When you publish academically, you clearly define the abbreviation on first use, and then you use it freely. If you only use it once, there's no need to abbreviate it. If you use a non-standard abbreviation, your peer referees will likely ask you to change it. If you use something really non-standard, it will probably be changed before proof by the page editor to match a style sheet.
Speaking of Style Sheets, IEEE has one, and it has a section on abbreviations.
Acronyms and abbreviations can be used to save time and space in the document. If the draft makes
extensive use of acronyms or abbreviations, a subclause within the definitions clause entitled “Acronyms
and abbreviations” may be provided. The acronyms and abbreviations subclause is not meant to take the
place of the definitions clause. If acronyms and abbreviations are included in the definitions clause,
Clause 3 should be titled “Definitions, acronyms, and abbreviations” and 3.1 and 3.2 titled “Definitions”
and “Acronyms and abbreviations,” respectively.
Within text, the acronym or abbreviation should follow the first use of the full term (the first time in the
introduction, then the first time in the body of the document, and then the first time in any annexes in which
the acronym appears). The abbreviation or acronym should be placed in parentheses when following the
Acronyms and abbreviations, followed by the full term only, should be listed in alphanumeric order. For an
example of an acronyms and abbreviations subclause, see the sample draft in Annex B.
Exceptions to the convention listed above are approved SI units. SI unit symbols are not abbreviations and
shall not be included in a list of abbreviations and acronyms. The treatment of letter symbols for units (e.g.,
mm for millimeter), letter symbols for quantities (e.g., R for resistance), and mathematical symbols (e.g.,
log for logarithm) is covered in IEEE Std 260.1 and IEEE Std 280 (see also Clause 14).
Acronyms and abbreviations should be avoided in titles of standards. However, if such use is warranted, the
procedure stated above should be followed.
Personally, I get the O/P I/P thing, but its usually a "what does that mean?" process to get it figured out, and I'd eventually arrive at the right answer -- I guess I need some basic contextual reading lessons (I suppose I can use that as an excuse to turn down reviewing grants when the NIH asks me to). If I were to edit the example question, I would probably define it (i.e., "output (O/P)" ) if it were used more than once; I would probably replace it with the complete word if it were used only once; and I would probably remove it from the title (where there isn't much context).