The rash of X Window-based hardware and software products either planned, or being released onto the market at present, are symptomatic of the computer industry’s increasing exposure to the infectious graphical windowing standard known as the X Window System. Wildfire transmission of the X bug is being driven by the increasing popularity of X Window – especially amongst the Unix community – and it is assured of being more than just a passing affliction judging by research from Dataquest, which suggests that the number of workstations running X Window is set to rise to over 2.25m by 1992 from around 40,000 now. The frontiers of windowing technology were pushed forward at Stanford University in the early 1980s and X Window itself began life as the W Window System in 1982 – a means of connecting terminals across Stanford’s campus-wide computer network. W Window was subsequently implemented under Unix – it became known as X Window in the process – and by 1985 engineers at the Massachusetts Institute of Technology had taken up the baton and were developing X under the auspices of Project Athena. As commercial interest in the possibilities of the system grew, the X Consortium was set up by leading manufacturers and educational bodies at the beginning of 1987 to fund and promote the development of X as an industry standard.
Upside down
The X Window system offers a user-friendly graphical windowing environment, giving users the ability to view applications running on the same, or different networked Unix systems within windows on the same screen. X divides the work between a server and client, communicating via X protocols, but here conventional distributed computing nomenclature is turned upside down. In X Window the X server resides closest to the user, for example on an X terminal or personal computer-based X display. The X application – the client – runs on the host system or on the same stand-alone system, where most of the processing is performed. The logic is that the job of a server is to look after, and manage, scarce resources – in this case the resource is the user’s attention. X11 Release 4, introduced at the end of last year, is the newest version of X Window from the Masscusetts Institute. It contains a range of new features, such as Shape, which enables the drawing of arbitrarily-shaped windows – before you were restricted to rectangles or squares and XDMCP, a new session manager supporting individually configured log-in windows to be set up more easily. In addition to the many bugs that have been fixed, there are new fonts and ICCCM support, as well as a complete XView toolkit which implements the Open Look graphical user interface and SunView application programming interface, and a new Athena widget set. Most importantly for X system manufacturers, the amount of code needed to draw and manipulate windows on screen has been cut in half, reducing dramatically the amount of X server code, and thus the amount of memory commandeered on X servers.
By William Fellows
The likely effects of X Window and its hardware side-show – X terminals – on the personal computer market is something of a hot potato with industry analysts at the moment, indeed the Technology Group of US broker PaineWebber has called the X movement the most important development in desktop computing since the personal computer. Most seem to think that manufacturers moving into the X terminal market can still make overall revenue gains, despite the loss in sales of some personal computers and workstations. X terminals themselves offer a number of end-user benefits, including multiple windows to different systems, multiple fonts, graphics, central administration and security. Both IBM and DEC have X terminals ready for introduction in the near future, which means that most of the major Unix manufacturers have now taken the plunge into X terminal technology over the last year or so, and are able to offer them as part of a computer system. However the technology – both hardware and software – is still in its infancy. At present X terminals are expe
nsive, at a minimum of $1,000, but more realistically $2,500, and up, but the market is set to expand rapidly. Although there are probably no more than 25,000 X terminals in use today, Dataquest forecasts a market worth $1,000m by 1993 with an installed base – including personal computers running as X-servers – creeping near to the one million mark within three years. The trade-off on X terminals is between memory and cost. A reasonable amount of memory – at least 2Mb – is required to handle application requirements and accommodate the essential networking and communications facilities. However one thing is certain – prices will fall. Firstly the presence of a host of manufacturers competing for market share will drive prices down, and secondly, the improved functionality in the basic X software itself should lower the minimum hardware specifications for X terminals, or at least free up more system space for other facilities. As well as IBM, DEC and the others, Japanese manufacturers will contribute enormously to the growth of the market. A consortium of 18 Japanese companies is currently developing a standard X terminal called the UWS – User-interface WorkStation – and other individual X-based hardware offerings from Japan will begin shifting in quantity during the course of the year – C Itoh Electronics has already pitched in with its CIT-X Network Display Station. This will be coupled with impressive implementations of X software from the Far East – Sony Corp is already reckoned to have one of the best X implementations yet available. At the moment X terminals fall into two categories.
Applications dearth
Dedicated terminals running X servers and extras only, and personal computer-based workstations running emulation software such as PC-Xsight from Locus, and PC-XView from Graphical Software Systems, or the more powerful XVision from VisionWare. The size of the market for X Window-based software looks like growing steadily, and this is guaranteed to provide a large target market for software developers. In particular, the push of personal computers into the technical environment and increasing Ethernet support should lead to a rapid take-up of personal computer-servers running under MS-DOS. The only blot on the horizon according to UK X specialist IXI Ltd is that different communications methods are sure to inhibit its spread. Although Ethernet and TCP/IP networking will be the major workstation connection, people will be using a whole range of methods, from 9,600 bps asynchronous lines to complex compressed protocols on high speed balanced lines – or ISDN – as well as different serial protocols across RS232 connections. So for the moment, most terminals will be used largely as multi-window displays for old software revamped for X. Also, but reckoned by IXI to be less important for its development, is the dearth of application software yet available for X. X terminal manufacturer Network Computing Devices reckons that there are currently around 38. The good news for users and potential X converts is that no one company has control over, or a clear lead in, X development. It means that firms can only justify more expensive X systems and software by adding true value with attractive innovations.
