I'm in favor of adding some more examples of decorative text:
- uʍop-ǝp!sdn
- ⓒⓘⓡⓒⓛⓔⓓ
- ¸,ø¤º°`°º¤ø,¸¸,ø¤º° Accents °º¤ø,¸¸,ø¤º°`°º¤ø,¸
- !!!!!!!!! Excessive exclamation marks !!!!!!!!!
- U̲n̲d̲e̲r̲l̲i̲n̲e̲d̲
- 𝕱𝖗𝖆𝖐𝖙𝖚𝖗
- 𝒞𝓊𝓇𝓈𝒾𝓋𝑒
- 𝘐𝘵𝘢𝘭𝘪𝘤
- 𝐁𝐨𝐥𝐝
- 𝙼𝚘𝚗𝚘𝚜𝚙𝚊𝚌𝚎
- Z̸͔̳̥͔̯a͙̥͡l͏g҉͕͇̻̝̪̠̰o̥̝̹̬̣̣
Note that the formatting above is different from normal formatting. Computers encode letters, numbers, and symbols as numbers, and these numbers are referred to as "characters". Normally, formatting is stored separate from these numbers; the text itself isn't italic, bold, or colored. It's sort of like a car: you can paint a car any color you want, but it's still the same car.
These kinds of text listed above don't have normal formatting. Instead, they're using obscure character codes to encode their formatting. Many of these characters only exist for compatibility purposes and aren't meant to be used unless absolutely necessary. This creates numerous problems. Search engines are a pretty good example. Let's say you want to search for the letter "A": your computer is going to search for the corresponding number in its memory, which is 65. Although the letter "𝒜" might look the same to you, your computer has no idea that it's the same letter as "A"--it has a completely different number. Some programs and services have giant lists of equivalent characters; for example, if you google "𝒜", it will take it to mean "A". However, none of these lists are perfect, and there are often ambiguities. It's significantly more reliable to just type "A".
Additionally, when formatting varies in titles--no matter how the formatting is encoded--it detracts from the usability of the site. It's harder to scan through a list of threads and pick out the relevant ones. Instead of drawing attention to threads, the formatting is probably more likely to frustrate people and cause them to skip over the inconsistencies.