For specialized applications requiring slots parallel to 0.00001" or 10 millions of an inch. We used an capacitance gage traveling the slot to measure and graph it out full size.
We used phenolic blocks sanded flat on a piece of sandpaper on a granite surface plate. To these block we used double sided tape and applied 400 grit and 600 grit sandpaper to. We change sandpaper often. We could easily use micron paper 10 to 100 times finer and get a surface that you can see your reflection in if we wanted too. In fact the reflection of zebra striped lines and checking the reflection quality is also a way of checking surface quality. Also we work in a temperature controlled clean room as dust can seriously effect readings.
Now the surfaces we got were just short of a mirror finish. When assembling the equipment even fingerprints and even drops of water will affect the slot width. I am guessing this is on the order of 10 to 100 times finer than scraping.
take you scraped block and measure it with a gage that can show 10 millions. you will see how flat it is compared to sanding. also we were working with titanium as iron and steel would loose surface quality due to oxidation within 24 hours.
now the true test of gage block quality is when clean, you can wring them together and they will stick to one another. I believe gage blocks are lapped similiar to making glass lenses by an abrasion process.
do your scrapped objects stick to one another ?