There was an article posted on HN a while ago about certain Xerox scanners that would change numbers on scanned documents (turn 6's into 8's, for example). So this is just a follow-up.
Huge problem: The Xerox scanner software is doing fancy (but broken) AI/OCR-type stuff to compress images that sometimes replaces a perfectly normal number with different perfectly normal looking number in a way that defies belief.
Once again, Xerox fails PR. Spend half your first paragraph blaming the customer and saying their concerns are not important, what could possibly go wrong?
I continue to be amazed by their response to this serious problem.
'when scanning “stress documents” to PDF—which can include very small font sizes, stray pixels and be difficult to read'
That's some impressive wording there. It makes it sound like only documents which have very small font sizes and are difficult to read are affected, but it actually says the exact opposite - that the issue can affect documents with large, clear fonts and no stray pixels.
6 comments
[ 4.4 ms ] story [ 25.8 ms ] threadI continue to be amazed by their response to this serious problem.
That's some impressive wording there. It makes it sound like only documents which have very small font sizes and are difficult to read are affected, but it actually says the exact opposite - that the issue can affect documents with large, clear fonts and no stray pixels.