This probably goes beyond the scope of your course:

Directly addressing screen positions like you do here depends on features a certain terminal type you use may (or may not) offer.

The UNIX terminal concept comes from times where "terminal" was a roll of endless paper running through a printer. It is easy to produce output in a stream-like fashion (just adding at the end), but the printer will have a hard time rewinding the paper and then print over parts it has already printed on. In fact this is what you are trying to do with all the "tput"-statements.

Before you even start debugging your code you might want to find out what terminal type you use:

Code:

echo $TERM

Terminals are handled in a specific way: every terminal type has certain capabilities and ways to use these capabilities. Usually these are some control code sequences which move the cursor, switch printing colours, etc.. All these capabilities are described in a database called "TERMCAP" (terminal capabilties) where the name of the terminal type is the primary key.

When you use a command like tput to position the cursor it first inspects the content of your TERM variable, getting, say, "footerm". Now it queries the termcap database for what the command sequence for moving the cursor in a "footerm" is, getting some control sequence as answer. tput now issues these control sequences and believes the positioning command to have taken place.

This is why you usually get a lot of garbage when you set TERM to a wrong value: tput will get control sequences which may mean nothing to your terminal type and are therefore displayed "as is".

After this little excursion back to your problem: if you want to program input/output masks like you seem to do then shell code is not the tool made with this in mind. This means: it is not impossible, but cumbersome and will not work as reliable as the things for which the shell is made. You might consider "curses" and/or similar libraries to do such things.

I hope this helps.

bakunin