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

Re: [Scheme-reports] Write procedure is not backwards compatible

On Mon, Jul 2, 2012 at 8:19 PM, John Boyle <johnthescavenger@x> wrote:
>> We definitely cannot use the SRFI-38 semantics, which
>> provide no equivalent to write-simple (i.e. no procedure
>> which guarantees to be fast).
> I see a use case, where a user generates code with shared structure and
> "write"s it to a file, then expects to be able to read that code back in and
> execute it.

There are two motivational factors behind write-simple:
speed and the desire to generate output that does not
use the reader label notation.  We both agree that
write-simple is necessary, you're just claiming that
speed is not as much an issue.

My point about SRFI-38, however, was that it says:

  This SRFI permits but does not require replacing the
  standard (write) and (read) functions.

which was very unfortunate, because you don't have
a guaranteed write-simple (with or without error handling).