While combing through English journal articles and semi-scientific articles, I noticed how problematic it is that in some cases those papers use a romanized, Latin script display of originally Non-Latin script words and expressions. The issue: In my case (Korean), there were several revisions of the romanization system and some of the texts I work with are either (a) older or (b) the authors simply don't comply to the correct romanization standards. Thus it is very hard (at least for me) to derive the original non-latin script (which is my goal) from an incorrect romanization.
Additionally, in many cases it is beneficial or even crucial to have access to not only the Hangeul, but also the Hanja, i.e. Chinese characters. At least that is my opinion.
Because I assert that only if I mention the Non-Latin script (in this case Hangeul/Hanja) "as is", can my academic work be precise, exact and unambiguous.
What do you think about that? Are there any standard rules for such a situation, that are widely used (i.e. expected from students and academics), e.g. in the USA?
Finally: Would there be cases in which you would not use the romanized, Latin Script text version at all? (Assuming that we only talk about a few (5-10) words per page.)
Note: The academic results I produce won't be published in any (international) journal anytime soon, but I still would like to follow the highest standard possible, without sacrificing common sense and preciseness.
The so-called standard rules are publisher-dependent. Sorry about that. Still, we can cite the rule of one of the Alpha-1 godfathers of computer science:
I want henceforth to give authentic spellings of people's names in their ``mother tongue,'' as well as the latinized form that is conventionally used in Western books and journals [...]
(D. Knuth, http://www-cs-faculty.stanford.edu/~knuth/help.html#exotic) Knuth also maintains a list of Asian names there.