Why then stars ? Any symbol, even any Unicode letter could be repeated and half-filled. Even logos (I've seen Apple logos used this way) or pictograms (I've seen film rolls for cinema rating, or trumpets for rating music, or beds for rating hotels, or forks/spoons/knives for rating restaurants, or flowers/leaves for rating a natural environment or its preservation, or seals for rating quality of documents, or diploma hats for rating universities and schools, or academic palms for rating researches, or droplets of water for rating the quality of waters, or fishes for rating fishing sites, or camps/teepees for rating campings, or open hands for rating welcoming acceptance, also various figures of hands with raised fingers, or street fires with a 0 to 3 lamps lit, or just one at the position of green or orange/yellow and red, or various road navigation signs. Also various fors of checkmarks or votes).
All of these are graphical substitutes for digits. All of them frequently have theur own local graphic design, local color conventions, local sizes. Even more frequently today, these prictograms display rich color and light effects. Using fonts for trying to display them is in fact the worse solution, and it will be just simpler to represent the existing digits (that are more precisely recognizable, even if they are not "beautiful"), and then leave a rich-text or graphic renderer using non-plain-text styling to perform the necessary substitutions (most often using a site-specific or document-specific choice of icons). The problem is clearly not that of encoding : the encoding is already solved with digits. You're just trying to add geometric shpaes to allow display poor icons that don't even match the need for colorful and beautiful, and distinctful icons. Even today, using the existing Unicode for the WHITE STAR character allows performing styling on it to render an empty, full, or partially filled star. This will be done via a stylesheet converting digits into styled stars. You don't need new Unicode characters for that. And the most interchangeable form for plain-text encoded in Unicode will remain the use of standard digits (now nothing prohibits us of using existing Unicode characters, but none of them have not been encoded for this derived use : don't complain if this does not work properly or if readers don't have a font to render them if all what they have is a renderer capable of displaying standard text with classic letters, digits and punctuation used in their language, with variable glyph design matching their own preferences, and their own requirement on font sizes for accessibility). If you start encoding a document using uncommon characters, automated Braille or aural readers won't know what to do with them : it may look smart for non-blind people reading your document visually, buth there should be a clear alternative way where the actual intended semantic is stored in the document and numeric figures are certainly best : they are accessible). It will be better and to read "3 stars" rather than "symbol, symbol, symbol" (where the same term "symbol" may be used for automatically reading distinct symbols occuring in the document, so loosing compeltely the semantic differences expected and making the document completely unusable). Now suppose that a smarter text reader tries to use the Unicode character name (in English) will it be understood in another language than English? Will it be smart when it will read "BLACK STAR, BLACK STAR, BLACK STAR" ? Will the reader need to detect these repetitions to infer "3 BLACK STAR(S)" ? Now what will happen if stars are becoming different ? "BLACK STAR, BLACK STAR", vs. "BLACK STAR, 33% BLACK STAR", vs. "BLACK STAR, 33% WHITE STAR", vs. "BLACK STAR, WHITE STAR", etc., when the actual meaning was a more compact and understandable numeric figure (here: 6/6, 5.6, 4/6, 3/6...)... I wonder how a reader for accessibility or automated parser will infer the correct meaning. In some fancy documents you may even use random symbols of roughly the same displayed size (apples, pears, peaches, bananas, raspberies, oranges, tomatoes... just chosen within a common family, and sometime using real photos rather than icons) : what will be significant is their presence or absence, or their total count. Not the glyph, and not the colors used visually... For me all the graphical substitutions of numeric figures are NOT plain text, they are presentational features for visual rendering, and part of the styling of a rich text document support (even when it is not electronic but engraved on a stone, metal or plastic/resin plate, or on the leaver of a precious book cover or printed on a T-shirt).
