kernodle@dg_rtp.UUCP (Mark Kernodle) (05/08/87)
Speaking of Xrlib, is there going to be such a thing with V11 X? Will it be supported by HP? Will it continue to be in the public domain? Does it exist now? We have a fair amount of stuff build on XRAY, and it appears that the cutover to Xt is going to incur considerable rewrite, thus my concern.
frank@hpcvlo.HP.COM (Frank Hall) (05/15/87)
> Speaking of Xrlib, is there going to be such a thing with V11 X? YES! We will be making our plans for X-ray more public in the coming months. As we stated at the X conference in January, the X10R4 release of Xrlib was the first phase of an ongoing development plan within HP for an X-based user interface management system. That development is continuing in close coordination with our work on the X Toolkit, and we expect to make significant further results public this year. For a partial synopsis of our plans for this phase, see the paper I gave at the X conference. These plans include Xtlib-compatible callback functions, arglists, easier development of field editors, and a simpler and more powerful panel manager. New messages will be added where existing messages cannot be compatibly enhanced. Our goal is to make X-ray easier and easier to use and more and more powerful, while maintaining a stable, upwardly-compatible basis upon which our customers, ourselves, and interested members of the X community can base software development. We wish to help stimulate the formation of meaningful ease-of-use software for the Un*x market. Due to the necessities of timeliness, stability, and ease of customer transition to V11, our current X-ray2 development is based on V10 and will be available in that form. At that point X-ray programs will be able to talk to both V10 and the new V11 servers (through V10->V11 Xlib and/or protocol converter). This may not be available on the first V11 distribution (V11R0) from MIT; depends on the timing. In the X-ray3 phase, probably in 1988, we plan to use V11 and the X Toolkit (which we expect to be sufficiently stable at that point) to greatly increase the flexibility and power of the X-ray API. Note that the purpose of X-ray is not the same as for Xtlib. Xtlib has a goal of flexibility, at times at the expense of programmatic ease of use, and a goal of remaining policy free. It is an API-implementer's toolkit, which is not necessarily intended for use directly by the application programmer. X-ray, on the other hand, is an API with the goal of simplifying the application programmer's life by providing interactive gadgets with a coordinated policy and appearance, accessed with a coherent programmatic style, which are integrated into a user interface management system that can evolve to support increased programmer productivity. We are adding our planned X-ray2 features in such a manner that the X-ray API will be efficient to implement and extend using Xtlib. > Will it be supported by HP? Will it continue to be in the public > domain? Does it exist now? Yes, yes, and yes -- with clarification. Xrlib is and will continue to be a supported HP product. HP has donated Xrlib to MIT for public use, but does NOT provide support to the general public. Much of HP's X-ray-related ongoing development will be open to public use, but some parts will stay proprietary for a variety of reasons (licensing options may be available, however). Since X-ray2 is V10-based and will talk to V11 servers by compatibility paths, it DOES exist now (in as much as the compatibility paths exist). > > We have a fair amount of stuff build on XRAY, and it appears that the > cutover to Xt is going to incur considerable rewrite, thus my concern. Our customers would have the same concerns, and we have to attend to our customers. It took us a while to see how Xtlib and X11 were turning out -- X-ray3 was what we were hoping to do in the X-ray2 timeframe. We won't satisfy those X-ray programmers (e.g., at universities) who may wish to immediately use simultaneous X11 calls, or the neat flexibility of Xtlib, but we have concluded that neither X11 or Xtlib are stable (or smoothly evolutionary) enough to be a primary platform for our customer base in 1987. It's easier to skate up a ramp than a staircase. We'll publish more on our plans at a later date. We have an updated Xrlib right now that adds some functionality to existing field editors, fixes some bugs (including the XrInit/XGetDefaults conflict), our manual with an expanded tutorial introduction, and a suite of tests for verifying Xrlib ports, which we could make available to MIT for public distribution if there is interest. -- Frank Hall hplabs!hp-pcd!frank