The view of the internet, 15 years ago

I was just sent a link to an article from 1995 on how the internet is over-hyped. It’s a fascinating read, not just in terms of the things it gets wrong but also of the ways in which the views expressed were plausible at the time. The article in question is “The Internet? Bah!” by Clifford Stoll, and appeared in Newsweek on 27 February 1995. For those whose memories of computer culture don’t stretch back this far, Stoll has form. He was a system manager at Lawrence Berkeley laboratory in California during a serious attempt to crack US military computers — one of the first examples of modern cyber-warfare. Rather than shut-out the crackers when he found them, he instead worked alongside a largely uncomprehending law enforcement community to help track them down, and brilliantly tells the story in his book The cuckoo’s egg. He then got concerned about the over-selling of computer technology for his next book, Silicon snake oil. His Newsweek article is in this latter vein. The crux of Stoll’s argument is that the internet will never replace traditional off-line activities like shopping for books, accessing a newspaper and the like. The internet is simply

…one big ocean of unedited data, without any pretense of completeness. Lacking editors, reviewers or critics, the Internet has become a wasteland of unfiltered data. You don’t know what to ignore and what’s worth reading.
It’s barely worth noting that many of these arguments have been invalidated by events. That’s hardly surprising, and while a technologist of Stoll’s standing should perhaps have been more wary about some of his predictions, the more important point is how the internet evolved to address points that, from a 1995 perspective, seem completely natural. Stoll’s comments on electronic publishing are perhaps the most interesting:
How about electronic publishing? Try reading a book on disc. At best, it’s an unpleasant chore: the myopic glow of a clunky computer replaces the friendly pages of a book. And you can’t tote that laptop to the beach. Yet Nicholas Negroponte, director of the MIT Media Lab, predicts that we’ll soon buy books and newspapers straight over the Intenet. Uh, sure.
And, of course, he’s right: who would want to read a book on a 1995 green-screen, or indeed on one of those then fairly new-fangled windowed displays? That’s only changing now, where displays have similar resolution to paper as far as the eye is concerned, and when e-paper displays can be read in direct sunlight — and when one can take an iPad or a Kindle to the beach, albeit rather carefully, and buy books not only straight over the internet but even completely untethered over the cellphone network. A similar argument can be made to take down the article’s discussions about e-shopping for airline tickets and restaurant reservations, e-government and access to information, and so forth. But the fact remains that Stoll’s analysis of the internet circa 1995 wasn’t too far off the mark. Where did things change? I suspect the clue is in the last paragraph:
What’s missing from this electronic wonderland? Human contact. … Computers and networks isolate us from one another.
Again, not an unreasonable view in 1995. No-one I can remember really suggested that social networks would flourish, and indeed come to almost define the web and internet in the early 21st century. And that’s rather surprising, given that the first “killer app” for the internet was e-mail, and not (as was expected) scientific data exchange: a social technology rapidly took off in a place where no such socialisation was expected. The surprise is that we were surprised again — and I include myself in that surprise — when the history of the internet clearly showed that it’s users see it as a social enabler as much as, if not more than, as an information source. Clearly we shouldn’t abandon the sorts of critical comments that Stoll was making, or worry that predictions about technology are almost always overtaken by events we had no idea were coming. But it does mean that whenever we hear comments on the social value of technology and the impact it will have on society — as is happening over internet reading and other technologies at the moment — we should pause and think whether the negatives identified are somehow intrinsic, or whether they rest solely on the systems as currently deployed and conceived. We’re familiar with the idea of a network effect. The strongest network effects are in the abilities of people to re-use and re-purpose technology beyond the bounds conceived of by its inventors. It’s only really surprising when this doesn’t happen.

The three academic stereotypes

You encounter a lot of different personalities in academia, but when you get right down to it they all seem to fall under three basic stereotypes. OK, I admit it’s a gross over-simplification, but here we go:

The young gun

Young guns are the change-makers of academia: the people who want to change anything that doesn’t work and replace it with something better — or at least something different that can be tried out and tested to see if it is better. This often makes them talented researchers (although not necessarily more so than stalwarts, the next stereotype), but they’re typically found in larger groups, taking on larger projects and collaborations. Young guns can be of any age. They tend to be young, of course — often junior staff who are keen and un-jaded, who want to move their discipline forward and teach it as well as it can be taught. They’re often found pushing for new modules, new degrees, new ways of teaching and assessment — and for promotion. But they don’t have to be calendar-years young: the gun-iest young gun I’ve ever met was in his late 40s when I first went to work for him, is now approaching retirement, and still has more ideas and energy than most people a third his age. These people feel young even when they aren’t. As is probably apparent, young guns can be hell to work with, since for them everything is potentially up for grabs. They’re often (but not always) better at idea-forming than at execution, and often (but not always) lack the long-term detail-orientation to make sure that their ideas work out. They are often (but not always) egotists, who may not recognise when a change they’ve become passionate about is wrong or not working. But without them there’s no-one to drive change forward and make sure that schools and disciplines stay fresh and relevant.

The stalwart

The majority of people in academia are stalwarts: people who have clear ideas about the things they want to do and how to do them, but are essentially positive about their activities. They might not lead change, but they’ll row-in alongside if they like it and will accommodate to anything that’s broadly agreed. This applies equally to teaching and research. The typical stalwart will teach modules carefully, perhaps changing the content and delivery only slowly if left to their own devices but being perfectly open to updating to include new ideas. In research they will typically be found slightly off the mainstream with a small group of students (often only one at a time), not going for large grants or big collaborations but being solid supervisors and contributors at a small-to-medium scale. Often they come up with great ideas, because they pursue a line of research solidly over a long period and so become world experts. The ideas might not be widely circulated, and so it’s easy to underestimate the sophistication stalwarts bring to their work. They may need prodding to publish appropriately, but they’ll then address tat problem as effectively as everything else they do. Stalwarts are essentially positive people, working within a well-defined comfort zone. They are the backbone of any research project or school, and need to be appreciated and rewarded appropriately. They also need to be listened to carefully, since they provide a stability and a sanity that young guns often lack, and will make sure that changes are properly thought through and executed upon.

The twisted nay-sayer

The first two stereotypes are basically positive, but it all goes down hill with the third: the twisted nay-sayer. (I can’t claim credit for the great name, incidentally, which is due to Paddy Nixon.) Twisted nay-sayers oppose all change, no matter of what kind and no matter how motivated, and will to continue to oppose changes even long after the decision has been made and the time for action has passed. On first acquaintance, a twisted nay-sayer often seems to be someone who’s stuck-in-the-mud, after an easy life, and not wanting to have the hassle of changing — a bit like a rather negative stalwart. But this is to overlook the twisted part, which will not only avoid change but actively scheme against it, or to reverse it afterwards. It’s this essential negativity that sets this type apart from others. They can, perhaps surprisingly, be excellent researchers, but they’ll also be constantly highlighting their successes to anyone who’ll listen, even when those successes are long in the past. When faced with a new field or innovation they’ll point out that it’s really just a poor re-discovery of something that was current years ago, or just an instance of some pet area of theirs that the new innovators should really have found out more about. The two things to remember about twisted nay-sayers are that they are egotists, and that they are made, not born — the young gun’s dark shadow. You make them by thwarting the expectations they have for their careers. This can happen in two ways: either their expectations were unreasonable, and reality has intruded; or their expectations were completely reasonable but were thwarted by circumstance, malice or indifference. A particularly common case is someone who’s been repeatedly passed over for a promotion they think (rightly or wrongly) that they deserve. The repeated denial of their aspirations eventually causes them to give up, turn their back on the future they can’t have — and then rail against fate and everything that comes afterwards. How, then, are we to deal with the different kinds of academics? Most people fall into some compromise between categories (stalwart with young-gun moments, for example), but clearly one needs to understand an individual’s primary motivations in order to know where they’re coming from. I think the trick is to make sure a school keeps, listens to — and occasionally reins-in — its young guns; recognises and rewards its stalwarts; and tries hard not to grow any twisted nay-sayers. All these activities are fraught with danger — especially in academia, where we lack most of the levers of control that normal organisations have. A school can’t typically promote on its own recognisance. Not promoting those who feel they deserve it risks overlooking a young gun and having them leave, or (worse) stay, but mutate into a twisted nay-sayer. On the other hand, many promotion boards over-value their young guns and ignore their stalwarts, who then feel under-recognised. Doing so can destroy the stability of a school and can lead fragmented research programmes, and to teaching being good at the edges but lacking a proper core. It’s also worth remembering that some people in senior academic positions are extremely conflict-averse and so will cave-in to pressure from twisted nay-sayers in the interests of consensus — not realising that consensus is neither possible nor desirable, and that acquiescing will only lead to more obstructionism, because it’s about the obstruction, not the particular issue at hand. It takes confidence to say yes to experimental change — but equally it takes confidence to say no when necessary.

Call for papers on managing federated networks

We are seeking papers on the issues involved in managing federations of systems, those that cross enterprise boundaries.

ManFED.COM 2011

1st IFIP/IEEE Workshop on Managing Federations and Cooperative Management
Co-located with IM 2011 in Dublin, Ireland.
May 23, 2011
Management approaches that can be applied across organisational boundaries are increasingly important in a wide range of application areas. These range from algorithmic approaches which adapt to the observed behaviour of third-party systems, based on game-theoretic approaches or other predictive models, to explicit organisational federations which adopt coherent solutions and management models to facilitate interoperability among multiple independent organisations. There are a number of significant, common, complex issues which must be addressed in all technologies and applications that involve federated organisations – how to enable secure governance in the absence of a single, central point of authority; how to achieve semantic interoperability in the absence of common schema; how to provide effective access control in the absence of common user and role models; how to provide analytics and support for effective decision making;  how to adapt to environments that can be highly dynamic as well as highly heterogeneous; how to construct and maintain a common inter-domain governance model in the presence of highly diverse local governance infrastructures . This workshop will, for the first time, bring together researchers from a broad array of application and technical areas who are concerned with cross-domain management. It will draw out common themes, problems and issues encountered, and the solutions being designed to deal with the problems of managing information systems that span autonomous domains. It will aim to provide the basis for a common understanding and common approaches to inter-domain management and governance that synthesises the insights and best of breed solutions being developed in the diverse areas in which these problems are encountered.
Cross Domain and Federated Management Issues in the following areas:
  • Governance mechanisms for federated environments, e.g. Policy Based Management
  • Collaborative management, algorithmic adaptation, game theoretic approaches, predictive modeling.
  • Modelling cross-domain relationships, i.e. Information models, formal specifications, languages for federation
  • Distributed trust management and federated security systems
  • Data federation
  • Semantic technologies, semantic mapping and linked data
  • Information security in federated environments
  • Cloud & grid computing management
  • Software engineering for federated systems, i.e. tool chains, design by contract, model driven engineering and design patterns
  • Model driven approaches for the generation of adaptive inter-domain relationships
  • Adaptive analytics for the support of managing complex, dynamic multi-domain solutions
Authors are invited to submit full papers (8 pages) describing the original work. All manuscripts must be written in English and should be prepared in IEEE style. All submitted papers will be reviewed by the ManFed.CoM Technical Program Committee. For the review, all the papers should be submitted in PDF format through the ManFed.Com page on the JEMS system (, filling every item and uploading the respective papers. The contributed papers, after being reviewed and accepted by ManFed.CoM referees, will be published in the Conference Proceedings that will be included in the IEEE Conference Publication of IM 2011 and will be available on IEEE Xplore. The papers will also be indexed and abstracted by several databases such as INSPEC, Engineering Index (EI), SCOPUS, Conference Proceedings Citation Index (CPCI), etc. Finally, the organizing committee are currently in discussions with the editors of leading network and service management journals regarding the publication of a special issue to include best submissions from the workshop.
  • Full Paper Submission: 15th December 2010 (midnight GMT)
  • Acceptance Notification: 30th January 2011 (midnight GMT)
  • Camera-Ready Manuscripts Due: 15th February 2011 (midnight GMT)
  • Kevin Feeney, TCD
  • Joel Fleck, HP
  • Rolf Stadler, KTH
  • Brendan Jennings, TSSG

Computer scientists as university principals

I was surprised to discover a computer scientist as principal of a world-leading university the other day. I was even more surprised to discover that there have been several others. Now I suppose I should first admit that I don’t know why I’m surprised by this — but I am. In most institutions computer science very much takes a back seat compared to other, more established Schools: physics, mathematics, biology, history and the like. It’s hard to see why this should be the case, especially given computer science’s central role in the new science and the fact that it’s often one of the highest-earning Schools in terms of research and innovation income (although St Andrews is unusual if not unique in having a humanity — international relations — as it’s highest-earning School, and from which we’ve drawn our current Principal). The person I came across was John Kemeny, who was president of Dartmouth College in New Hampshire through the 1970s. Kemeny is one of the co-inventors of the BASIC programming language, which provided me (and many others) with their first introduction to computer programming. BASIC was invented at Dartmouth, of course, so it’s perhaps unsurprising that he should have risen to such a position of influence. It’s hard to over-estimate how important BASIC was, in a world then populated by low-level assembly code or compilers of dubious quality: the first Pascal compiler I used used to spit-out assembly code so you could go through it, optimise and correct it by hand. As an interpreted language, BASIC provided a far simpler and more accessible introduction to what computers were capable of, and even on early 8-bit microcomputers was fast enough to be used for both serious applications and games. Having tweeted my surprise at this, I was then told about other computer scientists who’ve led — or indeed lead — universities:

  • Maria Klawe (algorithm design, accessibility), President of Harvey Mudd College
  • Tim O’Shea (computer-assisted learning), Principal of the University of Edinburgh
  • Ewan Page (mainframe pioneer), Vice-Chancellor of the University of Reading
  • John Hennessy (processor design), Provost and then President of Stanford University
  • Jane Grimson (databases, health informatics), Vice-Provost of Trinity College Dublin
  • Jeff Vitter (algorithm design), President of the University of Kansas
And of course an honourary mention for:
  • Paddy Nixon (pervasive computing), Vice-Principal for Research, University of Tasmania
Before anybody asks, this is not a tradition I have the slightest interest or intention of following in — or indeed the ability to do so. But it’s great to see that techies can and do aspire to the top job.

The first year

Today marks the first anniversary of my moving to St Andrews. What have I learned since then? It’s a strange feel to be reflecting on a year of my life, not least because it doesn’t feel like a year. Sometimes it feels like a lot less: I still feel very attached to Ireland, and I find that I spend a lot of time comparing the Scottish experience to y experiences of the previous twelve years. But in other ways it feels a lot longer than a year, in that I think I’ve found a professional home in St Andrews that’s exceptionally well-suited to my way of researching and teaching. So what are the differences? There are several things that spring out. Firstly there’s the size of the place — or lack of it. The university has around 6,000 undergraduate students and maybe 2,000 graduate students, so it’s significantly smaller than UCD — about the same size as Trinity College Dublin, I suppose. This has a corresponding impact on class sizes, where a 30—40 student second-year class, and maybe 10—15 (or less) in third and fourth years, is considered perfectly normal. That in turn leads to a more individual and interactive style of teaching. The second impact of size is in research, and especially in multi-disciplinary research. St Andrews is so small that one can know everyone (or at least anyone you want to), and can find and gain access to people doing research in whatever topics there are in the university that are of interest. In the past year I’ve interacted with marine scientists, astronomers, mathematicians, geographers, psychologists, medics and others, on a basis that will probably lead to some sort of proposal for collaboration or funding. One can do that in any full-spectrum university, of course, but size does make a difference: the intimacy of St Andrews, the fact that nowhere is more than a fifteen-minute walk away, makes it so much easier to interact. In a larger institution, and one with a larger computer science contingent, there’s a strong tendency to remain within a smaller comfort zone that’s not conducive to multi-disciplinary collaborations. I think we’ll be able to leverage our smallness. Thirdly, St Andrews takes teaching way more seriously than any other university I’ve worked in (or studied in, for that matter). All universities claim that teaching is a core part of their mission, of course, but it often doesn’t get treated with the same urgency or seriousness as research. That’s in part a function of how we’re evaluated: both individual staff promotions and the most popular global university rankings are heavily biased towards research excellence, and staff and managers inevitably respond to those incentives. But I suspect it’s more than that. St Andrews’ processes are very focused on teaching, as is the academic culture, in a way that’d uncommon in my experience (which is limited to the UK and Ireland, of course). I do all my own teaching, all my own marking, and participate in small-group tutorials both for and beyond the modules I teach myself. Moreover the processes of assessment, tracking and evaluation of students’ progress have tool support and are monitored from both the School and the centre. Other universities I’ve worked in don’t have this degree of monitoring — or indeed any monitoring. That doesn’t mean that teaching isn’t done well in those places, of course, but it does indicate where a university’s priorities lie. (In case this sounds like altruism, it isn’t: the National Student Survey results pull St Andrews up the rankings that take account of student satisfaction, and mean that the university can legitimately lay claim to offering an excellent experience to prospective students. It’s a good example of the university taking a broader and longer-term view than other institutions.) Following on from this, Scottish universities teach degrees with a “broad curriculum” in which students take a rather general two-year sub-honours programme before specialising into a further two-year honours programme. This provides  a broader base for students and avoids too-early specialisation, which I think is a good idea. I hadn’t quite appreciated what a difference it makes in practice until earlier this week, when I did a lecture on the history of the internet and what its evolution means for society in terms of publishing, privacy, trust and access to information. I’ve done technical lectures like this to computer science students before — I’ve used the evolution of the internet as a case study of large-scale systems design for software engineering students, for example — but this was an introductory for first-year students from across the university, both sciences and humanities. That’s not something that happens very often these days: in fact, in twelve years as an academic I’ve never lectured a broad-ranging class like that before. There’s something rather exciting about being able to address broader questions of technology’s impact on society, and to set essay-style questions, when one is used to the more technical style of scientific lecturing. Not only does it allow a more far-reaching and questioning style of teaching, and the associated invitation to oneself to think through the broader questions: it also feels like it might have an impact, however small and however subtle, on a wide range of students who’ll perhaps never encounter computer science again — but who will undoubtedly be affected by it profoundly as part of its impact on society. It was particularly nice to bring the recent discussions about the internet’s effects on learning to their attention, as well as to talk about sensing and its effects on privacy. Are there any things not to like? The lack of a senior common room is a little peculiar in a university of this age, only partially made-up for by a staff dining club that meets only infrequently. There are actually very few social occasions for staff across the university, which is a shame given the effort put in to student societies and the student experience: some more activities targeted at academics and researchers would be welcome. St Andrews itself is also somewhat remote from the rest of Scotland, more so than I first thought from looking on the map, but we’ve addressed this by moving to Edinburgh, which has everything one could want from a culturally vibrant city. I’m looking forward to what the next year may bring.