Broadband: how wide should the pipe be?

Broadband, digitisation, digital transformation, big data, it has been a long process, mainly driven by technical innovation. Still. Back in the day, in 2001 working for The Fantastic Corporation, I wrote this Spoof article on digital fascination and how it is driving business descisions. Well, in all honesty, I did not know it then, my focus was on using the scientific method as a tool (which it just is), so the article was meant to illustrate how a fool with a tool remains a fool (you can freely exchange tool for ‘new technology’ or ‘technology focus’, or….). Anyhow, here it is. Enjoy.

Broadband: how wide should the pipe be?

Investigating end-user constraints on big-data accessibility

Broadband, the next generation Internet is at our doorstep. Technical advances have created fast connections, so called pipes, with even the most remote location possible. Anywhere, you will have the information highway at your fingertips. But what will this new technology bring us? Will it improve the end-user experience? This paper describes an experiment investigating the influence technical aspects of broadband such as pipe dimensions on the end-user experience.


Broadband is maturing from an experimental stage and becoming within reach of every consumer. Fibre technology is dominating the market, allowing for data delivery with near speed of light. Over the last years network operators have invested a lot in high- speed glass-fibre technology, even more than there is demand for. As a consequence the prices for high-speed connections have dropped and have become affordable for every consumer [1]. Broadband is truly getting out of the labs and into the market place. The type of connection determines the speed in which you can send and receive data. The connection is commonly referred to as ‘pipe’. In particular, the dimension of the pipe determines the amount of data you have access to per time interval. There are two factors of the pipe influencing the transmission rate (see Figure 1), its lengths and its diameter.

Figure 1: an illustration of the factors pipe length and pipe width. Either a shorter pipe or a large diameter reduces transmission time. Length T results in larger transmission times. Pipe width D, due to the possibility of parallel transmission, results in reduced transmission times.

Figure 1: an illustration of the factors pipe length and pipe width. Either a shorter pipe or a large diameter reduces transmission time. Length T results in larger transmission times. Pipe width D, due to the possibility of parallel transmission, results in reduced transmission times.

The length of the pipe influences the access speed. The longer the distance, the more time it takes to receive or distribute data. Already running at the speed of light, the distance may be the least important factor. The diameter of the pipe also influences the access speed. A large diameter allows for parallel transmission of data, increasing the amounts of data per time unit received, whereas a small diameter allows for mostly sequential data transmission. Obviously, compared to parallel transmission the sequential data transmission results in a lower connection.

The end-user experience

The influence of the pipe dimensions on the end-user experience is still unknown. Not surprisingly, technical advances are driving the broadband market, ignoring the importance of information about the end-user experience. Contrary to technical aspects, the consequence for the end-user experience of these parameters never has been investigated. However, unless end-users experience benefits from technical advantages, all investments in these systems may be in vain. Reports from the market, such as [1], suggests the need for an alternative approach: using the end-user needs as basis for broadband rollout and development. This means answering questions like ‘What is the influence of the width?’, and ‘What is the effect of pipe length on the consumption of a so- called ‘compelling’ site?’ ‘Does the amount of joy the end-user experiences depend on the speed with which the data is consumed?’ This paper aims to systematically investigate and by doing so give insight into the effect of pipe dimensions on the end-user experience.

Experimental set-up

This paragraph describes the prototypes used and the experimental set-up and the task the users had to perform.

Figure 2: The largest (left, 80% percentile) and smallest (right, 20% percentile) of pipes normally available in the environment of consumers. These dimensions were used for the prototypes.

Figure 2: The largest (left, 80% percentile) and smallest (right, 20% percentile) of pipes normally available in the environment of consumers. These dimensions were used for the prototypes.

Apparatus used

As the focus for broadband is the consumer market, the sizes of pipes used for the experiment reflect those that commonly are available for house appliances. Figure 2 shows the two extremes (taken are the 20 percentile and 80 percentile of dimensions found). A prototype to control the information pick-up by end users was constructed. This prototype will be referred to as ‘GLobal Information SenSitive Experience transducerS (from here on referred to as GLASSES-TM). Figure 3 shows the prototypes used during the experiment. Various versions of the GLASSES-TM were created, to match the experimental conditions.

Figure 3: The four prototypes of the GLASSESTM, developed for the conditions of the experiment. Each prototype matches a specific combination of pipe width and pipe length.

Figure 3: The four prototypes of the GLASSESTM, developed for the conditions of the experiment. Each prototype matches a specific combination of pipe width and pipe length.

Experimental conditions

Table 1 shows the prototypes and the experimental condition for which it was used.


Subjects were presented twelve different tasks, and asked to grade the page on a seven point scale running from highly interesting to extremely dull. Subjects were given as much time as they considered required.

Dependent and independent parameters

As stimuli a twelve web pages were selected ranging from full multi-media experience to a page with text only. The stimuli were presented in random order. Six subjects participated in the experiment. All subjects were member of the Fantastic Corporation R&D department. After making participation to the experiment part of their MBO1, subjects were found to participate willingly and voluntarily. An aside, using the MBO to motivate the participation in such experiments might be the only useful application for the still controversial MBO [2]. Stimuli and conditions were randomised for each subject. Measured were the grading of the web pages and the amount of time used viewing the page.

Table 1: An overview of the experimental conditions and the GLASSES-TM’s used for each of the four conditions.

Table 1: An overview of the experimental conditions and the GLASSES-TM’s used for each of the four conditions.


This section gives the results of the experiment. Given the novelty of the broadband media, the data was recorded using Reverse Engineering techniques2 and analysed using Bistromatcis3 [3, 4].

Interaction: grading and pipe dimensions

Figure 4: results of the grading of web pages for the pages with multimedia content (left) and for the text only pages (right)

Figure 4: results of the grading of web pages for the pages with multimedia content (left) and for the text only pages (right)

Pipe dimensions were found to have most effect with multimedia content (see Figure 4). For text only pages only a small effect was found if the differences between conditions are extreme (a small and long pipe verses a short and wide one). For the grading of the pages, the best condition was found to be a short and wide pipe.

Conclusions and further research

For multimedia applications the dimensions of the pipe preferably is short and wide. For non-multimedia applications the benefits of a short but wide pipe are less clear.
During the experiment it was noticed that with narrow pipes end-users start to compensate the reduced information flow by active exploration, by making head- movements. It is known that information pick-up closely relates to possibilities of action [5]. Based on our observation it is believed that there is room for improving network and web technology by linking the behaviour of the user to the data transport. Further research will focus on how active exploration of the end-user can compensate for the dimensions of the pipe.
Also, additional applications of GLASSESTM are being developed. Especially promising are preliminary experiments conducted with additional filters GLASSESTM with a short but wide pipe. Depending on the end-user, the same information has to be presented in a different format to enhance it application and consumption. For the creation of these filters an Extremely clear yet Motionless Liquid (referred to as XML) was found, and successfully applied. Figure 5 – left shows a first prototype using a short and wide pipe encapsulating such a filter.
Further investigations will also focus on data filtering by using semi transparent reception, the so-called Partially OccLuding trAnsmission pROtocolair Information Device (also known as POLAROIDTM, see Figure 5 – right). However, as the results of the experiment show, to reduce data access, one could also use a longer and/or narrower pipe.

Figure 5: Two examples of possible applications of GLASSESTM. Left, a subject wearing a preliminary prototype incorporating a personalized XML filter for optimal information pick-up and right, subject wearing a prototype incorporating POLAROIDTM based semi transparent filtering techniques.

Figure 5: Two examples of possible applications of GLASSESTM. Left, a subject wearing a preliminary prototype incorporating a personalized XML filter for optimal information pick-up and right, subject wearing a prototype incorporating POLAROIDTM based semi transparent filtering techniques.

In sum, despite its slow roll-out [e.g. 1] broadband shows potential, and may have an extremely interesting and promising future, provided technical constraints are investigated and developed taking into consideration end- user needs and constraints.


[1] Karlin Lillington (2001) A ninety billion dollar mistake In: The Guardian. August 23, 2001
[2] Milkovich, G.T. & Wigdor, A.K. (1992) Pay for performance, National Academy of Sciences, Washington DC.
[3] D. N. Adams. (1982) Life, the Universe and Everything. Pan Books
[4] D. N. Adams (1986) Mostly harmless (the fifth part of the hitch hiker’s guide trilogy). Pan Books
[5] Gibson, J.J. (1979) An ecological approach to visual perception. Lawrence Earlbaum Associates, London. [Reprinted in 1986].

You can download the original paper here

7. Standardise, If everything else fails?

Don’t you hate it when old problems find new ways to sneak upon you? Donald Norman solved it decades ago, with his 7 design guidelines, and by now you would expect it to be absorbed in the fibre of interaction design. I doubt it is. The invention of new technology seems to invariably re-invent old problems.  But maybe this problem was never really solved.

The problem is ‘standardise’. It first hit me when discussing part of a interface of an iPad App, specifically how to close a pop-up window that only partly covers the screen? In an attempt to avoid a ‘windows like’ interface a cross to close the window was discarded and instead the user simply had to the tap anywhere (except on the pop-up window). The close button is the absence of a button, which is not yet standard therefore nobody recognises it and every user you see searching for the way to close the window.

Futjisi camera_OnOff
Image Source


I recognised the standardisation problem when, soon after discussing an iPad App interface, I had the pleasure of playing with the Futjisi X-camera. The discussion was about whether or not a pop-up window should have an x to close, i.e. adhere to a standard. In this context, the Futjisifilm camera is relevant because it does not have an on-off switch, but has the act of powering on the camera elegantly integrated in the act of usage. Elegantly beautiful. I love it. The Fuji X camera is switched on by simply opening/extending the lens, i.e. setting the zoom and focus. Turning the lens out switches the camera on. Returning the lens in, switches the camera off. This innovative approach did not survive the new versions, which again have a clear and recognisable on-off button. Zooming the lens has no impact on the status of the camera. An example of an innovative approach that did not match up to commodity. By the way, it has been discarded and new models of the Futjisi X-camera have an on-off button.

The_Design_of_Everyday_Things_(cover_1988)I feel this is the result from how we indoctrinate designers by considering standardisation is something ugly, something bad, to considered it as a ‘last resort’. We can trace this back to Donald Norman, who in his famous book abbreviated as POET, offers a list to guide interaction design. Number seven reads “when everything else fails, standardise” [POET, Page 189]. This has always  always bothered me, and I often referred to it in a mocking manner.

No, it does not merit a 5th postulate kind of debate, but I do like to challenge the ‘when everything else fails’ part of it.

Standing in the wasteland of interaction design, heuristics, formal or informal standards are not ‘a last resort’; they are a fundamental starting point. Referring them as ‘a last resort’ makes they are rendered invisible, ignored right form the start. Already at the start of the design process, standardisation is a fundamental constraint. John Flach formulated it for me like this; “the key is to make the constraints/opportunities/affordances/consequences visible to the user. This requires understanding both the functions of the device and the expectations of the user.” True. Absolutely, but this is a generic definition of what constitutes a good design process. I agree. But now what? A guideline must be like a road-map; how to get there.

Furthermore, standardisation is often not something you as designer can realise, which is suggested by by formulating the guideline as “when everything else fails, standardise”, with ‘standardise’ as an act, as something you as designer should be able to establish. That this seldomly true. Standardisation probably to happen in only a few situations, namely by 1) first movers, i.e. setting the standard,  2) by commodity; a grown heuristics that elevates itself to commonly accepted standard, or 2) by those in power, who can drive standardisation in a top-down manner.

  • An example of a standard set by first movers, is the side of the tracks most trains still ride. The first trains were developed in Great Britain, and in Brutian it is common to drive on the right side of the street, which tracks back to narrow London streets and most man being right handed. When trains were first introduced on to the main land, British engineers were asked for the implementation, which naturally resulted in the first trains running on the left track. As a consequence, even today, many trains drive on the lift side of the tracks. The power of the knowledgeable set the standard.
  • An example of a standard from commodity is the 12-volt plug in a car. Arbitrary at best [See for example ‘Little dongly things’ from Douglas Adams – Salamon of Doubts, page 142]. At the time, with more and more mobile electrical devices such as coolers and large radio players, were looking for power, the cigarette lighter revealed itself as opportune solution. Nowadays, you even find cigarette lighters in the rear of a car to offer a power to appliances if needed, unless of course the smokers are not only banned to the outdoors, but also to the boots of the cars (…seriously doubt that ). Microsoft’s based PCs also developed ground up. Due to relative low cost, and the modularity of components, Microsoft operated PCs became available in abundance, the power by the masses. Examples like Apple and Acorn, where hardware and software are much closer if not tightly linked, had a much more difficult time to survive. Eventually Apple did, but only after Bill gates committed to continuing MS Office for the mac development.
  • Examples for a standard set by those in power, there are a few. Of those who are (were) in the positions to define and set a standard, Steve Jobs is one of the most known ones. Not only with the introduction of the iPhone, but also with Pixar, revolutionising the animation industry.With every innovation, it is a balance between added value compared to the status quo versus the firmness with which the common way of working (commodity standard) is established. If there is no commodity standard, we are in the situation of the first mover; whomever grasps the largest market share in the shortest possible time-frame sets the standard. The more a commodity standard has been established, the more radical improving the innovation must be to become the new standard. The iPhones touch interfaces wiped away 20 years of telephone industry, almost overnight. USB ports are finding its way into the automobiles, but although there are many devices that power on USB ports, not all. The cigarette lighter may survive a bit longer as standard.

If you want to be successful as product designer, leaving the decision to adhere to a standard ‘only if everything else fails’ is not good advice. Although following standards does not guarantee a good product, standards in itself are not bad as they form a common language between users, and they offer recognition points for your device to gain acceptance. Standards (implicit or explicit), must be considered as primary constraint, right from the start. The problem with implicit or explicit standards is that hey have become part of our behaviour. And behaviour is difficult to change.

I propose to change the 7th guideline into something like: “To avoid everything fails, improve standards, or at least stick to them.” DO you have a better suggestion?



Norman, D. A. (2002). The design of everyday things. New York: Basic Books. (The re-issue, with a new preface, of The psychology of everyday things.)

A day in our digital life

In our digital life. we are connected constantly, but not all the time in the same way. To reach clients we need to be omni-present in the client’s digital life, whenever or however they connect .

Channel is a concept often used when talking about reaching clients.  A channel is an means to transmit data and information. Channels typically are connected to a platform of some sort. The newspaper is a channel. The television is a channel. The tablet is a channel. but are they really? How can the tablet be considered a channel if the newspaper is also a channel, since the newspaper is made available on paper, on the desktop computer and on the tablet. Or is the paper version of the newspaper a different channel compared to the digital version? As a reader, this is not how we expect them to work. We expect that if I start reading an article in the printed version, I can continue reading the same article on my smartphone, should the situation call for it.

The concept of channels, as information channel towards a (potential) client, may shift away further from the actual tool or technology used to establish the data and information connection, and become a specific aggregation of tools and technologies through which the target segment can be characterised.

At the moment, a channels’ definition is linked to the technology used. We refer to a channel as a single technological infrastructure, such as for example a tablet computer or a smart phone. Wikipedia explains; Historically, communicating data from one location to another requires some form of pathway or medium. These pathways, called communication channels, use two types of media: cable (twisted-pair wire, cable, and fiber-optic cable) and broadcast (microwave, satellite, radio, and infrared). Cable or wire line media use physical wires of cables to transmit data and information. Twisted-pair wire and coaxial cables are made of copper, and fiber-optic cable is made of glass. Nowadays, in information theory, a channel refers to a theoretical channel model with certain error characteristics. In this more general view, a storage device is also a kind of channel, which can be sent to (written) and received from (read), like for example a smart phone.

The way we use technology, this traditional definition of channel as technological platform does to suffice. In former times, people were either watching the television, or reading a book or listening to the radian. That where the simple times. To communicate to your target segment you had to select a (technology) platform, and maybe a time frame (e.g. run the advertisement in the morning or in the early afternoon? On sunday maybe?). Reaching your target segment meant pushing the information or message through the selected channel, i.e. platform. Placing an add on the television. Running an interview won the radio etc.

In current times, we have the i-family of products that we use imexchangably, and worse, sometimes all at the same time. At the moment I am writing this on a laptop while following twitter on my smartphone. In half an hour I will be on the train home, and probably first browse my mail on my tablet. Reaching your target segments in current times means sending the same message via different platforms that your target customers are using. A channel used to be a single platform, but tends to grow into a specific constellation of platforms that are used by the target segment in combination to basically perform the same activities.


This is illustrated with the image above; Nowadays,we are are connected throughout the day with the same source of information, but depending on the time of day and on the kind of activity, the platform and tool used to access this information differs even if our activities are the same or guided by the same interest. Take Jack for example. Jack lives in the city and commutes to work by tram. When he wakes up he first glances over our smartphone to check latest messages and news snippets. During breakfast he reads the news on a tablet. Commuting to work he again browses his smartphone, while at work it is his (desktop) PC that rules. Commuting back home he plays a game on his smartphone, and in the evening he watches a movie while chatting with friends and liking Facebook posts on his tablet. Then just before going to sleep, he glance a last time over his walls on his smartphone.

The point is that in order to reach Jack, in order to get his attention, it is not sufficient to address a single platform. To reach Jack, we need to send small notes or ramified information to his smartphone, maybe full texts with multi media to his tablet, and again small notes to his desktop (since this goes in parallel to ‘working’). The channel to Jack in fact consist of a smartphone, a tablet and a desktop pc. Now, maybe you are not Jack, maybe you are more like Peter. Peter has a smartphone and spends his time commuting reading the newspaper he subscribed too, on his tablet. The channel to reach Peter is different, not only the platforms it consist of, but also the manner in which Peter interacts with them, and probably also the amount of time spent on them.

‘channels’, as connections to transmit information and data, have grown into an aggregation of platforms that function in full symbiose, based on the behaviour of the target segment. In reaching this target segment we need to identify the platforms the channel is constructed from.

redefining what are channels, forces us to re-think how we reach our target segment, and invites us to move away from pushing data over individual platforms towards orchestrating information over various platforms in with the purpose of becoming more effective in informing and reaching the target segment throughout our digital and ’always connected’ day.

Making toast

ge_115t17_smallAn example of ‘elegant interaction’ is GE’s Hot Point Toaster (see Fig. source The problem, which this mechanism solved, was toasting a slice of bread on both sides. Modern devices have a simpler but less elegant solution: you simply slide down the slice of bread between two heating elements and a spring and bi-metal based mechanism releases it again when the slice has turned into toast. The ‘flopper’ mechanism was an earlier and I think a more elegant solution, at least from an interaction point of view. The ‘Flopper’ toaster had only one heating element, placed at the center of the toaster. In order to toast the slice of bread on both sides, you would need to turn it. Continue reading