Formal comment #115 (enhancement) Bytes<->vector procedures are requested Reported by: Andre van Tonder Component: bytes Version: 5.91 Pages: 62-63 Discussion Currently ...-list->bytes and bytes->...list procedures are supplied to convert bytes vectors to lists of numbers and back. However, where one has the choice, programming with lists of numbers tend to be less efficient than programming with vectors of numbers. It may be useful to supply ...-vector->bytes and ...-bytes->vector procedures in addition to, or maybe preferably instead of, the list-based procedures. Since these can be implemented extremely efficiently, it seems a pity not to provide them. RESPONSE: We agree that these may be nice, especially with the addition vector-map and vector-for-each (see the response to formal comment 78). A number of other converters might also be nice, like string->vector, vector->string, as well as other flavors of ...-bytes->list, list->...-bytes, ...-bytes->vector, and vector->...-bytes that store or extract other forms of data, like 16-, 32, and 64-bit signed and unsigned integers, single floats, double floats, characters. Rather than adding all possibly useful such converters, it might be better to retain lists as the canonical intermediate form and encourage compilers to optimize expressions of the form (list->xxx (yyy->list ---)) into code that does the yyy->xxx conversion directly. In the absence of a concrete proposal for dealing with this issue holistically, however, the editors have decided to keep the status quo.