack detail.
Badness: Audio and text are serializations of content which might not be very well suited to comparing
Badness: Inconvenient for a user
Badness: comparing is harder especially for complex images
Badness: Expensive operation since intermediate results are typically no longer valid.
Content, Device, Presentation Structure, Plot
Four images cannot be presented together. The suggested possible solutions happen at different stages during the generation phase. The substitution, for example, happens when the 4 concepts gets 'materialized', in Cuypers this is when a PS gets transformed to a HFO. Changing the plot, in contrast is done after the query returned its results which needs to be structured according to a narrative. The choice for an alternative presentation medium might also influence the presentation structure which needs to cope with time and synchronization.
| Author | Lynda |
| System | AHA! demo |
| Description | Scenario which describes conflicting strategies for author, designer and visitor. |
There are a number of pieces of text
describing parts of a painting. Each fragment is written from
the perspective of an author persona. There is a designer
responsible for the layout/look and feel of the final
presentation. The visitor is the flesh and blood person who
is exposed to the presentation. The user is an abstract
notion that can (to at least some extent) be modelled in the
system. A piece of text is selected
depending on the values of a number of parameters. E.g. the
knowledge of the user in the topic; the preference of the user
for visual or verbal information. By selecting a text
fragment that is appropriate we can increase or decrease the
motivation and/or ability of the user. (Opportunity is more
determined by external influences.)
Bad aesthetics can decrease the visitor's motivation.
Bad text legibility can decrease the visitor's ability to
assimilate the text. Bad choice of text versus image can
do the same.
(There are already
conflicts in creating the text fragment as if written by a
particular author persona for a user profile. These are
before we even start thinking about creating a presentation,
so totally out of scope of this scenario.)
Erica, a teenager from high school, has little interest in
history, and doesn't even like the dark, sombre painting she
is looking at. However, the girl in the foreground of the picture
caught her attention, so she became curious about the scene
and indeed why she has a dead chicken hanging from her waist.
Erica points to the picture of the chicken and expects to be
told something about it. There is a long piece of text
with an explanation, but this is too long for Erica's
attention span. There is no appropriate picture to display
with the new text.
Badness: The price is that the information will not be conveyed, but Erica stays motivated.
Badness: Erica might lose interest mid-way, but she can at least continue studying the picture.
Badness: The language style of the story will suddenly change drastically. This will make the visitor feel less comfortable with the system (= motivation decreases).
Badness: As it happens, Erica is also fluent in Dutch. The appropriate text is presented in that language - which she may find odd.
Need extremely detailed user model. Some of this can be input
beforehand ("standard" knowledge about how people work). Some
needs to be extracted from user (colour, learning style
preferences).
Domain model related to user model. (Standard AHAM model,
where user model is "shadow" of domain model.)
Model of design aesthetics related to user model. Need to be
able to model preferences of user in some better way than "I
like orange".
Need to know where the presentation is in the main progression
of the "story".
| Author | Frank |
| System | Amit's, Katharina's and Karolina's work |
| Description | Scenario which illustrates the importance of form before function and cases where function should come before form. |
The scenario assumes that the presentation engine has to generate a presentation about the art style "De Stijl".
The decision about the level of function or form within a presentation is an ongoing process that, however, is weighted for one or the other at a very early stage of the presentation generation (depends on the overall communication goal and user type).
Functional means here that the user is facilitated to perceive information units in a quick and concise manner. This requires, for example, that information units are legible and ordered regarding to importance. Functional access to information is related to communicational goals, such as learning, educating, collecting facts, specifying concepts, etc. The typical user for such information is the mature and active (get information fast).
A form-oriented presentation, on the contrary, is rather oriented to adopt harmony schemata or particular style parameters of the topic to be presented. Important here is the impression of the presentation. Here distortions to the information unit are possible. Form-oriented access to information is related to communicational goals such as entertaining, relaxing, leisure browsing, etc. The typical user for such information is passive - or a child.
The user has a colour deficiency => red and green cannot be assigned to items that need to distinguished.
The topic to be presented requires a certain use of colours => use the colours of the style to be presented (De Stijl = [red, blue, yellow])
Important information is of type "text" => make it legible (light and cool colours for the background, dark and warm colours for the foreground)
Important information is of type "text" => use a san-serif font
The relevant communication goal is "function" => make navigation as well as orientation elements distinct (e.g. use colours that are different enough to those assigned to information units of type text)
The relevant communication goal is "function" => keep layout of information, navigation and orientation spaces consistent
User model (age, physical abileties,...), discourse model, domain ontology
| Author | Lloyd |
| System | Topia/Noadster |
| Description | The user requests a presentation about a stated topic. The system finds informative resources conveying that topic. Thes system then composes these resources, both media objects and representations of abstract concepts with links to media, into a single hypermedia presentations. |
Established technologies let queries return matching resources, but as an unstructured collection. Coherent hypermedia presentations use compositional structure to relate the resources to each other. Typically, at least at the highest level, this presentation structure conveys the basic aspects of document structure: hierarchy, sequence and cross-references. Hypermedia structure, both high and low level, is currently only created by humans. For further automation of this process, the system must generate document structure around returned resources. It must also translate this document structure to hypermedia structure of presentation.
The returned atomic resources sit in an RDF-defined network. Each resources is either directly a media object or is an RDF-defined conceptual resources with triples linking it to media objects for conveying it. The system uses this network joining the returned resources to determine clusters of them. These clusters form the basis of a document hierarchy for the resulting presentation. They also form cross-references within this hierarchy. One way of generating sequence is based on the relevance measures for the resources given the query. Given a generated document hierarchy, there are various readily recognizable means of using hypermedia structure to convey this document structure around the media objects.
The knowledge base must have media objects for integration into a presentation. It should also have conceptual RDF-defined resources that link to media objects either directly or indirectly through other conceptual resources. Both types of resources need to be findable with automatic queries, such as text search on content of literal values of a resource as subject. In order to generate useful clustering, this graph should be well-connected, enabling as many different combinations as possible of selected resources within the linked graph to bring hierarchical clusters. Knowledge about the conceptual domain and user preferences can help guide the clustering. For example, varying weights of significance can be placed on certain types of resources and predicates to vary relevance of clustering for use in choosing among alternatives for clusters.
This technique is domain-independent, working for any RDF
repository. In addition to the requirements described
above, the presentability of such repositories benefits from
the user of
| Author | Stefano |
| System | IWA |
| Description | Scenario in which the plot is adapted because of lack of suitable material versus a scenario where poor material is chosen for not altering the plot. |
Unfortunately I do not have experience with
plot on a large scale, but I experimented on a small scale with plots
that had a structure of an argument (like an complex answer to one
question). My feeling is that in video form is much more important than
in mixed Cuypers-like presentations. You can really miss the point when
seeing my assembled video sequences, while when you hold a caption in
front of the user'nose for a reasonable amount of time you can safely
assume the viewer got the message (of course you can use captions in
video too).
This gives the idea that even if the direction or communicative goal is given, you compose
a sequence in an event-driven (or visual-driven) fashion, trying to fit
what you have in a library of plot template, but you choose the plot
based on the material, not the other way around.
The conflict for me is more of the kind "academic presentation" vs
"flashy superficial", i.e. have more visual effects and modern editing
or leave the people talk for hours. In video this can be extreem, but
also in presentation if you think you could make a slideshow with .1
sec per painting. You do not get to see the painting but you get the
idea the painter painted a lot and must be important.
Again you can choose either one of the two style, but if it fails because of the
material I would fall back on the other one.
I use a bottom-up approach for the plot, trying to create micro
aggregations that are further composed if a plot from the library seems
applicable, and as much as this plot is applicable. I do not think to
back track and change the material (maybe add more material) because
the micro aggregations are the scope where my semantics hold, and going
out of there there is no guarantee of "coherence". I guess this holds
for a computer, that has to bend the plot requirements (i.e. choose
another plot) to the data available while a human could
"creatively" find solutions to keep the plot the same.
The Semantic of the source +
the context where this holds (for ex. I can put video segments after
each others showing different people because my semantics does not
break down if the speaker is not the same, as in other genre of films).
Library of applicable plots
Knowledge of visual composition (this is video editing).
In video there can be a big gap between semantic of the annotated video
segment (in fact the semantic of the annotation) and the perceived
effect of the video segment when showed to a viewer. The time variant
nature of the video makes this more complicated. I have scenes where
people say "I do not think so" replying to the question "Is war the
solution?", duration 1-2 sec; another saying "It is a no-win war", they
say the same but if you put the first after the second and you remove
the caption of the question they contraddict each other. What under a
certain point of view can be considered poor material (too short to
give context) is a blessing under the point of view of manipulation. In
IWA one video camera went off while in the bag, recording a black
screen for 5 min plus the sound of our steps. I used this part as the
starting shot of my version of the documentary.
| Author | Katya |
| System | SampLe |
| Description | A scenario which describes conflicting workflows |
SampLe methodology consists of 3 major stages: stage 1 Topic selection, stage 2 Genre selection and adaptation, stage 3 Material collection and arrangement. The idea is that a presentation creator (or an author) can approach the process of presentation building in a way most suitable to his/her needs. Since we cannot predict what type of user is currently working with the system, we try to provide all possible options.
There are 2 possible workflows, since choosing or manipulating
a certain genre assumes its application to a certain topic/character.
The first type of workflow is the one suggested by the current demo, e.g. the one where:
workflow 1:
In order to support the first type of workflow the system should be able to provide a discourse structure for the presentation (extensions to the general genre template of e.g. essay based on the domain knowledge). Then in case an author wants to extend this discourse structure the system has to be able to verify whether the newly built structure still corresponds to the preselected genre.
In the second workflow the system should be able to fit the selected material into a genre structure. If none of the available genre structures are appropriate, the system will try to suggest the closest genre and point media items from the selected ones which do not fit in it.
Domain knowledge, discourse knowledge, knowledge on genre structures
| Author | Katya |
| System | SampLe |
| Description | A scenario which describes a conflit between adoption of plot vs. re-selection of the material. |
Unless the stages in the workflows in the previuos scenario are explained in a sequential manner, in practice an author can switch between the stages, change previously taken decisions, and see their consequences. Imagine that following the first type of workflow an author is selecting a genre = article, since she wants to make a presentation with a strong argumentation, where presentation of the essence is followed by the explanation. The author is switching to the content selection stage and tries to collect the media material. She finds out that available material is not enough to build the argumentation for her presentation.
One solution is to try and arrange the available material in the most benefitial way. In this case the presentation might still look not convincing enough.
Another solution would be to go back to the genre selection stage and choose for a differenr genre. As a result, the author might be not satisfied with the result, since her initial plan for the presentation fails.
Domain knowledge, discourse knowledge, genre structure knowledge
$Id: scenarios.html,v 1.10 2005/01/06 13:52:39 sbocconi Exp $