W3C Workshop on Delivery Context
Sun Microsystems, Inc.
In this paper we outline the apparent overlap in the scopes of the Device
Independence Activity, the Voice Browser Activity, and the proposed Multimodal
Interaction Activity. We go on to identify current limitations of
single web authoring approaches for multiple delivery contexts that impose
a tradeoff between interface optimization for limited devices and ease
of use for customization. We also discuss the need for current device
independent delivery systems to limit the level of content adaptation to
multiple delivery contexts for the sake of performance and maintainability.
We then raise the question of whether delivery context considerations need
to be incorporated into service and application specific preferences.
We conclude with a number of recommendations that we hope will provide
input into the rechartering of Working Groups within the Device Independence
The Device Independence
Principles  document states:
"The focus of the W3C Device Independence Activity is on making
the Web accessible anytime and anyhow, in particular by supporting many
access mechanisms (including mobile and personal devices, that can provide
access anytime) and many modes of use (including visual and auditory ones,
that can provide access anyhow)."
This focus seems to be closely related to the efforts of the Voice
Browser Activity  and the proposed Multimodal
Interaction Activity . While we acknowledge that the voice
browser and multimodal interaction efforts warrant the attention of Working
Groups, we are not clear on the need for three separate activities.
According to the W3C
Process Document , Groups are created within Activities and the
creation of single-group Activities should be the exception, not the norm.
We also would like to know how device independence is or is not distinguished
from multimodal interaction. Where does independence from a particular
mode leave off and the ability to use multiple modes begin?
Single Web Authoring Limitations
While single web authoring for multiple delivery contexts is clearly a
goal, current authoring approaches present a difficult tradeoff.
On the one hand, the multitude of devices that are now accessing the web
and the expected introduction of many more types of devices in the future
represent vastly disparate capabilities. Limited devices, where screen
real estate, bandwidth, etc., are at a premium require careful design of
interfaces to maximize their usefulness. On the other hand, the huge
amount of content on the web will continue to grow and change and authors
need to easily customize the presentation of this content.
Current authoring solutions either provide the flexibility to optimize
interfaces to specific devices while forcing authors to manage the complexity
of multiple delivery contexts, or impose a proprietary abstract intermediate
representation that cannot be adequately adapted to fully utilize the capabilities
of all devices.
Device Independent Delivery Systems
The promise of a device independent delivery system is that it can identify
the delivery context and adapt content appropriately and sufficiently from
the perspective of the user and the author of the content. This often
involves dynamic determination of the delivery context capabilities and
associated user preferences and then some processing to convert the content
into a suitable form.
To achieve the efficiencies necessary in current high performance and
maintainable systems, the identification and adaptation processes must
be constrained. Delivery contexts must currently be grouped together
and preconfigured instead of dynamically constructed for every request.
Also, content adaptation must currently be done primarily through a selection
process instead of a transformation process, again because of performance
requirements. Therefore, only a subset of all possible permutations
of delivery context capabilities and preferences can be handled in current
systems if high performance and maintainability are required.
Service and Application Preferences
While the CC/PP Working Group
 has defined how user preferences related to device capabilities are
structured and handled, it is not clear how to structure and handle user
preferences for service and application parameters that are device specific.
For example, a user may only be interested in new messages when accessing
his or her email from a phone as opposed to a personal computer.
Ultimately, device capabilities and preferences should be independent of
service and application preferences. But current systems often require
a tight coupling and manual user configuration of the relationship between
these preferences. How deep should awareness of the delivery context
go? Should services and applications change their behavior based
on different delivery contexts?
We recommend that the Voice Browser Activity and the Multimodal Interaction
Activity be merged under the Device Independence Activity and their respective
Working Groups be brought over under the latter; this may need a slight
modification of the Device Independence Activity, and perhaps a new name
We would also like to see further work in the Device Independence Activity
current limitations of single web authoring approaches,
the needs of high performance and maintainable device independent delivery
the role of services and applications in delivery context awareness.