[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Mail from Skona Brittain

The OPEN-KEYWORDS proposal which I just sent arrived in my hardcopy mailbox
today. I passed it through completely unedited; in fact, I didn't have time
to read it while I was typing. I'll comment later when I've had time.

There was also a cover letter that said:

    ... I still have not been able to do much of anything with my account.
    I tried to send mail but kept getting "no such local user" messages,
    even when just directly responding to a test message from a non-local
    friend. Nor have i received any X3-J13 mail at all yet. So if you want
    to respond to this, I would appreciate your trying my netmail address,
    but if your mail gets returned, please use [P.O. Box 747, Santa Barbara,
    CA 93102; (805) 963-3412].

    Regarding the order of arguments' evaluation, as in the 
    push-evaluation-order proposal, I still believe that this is specified in
    CLtL. I mentioned the reference on page 97 to "the usual left-to-right 
    order in which the various subforms are evaluated" but have since found
    a less oblique one: the entire last half of page 99.

    It is my impression that there needs to be a clarification of the
    effect of *print-level* and *print-length*, so if this impression is
    correct, I will volunteer to write it up. Basically, there seems to be a
    confusion about whether it is the actual components of the object that
    count or what the object looks like when printed in list notation. For
    example, the levels of 'x and (quote x) are considered different (cf
    page 373) but string-char arrays of rank >1 are affected by
    *print-length* even when printed in non-list notation (cf page 369).
    Other cases that are affected by this distinction include
    - a structure with n components has 2n+1 elements in its printed list
    - nil vs. ()
    - a rank 0 array has a component but prints with no parentheses whereas
    a 0xN array has 2 levels of parens and no components, etc.
    Since I am obviously not as well-qualified as the rest of the
    clean-up committee to judge such issues, and since I don't have access
    to any other preliminary feedback, there does not seem to be much point
    in my proceeding with writing up something like this until further
    A suggestion for a change that I have is that some of the defstruct
    arguments be strings instead of symbols, but I also don't know if
    there's any point in entertaining it.

    I was under the impression at the meeting Monday that thecleanu-up
    committee was going to suggest setting up several other committees,
    including one on the file systems handling, but when I reminded Larry on
    Wednesday, he said there weren't any that hadn't already been set up, so
    I am unsure of the status of the situation regarding a file
    subcommittee. ...