Further up the article, the author says that nothing came close to its speed, but when it was removed, no tears were shed - a bad smell and poor legibility.
I think the lesson there is that performance isn't everything. It was the fastest printer in the world, but nobody was sad to see it get replaced with slower, but less smelly / more readable printouts.
I dunno. I had personal experience watching an IBM 1403 printer when the carriage tape (which told it such things as where the top of the page was) broke. It could eject an entire box of blank fanfold paper very quickly indeed.
In the mid-late 70s our mainframe had a high speed line printer - it was driven by 120 hammers and a spinning belt of letters, the hammers hit the belt when the right character came past. People printing just cheques might put on a belt with just numbers and a few special characters, it would print faster. The metal belts were known to break, and embed themselves in the wall next to the printer, you didn't tarry when you walked past it.
The page size was determined by a loop of paper tape maybe 12 holes wide, it was as long as the number of lines on a page (paper had sprocket holes) if you put different paper on the printer (or checks) you put on a different tape loop. If you remember fortran carriage control a line that starts with '0' skips to the start of the next page - on this printer N at the beginning of a lines skipped until a punched hole in the Nth column of the tape loop came by (0 was by convention punched just once) ...... if the operators put a blank tape loop up, or had a column with no holes anywhere and someone skipped to it the printer skipped forever - paper came out horizontally and hit the wall rather than folding nicely.
Paper came in boxes, usually with silverfish, which we jokingly said were trained to eat the sprocket holes. The narrow paper made the very best paper airplanes
