The productivity paradox is the two apparently surprising facts that: productivity rates in the US have flattened out considerably since 1973, and this is the time that industry has invested in IT.

It would appear that investment in IT has not caused any particular improvement in productivity, indeed it may have reduced productivity.

Landauer argues that IT systems that have to interact with users, or which are intended to support users in their jobs (not completely replace them) are badly designed. They do not take their users into account well enough, hence users have difficulty using them efficiently and this is detrimental to productivity.

It has been argued that productivity is an inadequate means of measuring success in information industries, because information industries generate intangible products that are difficult quantify. Furthermore if new IT investment is coupled with new management structures that make the best of the IT systems then productivity improvements are apparent.

A well designed IT system is more likely to

Because of this users and their employers are much more likely to get a payback dues to tasks performed well and efficiently using the IT system. Badly designed systems are cheaper to develop and buy in the short term, but do not give the benefits described above.

Safety engineers work to reduce the impact of accidents by:

If the severity of a possible accident cannot be significantly reduced then its likelihood must be significantly reduced and vice versa. However, because the consequences of an accident for many safety critical system cannot be accurately predicted, work to reduce to likelihood of accidents is preferable.

The failure of a normal interactive system can be irritating and time consuming, but typically causes no lasting damage. If safety critical systems fail the consequences can be severe and permanent.

The failure of normal interactive systems only effects the user and possibly the user's employer. The failure of a safety critical system can affect a great number of players, external to and not complicit in the working of the system.

Free market maxims like ‘buyer beware' do not often apply to computer software, because it is often not the case that the buyer is the one that has to beware of the product being bought. Users of bad products usually have considerable leverage over developers by simply refusing to buy their products any more. In the software market, particularly for commercial products, the people who buy products are not the same people as use them, therefore the usual economic channel of feedback is blocked.

My viewpoint of the web is as follows: ‘The web is a vast array of information and with some effort the chances are that I'll be able to find some information about any topic I'd care to name. I am, however, suspicious of the accuracy of that information and use the web more as a tool to point to valid information, rather than as the source of that information. I find it enormously frustrating trying to find information on the web and rarely search around the web as whole, instead I have a collection of pages in my bookmarks that point to interesting places and 90% of my time is spent in those pages or one or two links away from them.

'The content of the web is claimed to be global and universal, but I am suspicious of that claim, I find the information there very culturally-centric around the North American way of life and viewpoint. Also the demographic of the web is largely affluent, white and male and the content of the web reflects this. I find many of the places on the web where interaction takes place with others to be unpleasant; chat-rooms, Usenet groups and mailing lists have an unfortunate tendency to be aggressive and pedantic, qualities I find unhealthy. I tend not to hang around long anywhere I encounter that attitude.'

‘I have two general strategies for finding things. The first one I call searching, where I know what I'm looking for, and I'll know when I've found it, I just don't know where it is precisely. My strategy in this case is to start off on a page from my bookmarks that looks as though it may be close to the desired page and then follow likely looking links. I'll do this for about ten minutes or so. If I haven't found the page by then, I'll resort to a search engine, type in search terms that are relevant and move on from there. If, after another ten minutes, I still haven't found the page I'll either give up or try and find a colleague to help.

The second strategy I call browsing. Typically I have a vague idea about what I want to find and no idea about where to find it. I'll usually start from a search engine and enter a few terms, read through the results and then alter the search terms if they don't seem to be generating the sort of results I'm after. I'll then start investigating the pages suggested by the search engine. I'll follow a link and then explore that page, following links from there to interesting look pages, or backtracking to the search engine results and then repeating the procedure from other likely looking pages. At times during this process I may notice links to pages that are interesting, but nothing to do with the original topic I was investigating. I may then completely change the goal of my browsing and search after that topic instead, or I may bookmark that page and come back to it later.'

'The content of the web is claimed to be global and universal, but I am suspicious of that claim, I find the information there very culturally-centric around the North American way of life and viewpoint. Also the demographic of the web is largely affluent, white and male and the content of the web reflects this. I find many of the places on the web where interaction takes place with others to be unpleasant; chat-rooms, Usenet groups and mailing lists have an unfortunate tendency to be aggressive and pedantic, qualities I find unhealthy. I tend not to hang around long anywhere I encounter that attitude.'

The following may be responsible:

The point is not that anyone in particular is responsible, but that the consequences of the failure of safety critical systems are widespread. In some cases it cannot actually be predicted who will suffer the consequences of system failure and therefore it is essential to reduce the possibility of the accidents happening in the first place.