Feedback on paper of Roussos et al.


FLAWS

This paper does make hypotheses -- too many to be properly evaluated in a paper of
this length. In fact, none of them is thoroughly evaluated, and some are hardly
evaluated at all. A large field is, fairly superficially, surveyed and appeals
are made to (often personal) experience of the various applications, but little
detail is given to flesh out the evidence.

Lots of unsubstantiated minor claims are made without any attempt to support
them, e.g., by citing some third-party work that provides evidence for them. 


HYPOTHESES

Below I pull out quotes from the paper that indicate the five main hypotheses I
detected. There is at least one claim for each of the main sections of the
paper. Below each claim I briefly summarise the proffered evidence for it. An
overall summary of these five claims is in the subtitle, that "mobile phones
could serve as information service end points, control devices for ubiquitous
systems, network hubs for personal and body area networks, and ID tokens".

1a. "From the pervasive computing perspective, the mobile phone's most interesting
use is as an end point for an information utility or service."

	Various potential applications are briefly mentioned and the mobile
	phone is suggested as "a short-term solution [to the problem of
	interacting with them] readily available to much of the global
	population at relatively low cost". Some home-grown realisations of such
	applications are briefly described. These are promising, but not really
	sufficient for the rather grand claim.


1b. "This view [that mobile users were simply mobile Internet users] has proven
erroneous."

	A study, by the authors, of the British Museum Compass and the Tellmaris
	systems showed that they would not meet the mobile user's needs, as it
	deluged them with information that could not be absorbed because of the
	mobile's small screen, etc. A mobile-specific recoding was more
	successful. Anecdotal evidence, but fairly persuasive nevertheless.


2. "The mobile phone offers a potential solution [to the problem that devices
cannot currently adapt automatically] by acting as a control device, similar to
remote-control devices used in consumer electronics."

	This appears just to be a personal opinion not backed with concrete
	evidence, but with some citation of relevant work to back some of the
	sub-claims.


3. "This approach [of using i-wands with an attached Mobiliser} removes the
mobile phone's form factor limitations and, taken to its extreme, lets us see
the whole personal area network as an extended telephone."

	The i-wand/Mobiliser work comes from other work of the main author and
	evidence for the claim comes from this work and from the Ericcson Phone
	Glove project. Suggestive evidence, but perhaps not quite enough for the
	'extreme' end of the claim.


4. "Mobile phones can address this challenge [of security and privacy
protection] by acting as a secure personal device that stores information used
to verify user identity and determines when to disclose this information."

	A speculative claim based on an extension of verification functions that
	mobiles already perform.  It did not seem to be based on real case
	studies and looks somewhat shaky given the dangers of phone theft and
	hacking.



YOUR REVIEWS

1. Many of you spotted the main hypothesis in the subtitle of the paper, but few
noticed that each part of this four-part hypothesis was then addressed separately in
each of the four main sections of the paper. Indeed, some people unfairly
claimed that the paper structure was disorganised. Identifying these five (1a
and 1b are both from the same section) sub-hypotheses would have helped you to
recognise the differing quality of the evidence for each of them.

2. Many of you saw the various case-studies merely as part of a survey of the
field. They were partly that, but more importantly they provided the evidence
for some of the hypotheses. They provided only anecdotal evidence, but for some
kinds of work and hypotheses it is as much as as we can expect. As a thought
experiment, consider what other kind of evidence it would be reasonable for the
authors to have presented, given the early stage of the field. 

3. People had a hard time separating the substantial hypotheses of the paper,
i.e., ones that the authors were advancing as their main contribution, from the
many minor, throwaway remarks. This was largely the authors' fault. However, one
could justifiably start from the main hypothesis stated in the conclusion and
then notice how the four parts of this were reflected in the main sections of
the paper. Some people misidentified problems and similar issues as hypotheses,
which was also forgivable given the authors' sometimes ambiguous presentation of
them.

4. Some people identified "exploratory investigation to suggest a hypothesis" as
a kind of contribution. However, I hope I've convinced you that the hypotheses
were all implicitly stated up front and did not need to be suggested by the
investigation. 

5. Many of you complained about the use of undefined jargon. But this paper was
published in a specialist journal whose readers would be familiar with these
terms and did not require them to be defined.