What I am reading is that the sector in question is physically damaged. A discontinuity in the matrix, perhaps.
Forgive me, it's been a few decades since I did hard drive repair, so my knowledge of disk operating systems is archaic and rudimentary at best. So this is my understanding from my work in the late eighties and early nineties.
As I recall, the system will normally mark those damaged sectors as offline because they are uncorrectable. The read/write software ignores the sector. That is, it does not attempt to write to the sector, nor does it try to read it (though I suppose you could direct a read from that specific sector, as I recall some early DRM attempts used specific errors at specific sectors to prevent copying floppy discs).
Now, back in the bad old days of 20MB hard drives, when defragmentation was a very necessary thing, we also had problems with--let's call it drift. The actual tracks on the platter were quite large compared to what is done now, and writing data to the sectors not as good for various reasons. We would have to refresh the data on the disk periodically (not like after so many days, but more like months), and that problem could extend into the actual borders that defined the tracks which contain the sectors. Suddenly, the drive would begin filling up with bad, damaged, or unrepairable sectors to the point that it is useless.
Sometimes, you could restore functionality to one of those drives by performing a low-level format, which writes those track lanes into the platter again, redefining the borders and identifying each sector for the drive map. Then after that format is performed, the system would do a read/write to each sector and determine whether the sector works or not.
Back then we often had to manually define the drive to a computers BIOS, telling it how many platters, tracks and sectors it held rather then having it automatically detect that information form the drive itself. Those BIOSes had a tendency to direct operations itself rather than rely on the drives own hardware/firmware, and you could get some really weird errors popping up because someone didn't define the drive correctly.
Nowadays, you never have to do low-level formats: the platters are more homogeneous, the sectors are smaller and better defined, the heads closer to the matrix, and the firmware is well-written. Usually, anything that will cause the number of bad sectors to dramatically increase is a catastrophic failure, usually in the hardware itself, and SMART oftentimes identifies these long before the drive completely dies.
Usually. I have recently suffered a hard failure that took out not only my laptop drive, but also the external drive I was backing up to; which is a good reminder to regularly backup and archive any data you want to keep, and have multiple copies in different places.
Forgive my rambling: tl;dr-- an offline sector is normally not read from nor written to.
