RWS@ZERMATT.LCS.MIT.EDU.UUCP (02/13/87)
Someone has asked me to summarize the protocol BOF, and where X is headed as far as standards and such. [Somehow seems like I've already done this before, but...] Those present at the BOF made it clear that it was very important that the core X protocol remain stable, and that there be controlled evolution of "blessed" extensions. Stability of language bindings (such as Xlib) and toolkits was equally important. Three possible routes to stability are 1) for MIT to continue in its role of the "owner", 2) for a recognized standards body such as ANSI, IEEE, or X/Open to become the owner, and 3) for an independent non-profit organization to be established or identified as the owner. Those present were mostly of the opinion that the fact of immediate and continued stability was more important than adoption by any official standards body. Furthermore, there was considerable concern that an official standards body would take five years to proclaim something a standard, and in the interim would be unable to resist the temptation to tinker with or totally revamp the specifications. Their opinion was that, while we should certainly pursue official standardization, in the near term an MIT-controlled organization would be very desirable. The other important aspect of the X future is the controlled integration and release of public software. Project Athena has been playing this role, but does not desire to continue this effort forever, and software is not something that existing standards bodies seem to deal with. Again, the suggestion was that the MIT-run organization could perform these tasks. Since the BOF, the subject of X has been brought up before the ANSI windowing committee; I forwarded George Champine's report on that to xpert a while back. MIT will be meeting with ANSI representatives in early March for further discussions. Since the BOF I have also had discussions with various people at MIT about setting up an X organization. In essence, the idea is to set up a very small organization, perhaps consisting of a manager/politician to oversee the standards side, a good technical person to do software maintenance and release, a lesser technical "slave" to do grunge work, and at least a part-time secretary. Actual distribution would continue to be handled by the MIT Microcomputer Center and/or other interested parties. The assumption is that all real "development" work would happen outside the organization. The organization would be supported by money and hardware from vendors; in return, the vendors would know that the software worked on their systems, and would have various privileges (like early access to new releases). On the standards side, the organization would promote and coordinate open discussion and participation in the specification of extensions, language bindings, toolkits, etc., and would also act as a registry for such things. Technical expertise on wide-ranging subjects (like 3-D) would not come from the organization itself, but from experts in companies and universities interested in the subject. An important part of the charter of such an MIT organization would be to figure out how to (within a year or two) cease being an MIT entity, either by having some standards body take things over, or by spinning off to a new or existing self-sustaining non-profit organization. Nothing is firm at this point; various discussions are continuing. Until something concrete is established, control of existing Version 11 components will remain firmly in place. If anyone would like to discuss things further, by all means contact me.