How anyone can comment on the FCD

Do you want to comment on the C++0x Final Committee Draft (that’s the link where it just went live online and is freely publicly available), but you aren’t an official member of some ISO national body?

Well, you can: UK is volunteering to channel your comment. Thanks, Anthony and the rest of the BSI panel!

(Open process: Indeed.)

6 thoughts on “How anyone can comment on the FCD

  1. This is extremely late in the process. The single most important feature (concepts) that I wished for was removed, and I am not too pleased on a number of other developments. Whatever comments I have now is realistically too late.

    Is the process open enough that I (or the next commenter) can simply go to the meetings, participate in the discussions, and casts votes? This is a rhetorical questions as I already know the answer to it.

  2. @Brian: I’m sorry, but you’ve been misinformed.

    This is extremely late in the process. The single most important feature (concepts) that I wished for was removed, and I am not too pleased on a number of other developments.

    But every working draft has already been freely available at the committee website since the beginning of the project; that’s typically three full drafts a year. Anyone could (and many did) submit informal comments on those at any time without waiting for an ISO ballot, for example via comp.std.c++ which since the 1990s has been an avenue to get public feedback and discussion of drafts at any time without waiting for a committee meeting or international ballot.

    No one can rightly claim to be excluded and not have had opportunity to review any draft and provide any comment. Proof by counterexample: I can name people in Europe who are not committee members and are poor, and who can’t afford to be committee members or to ever attend a committee meeting, who have nevertheless been extremely influential and valued participants who have influenced the language’s and library’s design and been huge sources of helpful feedback that have helped us refine this work.

    Anyone who wants to spend the time to participate could do the same without even waiting for a CD or FCD ballot. The point of releasing a full-on CD (fall 2008) or FCD (now) is to ask for official international review by national bodies to ask them whether they will approve the results of all of the above input and work as their national standard and as an international standard, and naturally that question waits until the document is consistent and complete enough to comment on as a whole. This avoids, for example, wasting their time deeply reviewing something that we know is in an incomplete and inconsistent state. But nobody had to wait for that if they wanted to participate earlier when the draft was still in a rawer form, as I’ve pointed out on this blog many times as well as above.

    BTW, concepts were removed because there wasn’t a consensus that they’re ready for C++0x, not because people didn’t wish for them — like you, everyone pretty unanimously likes concepts in principle and hopes to see them again the next time around.

    Whatever comments I have now is realistically too late.

    Even now, it’s not that late; we’re going to be working on comments for another year. Incidentally, this isn’t the first time UK has offered this avenue to enable general public comment. Even if you didn’t participate until now (see above), you now have several months to read the draft and send comments in, and we plan to spend a full year on considering all submitted comments (including any that UK passes through from the outside world).

    Is the process open enough that I (or the next commenter) can simply go to the meetings, participate in the discussions, and casts votes? This is a rhetorical questions as I already know the answer to it.

    The answer is and always has been Yes, including casting votes at the meetings in straw polls in technical subgroups (e.g., library working group, core working group, which can make or break features). Anyone who wants to make the effort to come is welcome to attend in person and participate fully in the debate. At nearly every meeting we do indeed have people who have never come before and are not members of a national body panel — for example, last month we had a number of local students from CMU and other schools attend and participate.

  3. Congratulations on getting the FCD out! I am looking forward to glancing it over very soon.

    @Brain: Isn’t that argument getting boring yet? Aren’t there better places to vent your opinions?

    @Herb Sutter: I was reading over the history of C++0x concepts and why they were dropped again recently, and I re-read about one (radical) proposed option which I thought was very appealing: Splitting the C++ Core Language standard from the C++ Standard Library standard. Interesting idea, and I would live to know your take on that thought.

  4. @Michal: Briefly, I think it’s very important for the C++ language and library to evolve and ship together. They could be separated, but there’s real value in the library adopting language features, and if we separate the two then the library would naturally lag the language which would be bad for a number of reasons including that we’d have language features that the standard library doesn’t use (at least for a while) and that the language designers would have less experience with how new language features affect and work with the standard library.

  5. @Herb: That was exactly my point: I argue that having library changes lag language changes could be a good thing. I would instead call it an “accelerated language standard” and not a “delayed library standard”, and would also add that independent library changes could be smaller and made more often than they are now (“more agile”).

    I do understand many (I dare not say all) of the negative ramifications, and understand the reasons why it is important for the two to evolve together.

    I only discuss alternatives because sometimes it can be disheartening to see the cost of releasing such massive, monolithic, (..high quality.., ) language-library-symbiotic standards. Sometimes waiting for a small-change big-impact feature (like say “auto” or “r-value references”) for over 5 years, each delayed not because of direct issues but because of unrelated features, makes the situation seem worse than just “unfortunate”.

    Perhaps my argument isn’t actually against a single standard. Perhaps the solution is simpler: What if new language standards were simply smaller in scope?

    C++11 is ridiculously packed with goodies. If it had come out as several separate standards several years apart, each would still have been hugely beneficial. Maybe that was the hope for C++0x, and hindsight is 20/20.

    So, ignoring the unforeseeable delays to C++11 (i.e., concepts), are opinions positive on the size of the overall scope? Was too much attempted at once?
    Are there any plans for/opinions on the scope of the next C++ standard yet? Any lessons learned?

    Thanks for reading and answering Herb.

Comments are closed.