At 17:45 14/02/02 -0800, Asmus Freytag wrote: >In principle this [not having a BOM] is a requirement for data being labelled >*external to the >data* as being in either UTF-16BE or UTF-16LE (ditto for UTF-32). These >formats *must not* have a BOM. >UTF-8 should *never* contain the BOM.
Even allowing this to be the case, I do not believe that any program aimed towards end users (eg. shrinkwrapped Windows programs) will **ever** insist on well-formed UTF so thoroughly that it complains if a BOM is present: there would be no benefit to the user in doing this.
