Good news from the standards front. Committees WG21 and X3J11 voted out the draft C++ Standard for a second round of balloting as a Committee Draft. See Dan Saks's column for more details. For my part, I was happy to see the committees exercise a modicum of restraint right before this important milestone. It took me only a week to update my library in response to the changes voted in at the Kona meeting. After a typical meeting, I usually have to invest twice as much energy after a meeting to do the same task.
In principle, I should have had even less work to do. At the previous meeting, in Stockholm last July, the committees promised themselves to stop making any substantive changes to the draft. Kona was to be just an editorial cleanup and a last chance to fix anything demonstrably broken. The effect of this pronunciamento was not quite as absolute as many of us would have liked. It simply encouraged some enthusiastic advocates of last-minute change to broaden their definition of "broken" to cover any undesirable aspect of status quo. Nevertheless, the real desire to stabilize C++ did manage to damp out much enthusiasm for change.
Standard C++ is now well defined and stable for years to come. Well, mostly. The draft will certainly be stable at least until July 1997, because that's the next meeting at which changes can be voted in. And those changes should be only to fix things deemed "broken" by an ISO national body. But I must observe that the committees still have a few open action items — aspects of the C++ language or library that some members still feel need refining. Some change is inevitable.
I must also echo Dan's observations that the draft C++ Standard is still not a model of clarity. It describes a very ambitious language that has evolved considerably in seven years of standardization. Few people, if any, will assert that this document is anywhere near as precise as the C Standard, which describes a far less complex language. As a consequence, we can expect a steady stream of Defect Reports, to clarify the meaning of parts of the C++ Standard, beginning as soon as the draft is approved as an ANSI/ISO Standard.
But what the heck. At least we now know, pretty much, what C++ should look like for the near future. Vendors are rushing to flesh out the last missing features in their implementations. You can expect fairly complete compilers and libraries to appear at a steady rate throughout 1997. That's something to look forward to.
P.J. Plauger