I put the results of a generation for comparison on <a href="http://macchiato.com/idn/UnicodePropertyResults.html">http://macchiato.com/idn/UnicodePropertyResults.html</a><br><br>A few notes:<br><ul><li>We've been forgetting to remove default-ignorable-code-points, so I added an exclusion. It only affects variation selectors.
</li><li>We probably want to remove Runic (Runr) as a historic script [although I didn't yet.]<br></li><li>I used the block notation instead of the raw ranges that Ken has.<br></li><li>This is from a program I used for testing properties, so the lines at the top, like the following, are actually executed to produce the results:
</li><ul><li> Let $baseId = [$gc:Lu $gc:Ll $gc:Lt $gc:Lo $gc:Lm $gc:Mc $gc:Mn $gc:Nd]</li><li># this sets a variable $baseID to the union of a number of property-based sets based on general category values.<br></li></ul><li>
The ## comments outline what is done to get the different results. I currently generate 3 lists. <br></li><ul><li>The base list, by range<br></li><li>Then taking out the historic scripts and symbol ranges that Ken recommended, by range
</li><li>A detailed version of the second, but skipping the big alphabets.<br></li></ul><li>The output is the standard Unicode data file:</li></ul><table style="border-collapse: collapse; margin-left: 40px;"><tbody><tr><td>
0030..0039</td><td>;</td><td>Zyyy</td><td> #</td><td>Nd</td><td>[10] </td><td>(0..9)</td><td><br></td><td>DIGIT ZERO..DIGIT NINE</td></tr></tbody></table><div style="margin-left: 40px;"><range> ; <script> # <general category> [<range count>] (<character(s)>) <name(s)>
<br><br></div>The results is an html file, although one could dump it as text. The characters are also shown, although you'll only see them correctly
if you have a reasonable collection of fonts. Firefox is better than IE
at falling back to whatever fonts are on your system. But it is also easy to pull into Excel (OpenOffice) for sorting or filtering by different fields, such as the script field or the general category field.<br><br>Mark<br>