Encode 50 million - Roman Numeral - Best practice

KP Mawhood

This should be the 50,000,000 symbol, a capital D flanked by 2 vertical lines and surmounted by 1 horizontal line.

4 character entries seems terribly convoluted, perhaps better than PUA for forward-facing accessibility?

• 007C VERTICAL LINE
• 216E ROMAN NUMERAL FIVE HUNDRED
• 0305 COMBINING OVERLINE
• 007C VERTICAL LINE

Any suggestions for a better approach? I cannot find this in unicode.

Thanks

KP Mawhood

Although given the structure of 499,000,000, nevermind. Sorry.

Chris Lozos

Take it away, PUA  ;-)

Igor Freiberger

Katy, I also think it is better to build a precomposed character in PUA.

The need is limited to 50,000,000? I mean, there are not only several other combinations, but also historic variations like circled X, barred V, etc. You may need other characters and also could consider a set of basic numbers with bar above to mount other high numbers if there is such demand.

With OpenType code it is also possible to transform regular numbers into Romans automatically. I did that some time ago (see the PDF below).

John Hudson

Igor, from your PDF:

It’s controversial if a pair of glyphs which does not touch each other must be

It really isn't: in digital text terms, a ligature is any glyph that represents more than one character.

John Hudson

It would be nice if there were an enclosing sign in Unicode that could be used to create higher denomination Roman numbers. There's been clever work done recently to handle cartouche enclosures of Egyptian hieroglyphs, so putting a box on three sides of a Latin letter or sequence of letters seems possible.

Igor Freiberger

Thanks, John. The note about ligatures results from a length discussion developed in Typophile then (2010). I did not save the page, but I remember there was opinions for both sides, especially regarding Dutch IJ. Good to know the issue is actually solved.

Andreas Stötzner

In the MUFI PUA repertoire there is a Roman D with bar above, F7B6. The other Roman numeral letters alike.

Russell McGorman

Igor Freiberger said:

..
With OpenType code it is also possible to transform regular numbers into Romans automatically. I did that some time ago (see the PDF below).

Thanks for sharing this, Igor. I've been working on a solution to this problem  which had been proving rather labour intense. 

Severin Meyer

@Igor Freiberger, correct me if I'm wrong, but it should be possible to implement the feature more succinctly like this:

@DIGIT = [zero one two three four five six seven eight nine];<br>@ROMANM = [zero.rm one.rmM two.rmM three.rmM four.rmM five.rmM six.rmM seven.rmM eight.rmM nine.rmM];<br>@ROMANC = [zero.rm one.rmC two.rmC three.rmC four.rmC five.rmC six.rmC seven.rmC eight.rmC nine.rmC];<br>@ROMANX = [zero.rm one.rmX two.rmX three.rmX four.rmX five.rmX six.rmX seven.rmX eight.rmX nine.rmX];<br>@ROMANI = [zero.rm one.rmI two.rmI three.rmI four.rmI five.rmI six.rmI seven.rmI eight.rmI nine.rmI];<br>&nbsp;<br>feature ss09 {<br>&nbsp;&nbsp;sub @DIGIT' @DIGIT @DIGIT @DIGIT by @ROMANM;<br>&nbsp;&nbsp;sub @DIGIT' @DIGIT @DIGIT&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; by @ROMANC;<br>&nbsp;&nbsp;sub @DIGIT' @DIGIT&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; by @ROMANX;<br>&nbsp;&nbsp;sub @DIGIT'&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; by @ROMANI;<br>} ss09;

I don't think it is a good idea to repurpose the .sc glyphs, because they have a different semantic meaning. For example, if you apply your feature on the number 1001, you would get M.sc zero.rm zero.rm I.sc. Consequently, an attempt to reconstruct the original number via the glyph names would give you M00I. With the above implementation, the semantic meaning of the sequence is preserved.
 
I appologize for going slightly off topic.