Posted by arlene

Properly speaking, ‘Internet‘ is the name of a communication protocol (IP = Internet Protocol), which unsurprisingly is central to the communication between your computer and the rest of the world. However, the name has also come to be applied to the global network of computers connected using this protocol; this is perhaps the most common contemporary understanding of the word. It denotes a huge, diverse assemblage of many different types of computers, communicating via a common set of protocols, but ranging from text-only monochrome terminals, at the low end, to powerful PCs, Macintoshes and UNIX workstations, with Windowing systems, high-resolution colour screens, and many configurable interface options, at the high end.

Some computers have multiple users, either at the same time or at different times, and some are used exclusively by a single user. With multi-user systems, many users may be connected simultaneously via a single host, so we must distinguish usernames at the host, since the hostname for each user is the same. An opposite extreme is the single- user PC or workstation, with a unique IP address that can be uniquely traced to the user. Other PCs are used by multiple users in sequence, and of course some users may share usernames in all manner of complicated situations, so it’s not advisable to associate hostnames or IP addresses with individual participants without good reason.

A common distinction that will be used here is between client and server systems: clients typically provide the user interface to a communications system, and servers handle communication and distribution functions, ordinarily on machines that are remote from the user’s client machine. The existence of standard protocols allows clients from multiple manufacturers to communicate with servers from other manufacturers, so Internet use is not tied to a particular hardware platform or software manufacturer. For each major Internet technology, such as email or the Web, there are typically several alternative client programs available for any given platform: while some are free, others cost money, and they vary in interface design and the range of functionality available.

Both hardware and software differences are apt to affect the way information is presented to the user. Web content is generally displayed as well as possible on any given hardware, allowing a certain degree of standardisation despite local variations, for example in display resolution or in the range of fonts supported. But interface differences between competing browsers may cause them to present the same pages differently, even on the same hardware. From a research point of view it is problematic that there can be a much greater degree of variation in the hardware and software of the client systems in use than in traditional approaches to computer-based experimentation, since this can introduce experimental error.

This diversity presents problems for certain areas of psychological research, in which the physical properties of the display are of importance, such as psychophysical or speech comprehension research. Researchers in such fields require precise timing of events, control of display properties, and possibly special interface equipment such as button boxes, and none of these things can be guaranteed across the Internet. But it is possible to conduct research over the Internet that is unaffected by this diversity, for example email studies and simple Web- based surveys. This approach is to be recommended if possible, since it can reach the largest possible population.’ In studies that are likely to be affected by hardware or software variability, it may still be possible to conduct Internet-based research, since it is possible to retrieve some information about client machines. So although we can’t control the client machines in advance, we can at least take account of variations afterwards to some extent.

A more recent coinage is ‘Intranet’, which refers to a local-area Internet network, in a company or educational institution for example, which may or may not be connected to the outside world, but which is private to some extent. Material that is accessible from within an Intranet may not be accessible from outside. This division can be useful if we want to restrict use of a Web-based presentation system to machines in our own department. Many departments in academic institutions have networks of similar machines for students’ use and for research purposes, and restricting a program to execute only on these well-known machines can help control against hardware variations. Also, when working with such an Intranet, we are likely to have tighter control of the participant population since we will probably meet our participants, whereas recruiting across the Internet can sometimes result in unexpected and unwelcome surprises.

This type of Intranet-based research is very similar indeed to what psychologists have been doing for years with local-area networks (LANs), namely running a program on each of a network of machines and sending the results to a central server. The difference is the technology: nowadays, the use of an Intranet-based system gives access to a range of attractive new software technologies, and confers a high degree of portability and re-usability on software: Web-based presentation systems have a good chance of running unchanged on other departments’ Intranets, whatever hardware they use, as well as being usable directly across the Internet if required.

LANs are also useful for controlled interaction and dialogue studies, and the high bandwidth of a departmental network makes it possible to use live audio and video links with a high degree of fidelity.

More or less any reasonably recent computer is capable of connection to the Internet, either directly via a local Ethernet network, or indirectly via the phone system using a modem or ISDN. Most dialup connections use PPP (Point-to-Point Protocol), which effectively emulates a direct connection over the serial modem link, so once connected the machine has essentially the same relation to the Internet as has a directly connected one, with the difference that the bandwidth of the connection is much lower, that is, the rate of data transfer is much lower. This limits the viability of high-bandwidth applications, such as live audio or video links, which might be considered by researchers in dialogue for example.

While many Internet users still connect via 56K modems, more recent ‘broadband’ (that is, high-bandwidth) access technologies include cable modems, which use cable TV network cables, and Digital Subscriber Line (DSL) systems, which are typically available from telephone companies since they use conventional copper telephone lines. Broadband systems typically offer data transfer rates on the order of hundreds of kilobits per second, although local details will vary.

Possibly related posts: (automatically generated)
Internet, Intranets, and System Diversity

• Introduction to Web 2.0 Website Patterns
• Combating Opinion with Evidence
• User-Centred Design: the Internet Raises the Stakes
• Making Internet Content Easy to Use
• Global Vs Local Branding
• Communicating with Graphic Designers
• Web Searching, I do not use Software and Protocols
• Shoppers vs. Customers
• A Variety of Organizations and Contexts Part 1
• A Variety of Organizations and Contexts Part 2