I have the same feeling, notably because the exposed documents are meant to be fonted to have its notations readable and consistent.
And most probably because it creates new irrelevant character distinctions within rich-text formats (SGML, HTML, ...) to manage these characters as well as other occurences coded with markup in order to produce consistent output (here of subscripts). So suppose we code some subscripts used by Indo-Europeanist, and not some others. How will a rendered document look like if some occurences are coded with new separate characters, and oters coded with markup and standard characters ? Suppose now that such text is to be generated/converted into plain-text. Some occurences will be left unmarked, and some others may be left with the new characters. There will be additional difficulties to insert a consistent additional notation in the plain-text format to convert both categories of subscripts. If this notation is not explained in the text itself, the document would become unusable. But even if a conversion system is adopted, there will be problems to have it produce consistent results throughout the text for all occurences of either separate subscript letters and of standard characters with subscript markup. I much prefer to keep the encoding conservative, only to handle the case of bijective mappings from important legacy (non-Unicode) charsets in which they were introduced in the early times where rich-text formats were not easily interoperable and plain-text was the only solution. Today we have lots of way to create easily interoperable rich-text documents (HTML, SGML, XML, DocBook, PDF, RTF, Word docs, ...) without needing such pollution of Unicode. Also I doubt they were ever used in a legacy interoperable charset encoding. Authors will tend to use one of the rich-text formats where subscripts are easy to produce from almost all existing characters. ----- Original Message ----- From: "Rick McGowan" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Thursday, February 19, 2004 2:45 AM Subject: Re: (SC2WG2.609) New contribution N2705 > As long as we're on the topic, I have to weigh in on the conservative side > in this argument, with Ken Whistler. Use of the existing subscript > characters is generally bad practice. Adding more subscripts would be > adding to the bad practice, and yield even more different ways to express > the same thing (markup versus direct encoding).
