In the earliest days of computers, there was some debate about whether the Most Significant Bit should be sent first, with successive "less significant bits" following until the byte was completed, or whether the Least Significant Bit should come first with increasingly significant bits following.

Most of the chip manufacturers followed one side, but a few others, notably Motorola, went the other way.

Apple was one of the computer builders who went with Motorola for their processors, and hence were - for a decade or two - mostly incompatible with computers (in the real world) that used the opposite system.

The computer must load a complete byte into the processor (or at least into the memory that feeds the processor) before anything can be done with the byte, but of course once it's read the (then eight bits) byte, if it reads them "backward" it gets a different result than if it reads them in the oppositely backwards order.

It's a fairly simple process to "flop" each byte (swap end for end) if you know which direction it's supposed to go, but it took quite a while before the ability was incorporated into a majority of computers.

The HTML standard(s) include bit order markers that can be used to tell computers that attempt to read pages how to interpret the bit-stream correctly. The markers are more frequently now, I believe, included in the CSS (style sheet) that the computer should see before it even starts to "translate" a page into a display form, but there are occasional omissions in what gets on the web, or sometimes all the "details" don't get loaded properly. It should be possible for the computer to determine which bit order is being used without reading a bit order marker but that process seems to be less than perfect.

The bit order markers are supposed to be "unprintable" so that they're not displayed in the web page, and the character string indicated is actually the "invisible marker read backward" by a computer that starts posting the page before it recognizes that it is a bit order marker and that it's reading the bits backward.

As long as the correct bit order is determined before the html interpreter gets far past the marker, no information is lost, since what it printed backward was supposed to be invisible anyway.

So don't worry 'bout it.

All this might have been avoided if Motorola (and a very few others) hadn't insisted on mounting their processors upside down, or if Macs had been popular enough so that we'd all have ended up upside down like they were. [t.i.c.]

John