[rfc-i] Scanning non-ASCII text
mrex at sap.com
Wed Sep 26 09:51:48 PDT 2012
Paul Hoffman wrote:
> Greetings again. The -00 draft gives as an argument for keeping
> the all-ASCII requirement:
> * In extreme cases of having to retype/scan hard copies of
> documents (it has been required in the past) ASCII is
> significantly easier to work with for rescanning and retaining
> all of the original information. There can be no loss of
> descriptive metadata such as keywords or content tags.
While the wording if this paragraph is debatable the content is
acutually quite valid and quite important.
> This doesn't hold up for two reasons. First, scanning software has
> handled non-ASCII characters just fine for well over a decade.
The scanning software ("human comprehension") that is involved in
most real-world consumption of I-Ds and RFCs by humans in a an optical
character recognition (OCR) process called "reading" has historically
been coping poorly with glyphs from other languages, is still coping
poorly with this today, and there is currently no indication that
this is going to change in the future. Numerous glyphs from Unicode
are completely indistinguishable when visually rendered on display
devices for processing by the "human eye" OCR system.
> Second, and more important: no RFC created in the future will
> exist only in a printed version.
You're mounting the horse from the wrong end.
Changes that are completely unnecessary, but will make printed copies
hard to comprehend, are to be avoided. There better be a very good
reason to break an installed base.
You also have to keep in mind, that there are certain popular tiny
display (i)Devices, where the manufacturer makes it arbitrarily
difficult to share information on them with your PC, and also
arbitrarily difficult for end users to work around these shortcomings
with own software, and there is currently no indication that this
manufacturer is going to change his hostile attitude towards the
end users who buy these devices.
More information about the rfc-interest