When the X Window System became an overnight success six years ago and just about all the big vendors adopted it as the basis for their graphics and windowing, it left the Massachusetts Institute of Technology in an embarrassing position. Here was an interesting research project that could now command big money backing from the industry, which at the same time needed new levels of support now that it had escaped the confines of academia. MIT’s answer was to sponsor the formation of a consortium that could be industry-funded. The result was the MIT X Consortium, which opened its doors in 1988 under the directorship of Mr X – alias Bob Scheifler, the principal architect of the X protocol and before that a designer on the Argus distributed system and CLU programming language at the Massachusetts Institute. Unlike other consortia, this one chose not to impose licensing fees onto X users, relying for its funding on membership dues and using only a small staff. Many of the innovations in X since then have come from member firms, who took their X sample source from the MIT X Consortium and then went off and tightened up the code for commercial use.
Reluctance
On the whole things have worked out well, with the X Consortium and X in general remaining vendor-neutral – despite some heavy involvement from individual companies, particularly Digital Equipment Corp, and an initial reluctance to get involved from Sun Microsystems Inc, which preferred its own PostScript-based NeWS Network-extensible Windowing System. What was lacking was a consolidated marketing and promotional programme, a role which the newly formed X Industry Association is to take on, and a long-term vision about what should and should not be included within X. Such planning has become more important as the use and scope of X has broadened, but also as Microsoft Corp’s Windows becomes ever more pervasive. The two solve different problems and can co-exist to some extent, but perception can be as important as technology, and X Window must define its purposes carefully if it is to survive the onslaught. So now the consortium is to be spun off into a separate organisation under the new leadership of Luther Abel, a veteran graphics developer who has spent the last two decades working at companies such as DEC, Applicon Corp, Via Systems Inc and Data General Corp. Bob Scheifler won’t be spun off with the consortium, and plans to go back to non-X related research at the Massachusetts Institute. The new X Consortium, says Abel, will work with the same business model as the old: If it works, don’t break it, he says. But Abel’s task is to keep X in the lead as the mainstream client-server windowing and graphics standard, pay more attention to its strategic direction, work with the X Industry Association, and define just what goes into the next releases, specifically X11.7.
By John Abbott
This year he must organise an orderly move away from MIT, a doubling of staff and the roll-out of the alpha version of X11.6, all by year-end. The beta and final releases of X11.6 will take place next year. Luckily for Abel, the departing Bob Scheifler has already overseen most of the development of X11.6 and at the Xhibition conference a few weeks back he outlined the main features likely to appear. These include LBX, a low-bandwidth version of X that uses a proxy X server close to the client that can be connected via high bandwidth link, enabling the remaining distance to the real X-server to be made over serial or wide-area network links. X-specific compression utilities and compression techniques are also part of LBX. Enhanced XIE rendering services, with a reduced subset package for document image processing small enough to be run locally on an X-terminal is promised – a feature particularly useful for incoming facsimile messages. Other areas outlined by Scheifler include graphics synchronisation extensions for video, animation and real-time use, a screen saver extension providing Macintosh-style flying toasters and fish on X-terminals, X-test extensions, enhancements to the PE
X PHIGS Extension to X graphics standard and support for deep frame buffers required for heavy three-dimensional graphics work. The Kerberos security mechanism will be integrated into the core X protocol, the X Server Xlib and (to some extent) the Xt toolkit will become multi-threaded, internationalisation support will be enhanced, and the first features of the Fresco C++ and structured graphics toolkit extensions should appear. Conformance to the Object Management Group’s Common Object Request Broker Architecture standard is planned. It’s not yet clear whether the Consortium’s drag-and-drop mechanism – said to go far beyond the functionality of the current Motif drag and drop mechanism will be ready in time, the fall-back position being the Open Software Foundation code. Abel is responsible for the development of X beyond Release 6. He says that although X11.7 has yet to be defined, it is possible to speculate on at least some of the areas where additions will be made. One will be inclusion of a Display PostScript extension, something that was once slated to appear in 11.6 but was held back due to a lack of development resources. The new generation Fresco toolkit should also appear in its fully fledged form. PEX version 6, reconciled with the final PHIGS+ standard and including advanced rendering, better performance and more extensibility will be included, but may cause unease as it’s a compatibility break with PEX 5. The Consortium says it is striving for compatibility. Support for audio and compound documents is also in the pipeline.
Legal entity
Content from our partners
Many of these features will be made window system-independent, so that development work can also be re-directed to work under other windowing systems such as Macintosh and Windows. At the insistence of its current staff, the X Consortium will remain in Cambridge, Massachusetts: it is already a legal entity and Abel is currently looking for office space with the intention of moving by the summer (summer never seems to arrive in the US). It’s Cambridge versus the rest of the world to our people he says: even out to Route 128 is the rest of the world. Yet X may have to look beyond the confines of its own little world to survive the competitive pressures of its next six years.
