Old and faulty

A podcast discussing old and faulty software and its implications.  I think the fears here are rather overblown - comparable to those which surrounded the Y2K bug - and of course includes the lame but obligatory "digital pearl harbour" quote (how many times have we heard that before).

It's claimed that there are zillions of lines of COBOL code (or some large number) out there running most of the infrastructure of modern life, and that as the knowledge of how to maintain these systems disappears from the job market due to technological change and the ravages or mortality, this could result in some great calamity (doom perhaps?).  I think this is unlikely.  Most businesses - especially large ones - know all about the problems of obsolescence, particularly of software and electronics, and businesses which don't plan for obsolescence are not really viable on any medium to long term time scale.

It may be the case that financial industries are not traditionally accustomed to planning for technological obsolescence.  It's worth remembering that the conversion of accounting from paper to electronic formats largely took place only within the previous three or four decades.  In a paper based world you don't need to think about these issues, and technology was limited to things such as typewriters which remained quite static over long periods of time.  Paper archives also didn't need to be periodically translated into different languages, and methods of processing paper, in terms of filing cabinets and indexes, probably also didn't change much from one decade to the next.

Large amounts of old code can sometimes also be replaced by much smaller amounts of newer code, since as time passes much of the low level details get abstracted away by use of libraries or higher level languages.  I actually have a qualification in COBOL programming, and can confirm that it is indeed a diabolical carbuncle, but probably made some sense in the 1960s and 70s.  As I remember much of COBOL was heavily oriented around a very particular sort of printer technology, which makes little sense in today's world especially if not much actually gets printed out.



So I don't think it's worth losing any sleep over the collapse of COBOLitarianism, and it's demise should be cause for celebration rather than mourning.  Most large business owners don't want to see their enterprises destroyed through negligence, and as the need to replace old back end systems arises they will be replaced, with little or no apparent change as far as the customer is concerned.

Wordpress postponed

For the present I've abandoned my experiment in wordpress blogging.  It was a valiant attempt, and I still might end up moving to wordpress at some point, but for the present trying to import from Blogger to wordpress in order to keep two blogs synchronised over a trial period of six months doesn't seem to be a viable option.  Currently the Blogger import plugin on wordpress 3.0.1, which was previously a one or two click operation, appears to be completely broken.

Trying wordpress is part of my overall "out of the cloud" strategy.  The thinking behind this is that if I have the facility to host my own content I might as well try to, and based upon past experience at the end of the dot com bubble I'm keenly aware that in straitened economic times free (as in beer) web services can change their terms and conditions, hastily erect paywalls or disappear overnight without warning.  If the worst came to the worst Google could say "You guys in the long tail, we're not making much money out of you - especially if you don't use Adsense - and we have some other fancy service we could use the disk space for.  Bye bye".  I don't expect that will happen, but it's always a good idea to have a fall-back option.

The third age of cybercrime

I never developed software directly for Siemens PLCs, although on one occasion I had to interface with them and communicate within their data block framework, and for another industrial application I considered using them but eventually chose a different sort of motion controller which was better suited to the job. I did write software on various other PLCs though, and the level of security on these devices was always either very cursory or non existent. Lack of security on such devices shouldn't come as any great shock though, because at least in the application domains I dealt with industrial control systems were only very rarely connected to SCADA systems or the general administrative IT system of a factory. In some cases security meant using a key to turn a dial to a particular number, so you would need physical access to the factory, control room and control cabinet or machinery together with quite specialised knowledge about what to do with the particular electronics. Factory level security - gates and security guards at entrances - was always highly variable, ranging from ultra strict to just being able to walk directly into a factory without being questioned by anyone, but security on control rooms and cabinets was usually adequate.

So if Stuxnet was specifically targeting Siemens PLCs this probably narrows down the possible culprits to control systems engineers. Those devices are fairly complex and expensive. It takes a non-trivial amount of time to learn how they operate (it's nothing like web or database programming on a Windows PC or a mobile phone), and it's not really the sort of thing which even an ardent technophile such as myself would hack on in their spare time outside of a work environment.

The Guardian article, which is a bit more in depth than others I've read on this topic, downplays the possibility of a financial motive, but a proposition such as "pay us $X, or else we shut your factory down" could be quite lucrative for organised crime groups, who might also have the finances to be able to recruit and train a specialist control systems engineer. Nationalistic/tribalist malevolence is also a possibility too, and nation states would certainly have the means and motive to generate this sort of malware.

Probably the moral of the tale is don't run mission critical applications, such as SCADA systems, on Windows, although I realise that currently the alternatives may be few and far between (maybe a gap in the market). Windows was designed primarily for home computer users, and wasn't created from the outset to be a secure networked system. Also in a factory environment it's probably a good idea to have some policy with regard to use of USB pen drives, and maybe it's wise to disable any unused USB ports in the BIOS on computers running critical applications.

Downsizing the state

I think that if the economy remains stagnant then 2011 could turn out to be a turbulent year in political terms. Presently we're in a kind of calm before the storm period. VAT hasn't yet gone up to 20%, which will inflate the cost of nearly everything, and the public sector employee layoffs havn't yet begun.

So how many people are going to become unemployed within the next year or so? It's difficult to come up with a precise quantity, but it's possible to do a back-of-an-envelope type of calculation. If the state is downsized by 25%, as indicated in this BBC article, and if it currently has around six million employees according to the Office of National Statistics, then if the number of jobs lost is in a similar proportion this works out at about 1.5 million. This is close to the current official claimant count for the unemployed for August 2010 (1.47 million), which means that potentially over the course of 2011 unemployment could double if the economic outlook doesn't improve.

1.5 million is only a very rough figure, but it gives some idea of the order of magnitude of the problem. Conservatively assuming that only a million of these newly unemployed claim state benefits, and that they're only able to get a miserly £50 per week to live on that adds £2600,000,000 per year of additional borrowing to the government's deficit. Will they be able to afford it, or will they resort to printing money again, instigating an extra round of "quantitative easing" or currency devaluation?

And apart from the numbers there's also the human element. I expect that many of those about to join the dole queue will be quite well educated middle class professionals who constitute the bulk of civil society. People like teachers, managers, administrators, councillors, social services staff, police officers and maybe some medical professionals (although it was claimed at the last election that NHS spending would be "ring fenced" to protect it from cuts). These people have a long way to fall in terms of living standards, and may not choose to slide quietly into poverty. They're also well connected in the classic social networking sense, and may be somewhat politically active.

So all the above looks like a recipe for a decidedly dicey situation. I wouldn't rule out the possibility of social unrest or rioting, similar to the poll tax riots of the late 1980s. In the best possible scenario the economy continues to recover and there is no particular problem in absorbing the consequences of a reduced public sector, with growing private sector business picking up the slack. Whether that will happen remains to be seen, but based on my own job searching the employment situation still looks extremely bleak with no obvious signs of improvement in the immediate future.

Festo arm

This is an interesting approach to the design of a robot arm, but I don't think that it's entirely novel. I've read of other designs from previous decades which used a similar sort of idea with many joints arranged in a snake-like fashion.



There are a few reasons why designs like this havn't been successful in the past. The main one is that the control of an arm with this design, in terms of the forward and inverse kinematics, is going to be significantly more difficult than for conventional robot arm designs. The fact that it's pneumatically driven also adds to the complexity of the control problem, since the rate of expansion of the various chambers may not be known precisely and may change over time as the system ages. Also, as I've experienced for myself, pneumatic actuators can be very noisy, to an extent where the noise problem alone might exclude many possible applications outside of a controlled factory environment. To really be practical the compressor needs to be small, capable of being run from a battery, and make as little noise as possible.

However, if the above problems can be adequately solved, then perhaps this sort of design will become much more popular. The all-plastic design has the advantage of minimising weight, which means that the arm can be longer than would otherwise be possible, and the compliant nature of combined plastics and pneumatics should mean that this design will be fairly safe to use in close proximity to people. Also a plastic arm is likely to have a much lower cost of manufacture than one traditionally made from metal, especially if it can be printed with something like a rapid prototyper.

An introduction to anthropology

Here is the first of a series of presentations by Alan Macfarlane on the topic of anthropology. I got interested in this topic due to the connection with Noble Ape and thinking about trying to simulate tribal cultures. Also it's difficult to understand the peculiar features of human cognition without also learning something about the environmental and social pressures which generated it.



If you believe that technology will eventually largely eliminate material scarcity it's interesting to note that some of the primitive tribal societies were living in a post-scarcity type of environment, below the carrying capacity of the environment. The future to some extent may come to resemble the past, with a greater emphasis upon reciprocity, gifting and individualistic status seeking. Post-scarcity is already well on the way to being achieved within the realms of software and music.

Is Google Instant a Spivey experiment?

Google have added a feature which allows search results to be returned as you're typing a query. I'm not sure that this is immensely useful, although it's an impressive technology demo showcasing how fast their system runs.

One thing which occurred to me is the question of what, if any, additional information this sends to Google and how they might try to monetize it (attempting to adopt the mindset of someone trying to sell advertising). In this instance each keystroke probably is invoking a query, or maybe the system just polls the query box every few hundred milliseconds - probably the keystroke event would be more efficient in terms of coding and processor usage than continual polling. This could convey - directly or indirectly - time stamped information about when keys are being hit, in addition to the partially typed search query itself. I think something similar occurred with the now defunct Google Wave, and possibly they might have reused the same code.

Does the temporal sequence of keypresses convey any information about the user or their intent? According to a company called Digital Proctor, it does. I'm not sufficiently expert in this area of psychology to know whether Digital Proctor represents real innovation, or is merely snake oil/pseudoscience, as many psychometric type tests are, but the notion that you might be able to identify someone from their typing style doesn't seem completely unreasonable. Both handwriting and speech are really just motor patterns, and those are relatively unique, so typing patterns might be also.

Here's the elevator pitch:



If their claim is true then possibly Google may be able to identify users, even if they're not logged in, via their typing style. A more likely scenario is that typing style is not completely unambiguous but can be used along with other weak classifiers such as browser info to make a strong one.

But even if typing style only provides such a weak signal about the identity of the user that it's not worth considering, other information may still be latent in the semantic-temporal pattern. Are there systematic relationships between the semantic content of the query, or the intent which the user has, and their physical movements? According to the psychologist Michael Spivey, there might be.



Spivey describes his ideas in a book called The Continuity of Mind. His basic idea is that there are multiple cognitive processes going on in parallel which are in some sense vying for control of your motor output, and that this dynamic struggle is to some degree apparent in your movements. In his view the decision of which action to take isn't the final output of a sausage machine but is created and modulated in flight by multiple other processes occurring simultaneously. This is quite consistent with a Brooksian notion of cognition as multiple parallel streams rather than a strictly sequential or "club sandwich" cognitive architecture.

So maybe with enough mining of a large semantic-temporal database it might be possible to extract any systematic correlations between search queries or particular words and temporal typing patterns, and this could be saleable to advertisers or useful for displaying contextual adverts. Whether Google realize this or not I don't know. Maybe they just wanted to show off the speed of their search engine, and hadn't considered that keystroke sequences might contain additional information.

Knuth in action

The remote controlled economy

An article written by Marvin Minsky in 1980 called Telepresence: a manifesto shows how forward thinking he was about telerobotics, or perhaps how slowly technology has progressed. At that time computing and communications technology were still in a relatively primitive state - for example, home computing was only just beginning - and I'd say that it's only really within the previous five years that telerobotics has become a really practical prospect.

For telerobotics you really need:

• Video cameras. It was really only from the late 1990s onwards that capture and transmission if images digitally became something which could be done cost effectively with consumer hardware, and only by the mid 2000s that enough internet bandwidth existed to do this well (broadband) without a lot of jerkyness. The transition of camera technology from analog to digital within the previous decade has been a largely unnoticed revolution.

• The internet, or some comparable system for high speed bidirectional communication capable of transmitting streaming video. Even though the internet existed in some form in 1980, computing speed and bandwidth would have been far below the minimum requirements.

• Wireless communications. It took at least another decade for wireless telecommunications to become possible via mobile phones, and another 10-15 years on top of that for the infrastructure to become ubiquitous and have enough bandwidth to support video.

• Small low cost computers which can act as telepresence relays. Embedded computer hardware fast enough to stream video and audio has only come into the price range of a few hundred dollars within the last five years.

• Low enough communications cost. It seems like a distant memory now, but 5-10 years ago being connected to the internet or making extensive use of mobile phones could be quite a pricey business. It was only around 2000 that flat rate internet subscription was introduced, prior to which communication would be paid for by the minute and costs could quickly mount up. Over a decade ago not many people could afford to be connected to the internet full time, without being cut off by their ISP or running up massive telephone bills.

Minsky identifies some of the main areas where telerobotics will be useful, and he also talks about the "men in suits" problem with regard to space based construction. The development of the International Space Station within the last decade has demonstrated beyond any reasonable doubt that having (mostly) men in space suits bolting things together is not a scalable way to carry out construction work in space. To reduce risk and cost and apply economies of scale it would be far better to use telerobots operated from the ground, with the construction sites being in geostationary orbits. If the robots were semi-autonomous it might be possible for one operator to work on more than one construction task at a time, and the work could be done by anyone with minimal training, as opposed to the extensive training which astronauts must currently go through.

But the bulk of the telerobot revolution won't take place in space or even in nuclear power stations (which presumably was a fashionable topic in 1980) but instead in the more mundane settings of homes, office spaces, streets, hospitals and ground based construction sites where there is far more potential for economic and social transformations to occur. Some of this could be controversial, since the ingrained tribalism of human psychology is often opposed to "foreign" workers and I'd guess that the legal and taxation frameworks of most countries are not geared towards a situation where potentially anyone can telerobotically do physical work at any location within a few seconds of logging in, although there may be benefits in terms of a much reduced need for migration and essentially perfect automatically policed immigration control. Telerobotic working also enables an expanded workforce which includes people who previously would have been considered to be economically inactive, such as the disabled and elderly.

Also telerobotics makes no assumptions about the progress, or otherwise, of research in AI. Even if you take a very pessimistic view in which AI remains a near intractable problem taking decades or centuries to solve with no major breakthroughs along the way this doesn't prevent the rise of a telerobotics industry in the near future. Many of the services for which telerobotics is a facilitator may also be fundamentally human-to-human in nature, such as health care or holidaying, so this isn't necessarily a robot takeover situation even if AI does make significant progress. For many such scenarios you can just think of the robot as another kind of communications channel.

Open core?

It will be interesting to see what happens when an app store selling "commercial" (i.e. proprietary) software appears within future versions of Ubuntu. I'm not really ideologically opposed to proprietary software - I've written plenty of it in the past - but closed source methodology has various pragmatic disadvantages and typically a low bus factor.

There has been proprietary software in Ubuntu for years, in the form of various drivers and codecs, but so far no explicit closed source applications are installed by default or are overtly advertised as far as I'm aware (unless Ubuntu One counts). One possible problem if big name software companies show up to display their warez within the app store is that the OS could begin to suffer from the same issues as occur with "open core" systems - assuming that Canonical takes a cut from each app downloaded from the store, as Apple do with theirs. That is, it could be considered financially advantageous to ship low quality or broken default free applications with the distro, in order to make the purchasable proprietary equivalents a more enticing choice. If I was a conspiracy theorist I might contend that's why they ditched the Gimp if they believed that a version of Photoshop might feature in a future app store, but this would of course be a completely scurrilous speculation.

Also, merely prominently displaying proprietary apps in front of the free software within the store could itself be contentious. In my crystal ball I can foresee much wailing, flame wars and gnashing of YouTube videos if the introduction of an app store isn't managed with due care.

Probably the path of least resistance would be to stick to proprietary games within the store, at least initially. Although there are open source games - some of them quite good - for the most part people don't expect to have the four freedoms with respect to games, which can be considered to be more artworks than utilitarian by nature.

Augmented life

Eventually people will no longer think of this sort of thing as "augmented reality" and just think of it as ordinary life. The technology needed to do all of the things shown in this video exists today. What's missing are commercially available Eyetap-like spectacles, which could be wirelessly connected to a phone.

Augmented City 3D from Keiichi Matsuda on Vimeo.