The mark is used very often as a means to go back to a previously stored position. For this use it is often a restriction that there is only a single mark per buffer. Therefore 10 tags have been introduced. Tags are somewhat loosely organized marks that are not bound to a specific buffer. When a tag is placed it is bound to the current buffer and the current column in the current line. It is then possible to go back to that position from any buffer with the ># command in the command line (# stands for a single digit of the normal keyboard). The tag is placed with the <# command.
Once a tag is placed it will survive anything the editor may do internally (like garbage collections), but when the user changes the line to which the tag is bound this may make it unbound and pointing to a fictitious address. If the user wants to go to this tag afterwards there are two possibilities: Either the tag points now to a different line in the same buffer as the original one (this could be caused by garbage collections moving another line in place) in which case this is considered to be the tag position, or the above is not the case and the message 'tag not found' will be displayed. This is the price to be paid for a significantly smoother performance than when the tags would be arranged like the marks. In practice the above `erroneous jumping' occurs very rarely, although this may depend on the habits of the user.
When a tag is hidden inside a closed fold stedi will also not be able to find the tag and go to its position so it will also report 'tag not found'.