I've used dd hundreds of times and still do not understand it.

In theory (at least my theory) is that a stream multiplier should copy information bit for bit so if I want to make a boot flash image from an iso file it seems that all that would be required would be

I've successfully done it in the past but sometimes it doesn't work

I remember last time I gave up hope I was installing fedora and eventually used a windows based gui program distributed by fedora. After closer examination I realized it was just using a windows port of dd. However they recommended first formatting the drive as fat16 before running the program.

Now, today I already tried creating sdb1 as a fat 16 partition and then using dd. No luck.

Honestly, why would you need to do that? I don't know a whole like about low level drive mechanics, but shouldn't dd write whatever bits are necessary to encapsulate partition, thereby creating partitions itself?

If anybody could tell me what preparation I should be making on the flash drive before running dd I would greatly appreciate it.
Any other technical explanations would be welcome as well.

No ideas?

well I figured it out anyway... Sort of

based on a thread I found I tried the following line without any prior prep to the drive

So the question is why did it work. "noerror" is a debugging message"
It does seem sort of logical to make writes occur slower than reads.... Sort of, actually I'm not convinced this has a whole lot of effect. If you were slowing down writes perhaps that would make more sense but default write is 512 or 1b

I'm thinking the difference would be the "notrunc" flag. For no other reason than I exhausted all other options. I don't really see why any modification to an offset somewhere in the middle of the drive would affect the MBR

Any ideas?