Why do we have cross-over cables.

You're either trolling, or extremely stupid. Either way, the solution is:

She IS stupid .

It would surely be much nicer if equipment at both ends of the link used just a single pair to communicate in both directions, full duplex, too. If the plain old telephone system can do it, why not data communications equipment?

No distinction between DCE or DTE, only one single pair of wire, only one type of cheap cable, cheap and simple connectors, no fuss, no confusion.

Maybe it's this way because it was designed by a committee?

Jeroen Belleman

Jeroen Belleman wrote:.

No, because it was designed to work, and meet the requirements at the time.

Of course it works. That's not the point. The question was about why we don't come up with an interconnect that doesn't require cross-over cables. It's so nice, so simple.

As for requirements; We all know that 'requirements' are really a mere description of the way the prototype behaved the day we got it to work. ;-)

Jeroen Belleman (Standards are merely intended to cripple competition.)

Consider the speed limitations of such a system.

Even fiber optic transmission (transception)has two paths.

I'll bet that it has to do with the speed of the wire limitations at the time.

This is a question for Floyd Davidson.

No. The original post asked why we don't have an interconnect that _only_ uses cross-over cables. The question has been answered on both the historical basis of a clear distinction between "transmitting" and "receiving" equipment and ease of manufacture of straight through cables.

The modern trend has been to use a bidirectional interface such as USB but this is necessarily slower than a similar interface using three wires instead of two and a cross-over cable. Nobody has actually answered the question in the current context.

Ah yes, that was indeed the question.

I suppose you are thinking of half-duplex traffic. It's quite possible to have full-duplex traffic on a single pair of wires. There would then be no speed penalty.

Jeroen Belleman

Except that, to have full duples on a single set of wires, you are assuming that one or both directions need to be modulated, with the added expense and costs of that modulation in the interface.

Charlie

Actually, there WOULD then be a speed issue.

RS485?

RS232 was the solution to a different problem (how to conect modems to terminals)

Not at all. I was thinking of hybrid transformers to separate outgoing data from incoming. The extra cost should negligible if the interface already uses a transformer to couple to the wire anyway.

Jeroen Belleman

Oh, that's very helpful. Would you mind motivating that statement just a little?

Jeroen Belleman

I have seen a few or more implementations of full duplex over a single pair or coax. It is not even that difficult. I don't suppose you have heard of a hybrid transformer? Echo cancellation is rather more difficult. .

Just because you have not seen bidirectional full duplex on a single fiber does not mean that it cannot exist. I have used such devices. They typically use wavelength division multiplexing. .

Hybrid transformers are just like a circulator, they reduce the energy coming from the transmitter site back into a receiver at the same site, but they are not magic! A lot of the energy still gets through. This is fine, if you are using voice, as speech communication tends to be simplex - you stop speaking when the other party is talking. But if you want full duplex data, then the modulcation schemes between the two channels has to be different. Modems used different frequencies for uplink and downlink, fiber can used different wavelengths for the two directions, but both require modulation of the data.

Charlie

Again, yes, you can do full duplex, but it does require different modulation schemes between the two directions. If you just try to apply digital data to a wire, the wire doesn't know which direction it came from!

Charlie