CPHC/BCS Distinguished Dissertations competition 2013/14: call for submissions

The 2013/14 CPHC/BCS Distinguished Dissertations competition is now open for submissions.

Closing date Tuesday 1 April 2014. Further details can be found below and on the web at http://academy.bcs.org/content/distinguished-dissertations

The Council of Professors and Heads of Computing (CPHC), in conjunction with BCS, The Chartered Institute for IT, annually selects for publication the best British PhD/DPhil dissertations in computer science.

The scheme aims to make more visible the significant contribution made by Britain - in particular by post-graduate students - to computer science. Publication also serves to provide a model for future students.

The selection panel on behalf of BCS/CPHC consists of experienced computer scientists, not more than one from any institution, each normally serving on the panel for three years. The panel members for this year are: Russell Beale (Birmingham), Simon Dobson (St Andrews, chair), Michael Fisher (Liverpool), Joemon Jose (Glasgow), Steve Pettifer (Manchester), Iain Phillips (Loughborough), and Perdita Stevens (Edinburgh).

Any dissertation is eligible which is submitted for a doctorate in the British Isles in what is commonly understood as Computer Science. (Theses which are basically in some other discipline but which make use, even very extensive use, of computing will not be regarded as eligible.) However, there is a limit of THREE dissertations per year per university, and one per research group within any university. To be considered, a dissertation should:

  • make a noteworthy contribution to the subject;
  • reach a high standard of exposition;
  • place its results clearly in the context of computer science as a whole; and
  • enable a computer scientist with significantly different interests to grasp its essentials.
It is reasonable to submit a thesis to the scheme if it has all of the above qualities in good measure, and if it is comparable in standard with the top 10% of dissertations in the subject. Long dissertations are not encouraged; if the main text is more than 80,000 words, there should be good justification.

The dissertation should be submitted electronically (as a PDF file) by the author's examiners, or by the Head of Department with the examiner's advice. The submitted version of the dissertation must be the final version after any required corrections have been made. The competition period for the 2014 competition is for theses accepted from 1 January 2013 until the closing date of 1 April 2014. A dissertation cannot be submitted to the competition more than once.

The dissertation should be accompanied by a written nomination comprising the following information:

  • a justification, of about 300 words, by one of the examiners - preferably the external - explaining the dissertation's claim to distinction (against the criteria listed above);
  • the name of the primary supervisor and the research group within the university to which the student was primarily affiliated;
  • an assurance that within the competition period the examiners have recommended to the author's institution that the doctorate should be awarded;
  • the names and contact details of three suggested reviewers who are not in the same Department as the nominated thesis and who are independent of the supervision and examining of the thesis; and
  • an indication should be given if the dissertation is being considered for publication elsewhere.
The nominated reviewers must have confirmed that they are willing to provide a review. In addition the author's written agreement that their thesis may be considered for the Distinguished Dissertation competition should be emailed by the author to disdis14@easychair.org.

Submissions should be made on-line via http://www.easychair.org/conferences/?conf=disdis14

The first author name submitted should be that of the thesis author; the individual submitting the nomination should list themselves as the second author. The title and abstract should be those of the thesis being nominated. The first file uploaded should be the 300 word nomination; the thesis document should be uploaded as an attachment.

If any problems are experienced, or you have any questions, please email disdis14@easychair.org for assistance.

The deadline for submission is 1 April 2014.

Graduation address: "Every success is everybody's success"

I was honoured to be asked to give the graduation address at this year's St Andrews Day ceremony. The speech is below.

Chancellor, Principal, colleagues, friends, ladies and gentlemen:

Graduations are a celebration of hard work and success. And the efforts you've all made to be sitting here today certainly deserve to be celebrated: whatever your course of study, you've shown the determination, dedication, intelligence, creativity, and drive to succeed.

But being asked to give this graduation address got me thinking about the nature of success, and I'd like to share a thought with you: that success is not something we can readily ascribe to anyone individually. Rather, it's a jigsaw that assembles itself from the actions of those people you meet and by whom you are influenced. Indeed, when you get right down to it, every success is everybody's success.

To see what I mean by this, think about how much had to go right for your studies to take place. You had to be born and brought up in a way that made you emotionally able to leave home and thrive on your own, possibly mastering a different language and culture, to become a rounded individual with the skills needed to take on a university such as this. This is no mean feat on the parts of yourselves and your parents, as I hope you appreciate; not to mention your earlier teachers, friends, neighbours, and all the other people who influenced you down the years. When you came to St Andrews, I'm sure you discovered that learning and research don't occur in a vacuum. Most of you will have worked as part of a team, either in a lab or a seminar, where you came together to do something that perhaps none of you could have done individually. If you think back, I'm sure you can remember plenty of things said or done that have contributed directly to your being here today.

We can cast the net wider. The university is clean, secure, and well-managed, thanks to the efforts of porters, cleaners, secretaries, administrators, and a host of others -- efforts that tend to be hidden away and are easily forgotten, but that contribute to your studies at least as much as the efforts of your lecturers. The lights are on, the labs and seminar rooms are warm (more or less). There are coal miners in eastern Poland whom we will never meet, and who will never know to what they contributed -- but without them the wheels would not have turned, the lights would not have burned, and none of the functions that we perform in this university would have been possible. So every success that happens here is their success too.

And of course we should look through time as well as space. With this graduation we're coming to the end of celebrating the 600th anniversary of the University of St Andrews. Think what has happened over those six hundred years to get us here! All the discovery and learning, all the patient, careful scholarship down through the years, slowly building knowledge, slowly building the reputation of this University, and of Scotland, as a place to come to learn and to teach and to do research -- sparkling at this graduation today before ricocheting off into the future. Any successes any of us have here owe a debt to those who have come before us, who made this (frankly very unlikely) place possible.

If there's any substance to these musings, then it's this: success isn't the singular, individual thing that we sometimes like to think it is; but nor is it an atomised, isolated thing occurring outside a particular place in space and time and the flow of humanity. The modern world tends to focus on measurement, and the corollary that anything that can't be easily measured either didn't happen at all or at least can be safely ignored. But a moment's reflection will convince you that this is nonsense: the successes we're celebrating today have been guided and driven by influences that we would struggle to identify and certainly couldn't quantify in any meaningful way, but without which we would not all be sitting here.

This has some quite profound implications. It means that whatever you all do from today, whatever successes you enjoy in the years to come, are of real importance, no matter how small they may appear to you. Your research project may not give rise to the next Google: but it might be read by someone, who writes a thesis, that's read by someone else, that gives another person an idea, that someone else uses to change the world for the better. We'll never know the exact details of this process -- we'll certainly never measure it or report it -- but lack of public credit doesn't equate to lack of value, and that's something that can help keep us all motivated and generous with our time and our ideas.

So as well as celebrating your own individual successes today, I hope you'll also celebrate the contributions you've made to the successes of others through friendship, collaboration, advice, mutual support, tutoring -- or just simply being here. Every success really is everybody's success. Thank you, and enjoy the rest of the day.

(And in keeping with this spirit, I'd like to thank Al Dearle, Steve Linton, Linda Rafferty, and Lisa Dow, for their comments that made this speech so much better than when I first wrote it. The official press release version is on the university web site.)

Too many numbers

I've never really noticed before how over-indexed even common documents are.

Maybe it's just that we've become more sensitive to these things recently, but when I recently renewed my car's tax disc (for non-UK readers: the document that shows your car is legally on the road) that I realised exactly how much numeric information appears on the document:


There's some "expected" information that really has to appear: names, addresses, car registration numbers, fee, and the like (which I've blurred). But the real action is on the counterfoil -- a document you don't have to keep, are never asked to produce, and will basically never be seen by anyone again.

Let's start with the long sequence of numbers (1)  at the top left. Two groups of these numbers are repeated bottom right as (3) and (4); three groups are distinct and don't appear anywhere else. There's a long issue number for this document bottom-left (2). At the bottom right, (6) also appears on the disc itself as (8) -- in fact the only number  that makes it onto the tax disc itself, although there's also a barcode in the centre.

As if this wasn't enough, and despite all the numbers being printed in what is clearly a machine-readable font, there are two QR code. In the interests of science I scanned them both. (6) repeats (5) (and therefore (8)), but (7) was too small to scan with a cellphone QR code reader: it's visibly different to (6), though.

So this is eight distinct pieces of information, in the main all dutifully recorded only to be discarded when one detaches the tax disc from the counterfoil to fit it. What is it all? Since we don't have ID numbers in the UK, none of the numbers relate to me directly. I can understand a single registration number for the tax disc -- although even that's a bit redundant when you can query the tax status of a vehicle online to check whether the disc is genuine or not -- but the rest mystifies me, as does the use of three machine-readable formats on one document.

I'm not worried about the volume of information per se, as it's being discarded and -- more especially -- it doesn't seem to relate to me or my identity in any way, but I am curious as to why it all appears in the first place and what purpose any of it serves.

The closed cycle of MOOCs

Are the international rankings indirectly fuelling the rise of MOOCs? And is there a positive feedback cycle at work?

The emergence of Massively Open On-line Courseware or MOOCs is being hailed as a disruptive moment in education, similar to the revolution that overtook the music industry. By unbundling modules from degree programmes and the universities who deliver them, the promise is to allow more personalised, varied, accessible, and (most importantly) cheaper education.

It's easy to see how MOOCs benefit some students in some disciplines. Students living in remote places, from the developing world to the rural US (or rural Ireland, for that matter) can access courses from leading universities they would otherwise not be able to, or want to, attend physically. Students with disabilities, or those with significant work, family, or care commitments can more easily stitch education into their lives, freed from the constraints of structured and time-bounded degree programme. Modules available either free or at massively reduced cost will certainly broaden access and reduce the real or perceived elitism of the top institutions.

I think there are some serious caveats with the techno-utopian vision that's being propagated, not least the suitability of MOOCs for many subjects and the way that much of the innovation is more about control than about openness -- and these are topics I intend to return to. But for this post I want to focus on a narrower hypothesis: is there a positive feedback cycle in the economics of students and rankings driving at least some of the push towards MOOCs from some institutions?

What got me thinking about this is the release of the latest QS World Ranking of Universities. These are influential sources that (I can say from personal experience) influence students' (and their parents') choices about which university to attend. This is simply a fact: one can argue that position in a research-based league table will have only a moderate influence on an undergraduate's university experience and later employability, but that's doesn't stop people considering the them important. Indeed, it is such an important factor that many universities publicly make being in the top 200, top 100, or even top 5 an institutional strategic goal that influences all their decisions. (For full disclosure, St Andrews comes comfortably in the top 100 universities in the world in this ranking, and doesn't use ranking position as an element of its strategy. It's nice for us to be highly ranked, but this is a consequence of our activities not a determiner of them.)

Rising up to the top of the tables is expensive, and I suspect that it gets exponentially more expensive to climb higher the higher up one is. The top-ranked institutions are amongst the richest in the world: Harvard, Stanford, MIT, and the like all have enormous economic power alongside their obvious and uncontested academic excellence. They are also major players in the MOOC space.

At first blush, this seems counter-intuitive. MOOCs are about cost reduction and breaking the power of traditional providers: why would the top universities be involved in this, when it seems at least possible that such a move will cannibalise their bricks-and-mortar students? It's likely they don't see this as a real threat, since they're over-subscribed by several orders of magnitude more students than they can possibly accommodate. It's also possible that they see MOOCs as a branding exercise rather than one of education, raising their profile at essentially no significant cost.

However, a more economic driver also occurs to me, at least when one moves down out of the elite stratosphere. Many institutions want to move up the league tables, which often involves going shopping for star academics: people whose research excellence enhances the reputation of their employers. As an institution gets more highly ranked, the quality of academic they need to have any impact on their position also increases, and as star academics are generally more expensive, improving ranking involves increased expenditure at each step. Increasing staff cost is therefore a consequence of a strategic decision to climb the rankings.

These stars and their research infrastructure have to be paid for, and in many systems these costs more or less trickle-down to student fees. Education inflation is running well ahead of general price inflation in the US (see for example this article from 2012), and a large chunk of this comes from academic salaries. (Admittedly an even larger chunk comes from increased administration.)  The problem is that this inflation sets up a countervailing pressure, as students look at the costs of their education in terms of accrued student debt and contrast it against expected lifetime earnings -- and in some cases decide it's not a sufficiently valuable proposition. Physical institutions can't simply grow their numbers, since students attending a university have overhead costs: they have to be accommodated nearby at a price they can afford to pay, if nothing else. The pressure therefore builds up to reduce student costs while keeping the size of the student body roughly constant.

This is where MOOCs might come in. Star academics have celebrity that can be leveraged by getting them to develop MOOC courses that can be sold worldwide. Even a trickle of income (from up-front registration or charges for certification) provides a revenue stream that can be used to reduce the costs for traditional students. MOOC development is sometimes seen as cost-free (since the staff are in place already), and so the revenue feels like money for nothing. But as MOOCs become more popular, institutions require more MOOCs, and more star academics to make them, and hence more revenue to pay for these individuals and their research, and so more MOOCs: a feedback loop that might actually become self-sustaining, a bubble in MOOC provision driven by a desire for increased international rank with a stable bricks-and-mortar student body.

Any such bubble will be an issue for universities of less exalted status, since the provision of free courses from elite institutions looks set to change how students seek out knowledge, despite the fact that not everyone is an autodidact who can learn by themselves. It's still not clear what impact MOOCs will have on education in its broadest sense, and there are certainly many positive aspects that we're interested in exploring ourselves. A self-fulfilling bubble is however not something we should be indifferent to.

Scratch: visual programming for Arduino

For some people, traditional programming is daunting. This needn't be the case, and there are several more visual programming studios available -- one of which has now come to the Arduino.

Scratch is a long-running programme from MIT that lets non-programmers -- and especially kids -- get started writing complex programs. It comes from the family of languages that include Squeak and Etoys, all intended to demystify programming and computers.

Scratch has now come to Arduino by way of S4a, that lets people develop sketches using a very visual approach:

A Scratch program

It's hard to describe exactly how this sort of visual programming works, and it's definitely not suitable for all tasks: typically it's better for exploratory and experimental, "play" approaches rather than more detailed and complex processes. But it'll be great to see whether this is appropriate for many of the applications the Arduino finds a home in -- and even more so to explore it for what we're trying to do here at Citizen Sensing.

A non-prophecy from ten years ago

Something I said ten years ago comes back to not haunt me: in fact, I was quite pedestrian about the whole thing.

My friend Rich was house-cleaning and came across an interview that I did with the Irish media about ten years ago, while we were running our company, Aurium. One thing really jumped out at him was the answer I gave to one particular question:

Q: What will we be using to access the Internet 10 years from now?

A: I don't think we'll even think about it in those terms. It's a bit like asking what sort of device you use to access the telephone network. In 10 years the internet will be so ubiquitous that we won't even think about it.

The interview

I remember giving this interview, and it feels weird in a number of ways. Firstly, when you see an "expert" being interviewed in a newspaper, don't assume that he's there because of his expertise. He could simply be a randomer with a good PR firm, which is what I was: the article was "placed" as part of our public relations campaign. Secondly, I was right in terms of where the technology was going, but that wasn't due to any vision on my part: it was simply a function of being part of the development of that technology and seeing from the inside where it could go. There are plenty of alternative futures where things happened differently and the internet didn't take off as it did. Had we had a major security crisis or breach of privacy in the first few years, that might have killed people's confidence enough to damp-down the uptake.

What really struck me, though, was exactly how pervasive the technology did become. As it happens I'm in London on business with my colleague Graeme, and it's instructive to look at all the things we did on the internet -- and indeed from mobile gadgets. Firstly, I checked-in on the British Airways cellphone app and so didn't need a printed boarding pass. (This is an application we first suggested ten years ago, incidentally: a boarding pass is just a token, so why not text someone a long number to identify themselves with?) Then we arrived not knowing where the hotel was, but a combination of the Tube Map app and Google Maps soon directed us via train and foot, from Gatwick airport to Pimlico. We looked up places to drink and a place for dinner the next day using TripAdvisor, again navigating there with Google Maps, and made recommendations for the various places for other travellers. (Oh, and incidentally checked in for the return flight while sat in the pub -- very civilised, I must say.) We found a coffee shop while we were waiting for our meeting, which itself took place in a venue whose location we also didn't know relative to our hotel. I collected the different bookings and details in TripIt, which shared them into my Google Calendar so I didn't have to take note of them. And I took some pictures and shared them on Facebook.

But actually the most surprising thing isn't the technology, or the mobile device, or the fact that it all actually works together in practice: it's that real people actually do it, and I don't think I'm unique in using all this mobile internet technology when travelling. We pretty much take for granted the idea of finding our way in a new place without preparation: neither Graeme nor I ever even thought about how we'd find the hotel or the meeting venue beforehand, we just got up and did it.

Looking back, I'm not surprised we (the technology community) got the technology to work; I'm not even really surprised at the availability of mobile internet and a load of apps to make use of it; but I must confess to being slightly surprised that at the acceptance of all the gadgetry amongst the general population, enough to generate an ecosystem of companies who work together and create more value from their interoperability. It's something we always said would happen, but it's quite strange to see it in operation, and it's a positive achievement we shouldn't forget about.

Uploading sketches to a breadboard Arduino

It turns out there there are quite a few versions of the "same" components out there. Uploading sketches to an Arduino-on-a-breadboard is trickier than it first appears.

The Arduino-on-a-breadboard showed that we can get a lower power version of the same architecture. However, in doing the measurements I used a microcontroller already loaded with the code I used for the power measurements (sleeping and blinking). Trying to change this code and upload it via the USB breakout board didn't work -- repeatedly.

It turns out that the breadboard tutorial on the Arduino web site is actually flawed for the current versions of the components concerned. There are actually two problems: the microcontroller needs to be manually reset before uploading a sketch; and the USB breakout board needs slightly more supporting electronics to talk to the microcontroller.

The first problem stems from the microcontroller needing to be reset before code can be uploaded to it. Essentially the reset makes the bootloader wait for code for a few seconds, and start the existing program if none arrives. On older Arduino models you have to physically reset the board using the reset switch just before uploading a sketch; on newer models, this reset happens automatically. Setting up the breakout board to reset the microcontroller immediately before it tries to talk to it will solve this.

The second problem is more subtle. The USB breakout board is actually a USB to serial converter. The tutorial suggests that it is enough to connect the transmit and receive (Tx and Rx) lines to the microcontroller, but this turns out not to be the case: one also needs to connect some handshaking lines to make the system synchronise and communicate correctly. I eventually found a post that explains this: however, that post is flawed too, because it relies on a particular pin-out for the USB breakout board that's different tothe one I have. So here's a debugged explanation of what needs to happen.

FT232r breakout board

We need to connect the basic TxD, RxD, Vcc and Gnd lines on the breakout board as you'd expect. The picture to the right shows the the underside of my breakout board, with the pins named. If we number the pins counter-clockwise from the top left (so DCD is pin 1, TXD is 9, TXLED is 11, and VCC is 13), we connect pins 3 and 10 to ground, pin 13 to power, pin 9 to pin 2 of the ATMega microcontroller, and pin 5 to ATMega pin 3.

What now also need to happen is that we need to connect the CTS and DTR lines to something. DTR (Data Transfer Ready) is sent low when the USB has data ready: we want this to trigger a reset of at ATMega. We then need to send CTS (Clear To Send) low so that the board starts sending data. This is basic serial-port handshaking. The timing can be accomplished using an RC circuit consisting of a 100ohm resistor and a 100nF capacitor attached appropriately. Putting this circuit onto the breadboard sorts out the handshaking, and the Arduino IDE happily uploads sketches just as it would to a "real" Arduino.

The net result of this is to add some more wiring to the USB end of the Arduino breadboard:

USB end

Note the resistor and capacitor. (The red wire crossing the breakout board is a Gnd connection, needed because my breadboard only had single power rails top and bottom.) The circuit involved is as follows:

Handshaking circuit

For my particular breakout board shown above, this means connecting pin 7 to the capacitor and pin 15 to the following resistor. (It's this last step that the post gets wrong -- or at least uses a different pin for CTS.) The net result is an Arduino-on-a-breadboard that looks like this:

Second Arduino-on-a-breadboard

Somewhat more complicated, but rather more functional.

I think you have to maintain a sense of perspective about these issues, annoying as they are: in many ways it's good that the components change and evolve rather than staying exactly the same, as it means that they're being developed and refined over time. On the other hand, it means you have to be very circumspect about following blindly the tutorials and explanations on blog posts from even a relatively short time ago.

Low-power Arduino-on-a-breadboard

Putting an Arduino together from scratch lets us look at where the power consumption might be reduced -- and is just an interesting thing to do anyway.

One of the most exciting things about the Arduino is that it's open-source, so you can build them yourself -- and potentially vary the way they're put together for specific projects, which is very useful as a starting point for people (like me!) who aren't hardware engineers.

The main challenge for sensing with Arduinos seems to be their power consumption, and the obvious way to address this is to see whether there are things to be done to reduce the power drain, for example by addressing the issue of the quiescent current of the power regulator.

As a starting point, I used an on-line guide to build an Arduino on a breadboard:

An Arduino build from components

Actually this isn't a "full" Arduino as the analogue to digital converter (ADC) isn't properly set up, but it has the basic components of microcontroller (the same ATmega 328P as on an Arduino Uno), LED, reset switch, power, and USB. The breakout board at the left-hand side is the USB adapter, while the cluster of components on the right is the power regulator. At present I'm powering from batteries; one can also power from the USB, or from a wall power supply via another breakout board, but this way allows the same power measurement regime as earlier.

Measuring power for a simple "blink" program gives the following result:

Activity Power mode Current
Nothing Deep sleep 4.5mA
Flashing LED Awake 17mA

So in deep sleep mode the system draws about a seventh the power as a "real" Arduino. This is all down to the choice of voltage regulator: an L7805 with a design maximum quiescent current of 6mA. To put this into perspective, a system that could last a week on a standard Arduino board would last the best part of two months in this configuration. Put another way, we can build a sensor mote with an Arduino architecture and dramatically increased lifetime by changing a core component and using SleepySketch to keep the system asleep by default.

Ambient backscatter

An interesting article on how to power sensors and other "Internet of Things" devices.

A group at the University of Washington has developed a way of making use of "stray" radiation to power simple radio transmitters and receivers. Rather than use a dedicated power source, whether on-board like a battery or transmitted as in near-field communications, this technique makes use of the ambient radiation of cellphone signals, wifi networks and the like to provide enough power to energise a simple radio link.

Recycled Energy: Ambient Backscatter Allows Wireless Communications with no Batteries

If it works reliably, this'll be a huge contribution to low-power environmental sensing as well as to the applications the authors are targeting.

How to write an abstract

An abstract is an advertisement, not an introduction.

I've spent much of this week working with MSc students writing their dissertations, and this has inevitably led to the part of a dissertation that often causes the most pain to write (and read, for that matter): the abstract.

What is the abstract of a report or paper? What it isn't is an introduction or guide to the rest of the document: that, unsurprisingly, is what the introduction is for. The goal of an abstract is much simpler: it's intended to persuade the reader to read some or all of the rest of the document. It may be surprising that this is an issue, but in a world in the grip of an information explosion it's clear that readers' attention is a limiting quantity in information processing. As a reader, should you bother to read a paper? Or not? And how do you make this decision?

Many new researchers, when confronted with a paper, start at the front and work forward. More experienced researchers know this is a mistake that leads to reading whole tracts of irrelevance or nonsense. (The former is worse: nonsense at least has occasional entertainment value.) This leads many people (myself included) to adopt a non-linear reading style:

  1. Read the abstract
  2. If still interested, read the conclusion
  3. If still interested, read the introduction
  4. If it's really interesting, read the rest of the paper
(There are several variants, a couple of which are described here and here.) The point here is that we can filter out papers of lesser interest to reserve time and head space for the most interesting. Only if the results grab your attention do you need to spend time discovering the detail of how they were obtained -- rather than doing that work only discover that you don't care about the results being reported.

The abstract is the key: without a decent description of what the paper is about, the discerning reader will not proceed even to the introduction, and therefore any work contained in the paper will remain unread -- and therefore be largely worthless.

So how does one write a decent abstract? Most experienced scientists have their own technique, and the approach does need to vary from field to field and for different paper styles: a review paper is different to a piece of primary research. The approach I've come to rely on for research in computer science and mathematics can be described succinctly as the five-sentence abstract. I've found five sentences seems to be about optimal, structured as follows:

  • The area of the paper (1 sentence). The problem area to which this paper makes a contribution.
  • The issue the paper addresses (1 sentence). Presumably the area is not yet fully explored, and you've found a problem that needs tackling -- otherwise what's in your paper?
  • What you've done, the results you've obtained (2 sentences). The key contribution of the paper, what you've added to practice and/or knowledge
  • What this means (1 sentence). Why should anyone care?
Let's do an example, from a paper I've always wanted to write about finding unicorns:
There are lots of interesting animals out there, many of which have horns. No-one has yet reported observing any one-horned horses, however. We describe our research survey of the horses of the West of Ireland. While we found many horses, and many other horned animals, we failed to locate any horned horses. We conclude that further research is required to find unicorns, preferably in an equally pleasant holiday destination.
OK, perhaps not a great example. Let's try another, from a real paper:
In the domain of ubiquitous computing, the ability to identify the occurrence of situations is a core function of being context-aware. Given the uncertain nature of sensor information and inference rules, reasoning techniques that cater for uncertainty hold promise for enhancing the reasoning process. In our work, we apply the Dempster Shafer theory of evidence to infer situation occurrence with the minimal use of training data. We describe a set of evidential operations for sensor mass functions using context quality and evidence accumulation for continuous situation detection. We demonstrate how our approach enables situation inference with uncertain information using a case study based on a published smart home data set.
(Taken from McKeever, Ye, Coyle and Dobson. Using Dempster-Shafer theory of evidence for situation inference. In Proceedings of the 4th European Conference on Smart Sensing and Context (EuroSSC). Volume 5741 of LNCS. Springer-Verlag. Guildford, UK. 2009.) The abstract goes from domain to challenge to approach to significance: having read it, the reader hopefully has a fairly good idea of what the paper contributes to which domain, and why this contribution is significant (in the authors' minds, at least).

Shorter is often better, of course:

Wireless sensor networks are attracting increasing interest but suffer from severe challenges such as power constraints and low data reliability. Sensors are often energy-hungry and cannot operate over the long term, and the data they gather are frequently erroneous in complex ways. The two problems are linked, but existing work typically treats them independently: in this paper we consider both side-by-side, and propose a self-organising solution for model-based data collection that reduces errors and communications in a unified fashion.
(From Fang and Dobson. Unifying sensor fault detection with energy conservation. In Proceedings of the 7th International Workshop on Self-Organising Systems (IWSOS'13). Palma de Mallorca, ES. May 2013.) In this case my student Lei messed with the sentence structure because we wanted to get across the idea that the main problem was the totality of the existing approach to the problem, which we wanted to address as a whole. The abstract still follows the basic structure, and I think is stronger for being shorter.

No rule of writing is hard-and-fast, of course, and so you'll often find great abstracts that adopt a completely different approach. I don't think this matters so much as ensuring that the abstract is fit for purpose: an enticement to read further.