YAPP ignores subscripts/superscripts, causing misleading required parts of answer

[hftf]

Here are 4 examples just from two adjacent packets:

1. The 2 should be underlined in Packet 8, bonus 13
   Packet: N₂O
   YAPP:   N2O (“N-2-O”) [or nitrous oxide] ⇐ <b><u>N</u>2<u>O</b></u> (“N-2-O”) [or <b><u>nitrous oxide</b></u>]
2. But for some reason the 2 is underlined correctly here
   Packet: NO₂
   YAPP:   NOx (“nox”) compounds [accept NO AND NO2 ⇐ <b><u>NOx</b></u> (“nox”) compounds [accept <b><u>NO</b></u> AND <b><u>NO2</b></u>

   

3. The 2 should be underlined in Packet 7, bonus 9
   Packet: CO₂
   YAPP:   CO2 (“C-O-2”) [or carbon dioxide] ⇐ <b><u>CO</u>2</b> (“C-O-2”) [or <b><u>carbon dioxide</b></u>]
4. The 2 should be underlined
   Packet: H₂
   YAPP:   hydrogen [or H2 (“H-2”)] ⇐ <b><u>hydrogen</b></u> [or <b><u>H</u>2 </b>(“H-2”)]

   

For what it's worth, I have strongly recommended against any use at all of subscripts/superscripts in quizbowl packets for many years, and this issue demonstrates two of the reasons: they mislead about the required parts of answers and interact poorly with underlining, and technology is often not designed to handle them correctly. Like #49, this issue should be classed high severity.

Note also that invalid tag nesting is produced throughout: <b><u>hydrogen</b></u>. Can file a separate issue for this if you want, but it's possibly the same root issue. See:

YetAnotherPacketParser/YetAnotherPacketParser/YetAnotherPacketParser/Compiler/Html/HtmlCompiler.cs

Lines 103 to 156 in 00c6f58

	private static void WriteFormattedText(FormattedText node, StringBuilder builder)
	{
	Verify.IsNotNull(node, nameof(node));
	if (!node.Segments.Any())
	{
	return;
	}
	bool previousBolded = false;
	bool previousItalic = false;
	bool previousUnderlined = false;
	foreach (FormattedTextSegment segment in node.Segments)
	{
	if (previousBolded ^ segment.Bolded)
	{
	builder.Append(segment.Bolded ? "<b>" : "</b>");
	previousBolded = segment.Bolded;
	}
	if (previousUnderlined ^ segment.Underlined)
	{
	builder.Append(segment.Underlined ? "<u>" : "</u>");
	previousUnderlined = segment.Underlined;
	}
	if (previousItalic ^ segment.Italic)
	{
	builder.Append(segment.Italic ? "<em>" : "</em>");
	previousItalic = segment.Italic;
	}
	builder.Append(segment.Text);
	}
	// Close any remaining tags
	if (previousBolded)
	{
	builder.Append("</b>");
	}
	if (previousUnderlined)
	{
	builder.Append("</u>");
	}
	if (previousItalic)
	{
	builder.Append("</em>");
	}
	}
	}
	}

[alopezlago]

I've been meaning to file a bug and fix super/subscript issues for a while and never got around to it. We have to read the super/sub script tags from docx and from HTML to know that it exists, persist that information, and then write it later (and have MODAQ know about super and subscript tags). I don't really care that much about mixing the tag ordering that much, but I could fix it separately.

[alopezlago]

This will be fixed in the next version.