[Larceny-users] ERR5RS Default Record Printer

[William D Clinger]

Ray Racine wrote:

> If ERR5RS records do not have a defined method for defining a default
> printer (someone please point me in the right direction if they do) then
> shouldn't they?

That is certainly a desirable feature, but standardization
efforts should distinguish essential from desirable, and
should make explicit (or at least conscious!) decisions
as to what is which for their particular purposes.  (That,
of course, presuppposes conscious decisions concerning the
purposes of the standardization effort, which is one place
the R6RS came up short.)

> Being able to define a default printer to a ERR5RS record seems a
> desirable feature to standardize.

Agreed.  Maybe not essential, however.

> Just wondering if this was an ERR5RS record design oversight or a
> conscious decision?

Conscious.  ERR5RS omits a great many desirable features,
or characterizes them as optional rather than essential.
This comes out of an explicit policy that ERR5RS should
not be "gratuitously difficult to understand or to
implement.  As the R5RS is extended, every new feature
should pay its own way by improving portabiity of user
programs more than it degrades portability by confusing
programmers or by discouraging implementors from
implementing ERR5RS."

If ERR5RS required implementations to allow customizable
printing of records, then it would become more difficult
to implement ERR5RS in systems whose implementors are
uncooperative.  One of the big things ERR5RS has going
for it is that we can implement it on top of most major
implementations of R5RS or R6RS Scheme, whether the
implementors of those systems like it or not.

ERR5RS in uncooperative systems will not be as nice as
ERR5RS in Larceny, but I can live with that.

Will