However, if it does work, I don't know how many 1's & 0's have to get scrambled before there are consequences.
If you connect the slave to the end connector and the master to the middle connector-
nothing will happen. the connection method is, pure and simple, a convention. (at least, pre ATA-133, I believe the 80-conductor cable and the various protocol and standards defined therein make it a bit more strict) neither drive can determine which connector it is in.
the actual controller on the motherboard (which, is an appropriate term, since we are dealing with multiple drives, each with their own controllers, something needs to manage the communication between the multiple devices and the single motherboard- thus the term controller, which coincidentally was also used for the add-on cards used in older PCs before on-board controllers were placed on the motherboard, you'd have an ISA card that would interface with your IDE drives. since it was in control of the I.O between all IDE devices and the motherboard the term Controller is nothing if not accurate.)
The key to knowing why you can connect either one in either spot (as long as you don't have multiples of each) is in the Jumpers.
When the computer powers on- it of course goes through it's post, where it checks everything. included in this check is the investigation of the ATA devices. the BIOS simply sends a ATA Identify command to the master, and then to the slave (each ATA command has a "channel" which defines wether it is to be accepted by the master or slave device- note that all messages sent from the motherboard to either device is received by both, so both the master and the slave receive two Ident commands. Wether the drives respond to the master or slave id commands depends on the jumper settings, which determines how the integrated electronics are to behave. (essentially, the jumper settings define a "channel" to use.)
if you have two drives jumpers to master, when the BIOS tries to identify the master drive, it receives two responses, since both drives have been set to the master "channel" (via the jumper settings). What happens depends on how the BIOS is configured to work with this- sometimes it detects whichever response arrives first- other times, it might even accept the second Master response as a response to it's later query for the slave drive configuration. Whatever the case, issues will result, especially as every single ATA command (including reads, writes, and so forth) are sent to both drives. Generally, because of differences in drive geometry, this quickly causes a hang. in practice, the most common result is the for the detection routine to hang, as the two responses are almost never contiguous, but rather "mixed together" so that the BIOS detection routine cannot properly parse it, so it waits forever for a "proper" response.
The same is true with two drives configured as slave/secondary.
With regard to using a cable "backwards"... you cannot plug it in backwards anyways. the connection for the motherboard on an ATA-133 (80-conductor) cable is keyed with a plugged hole so that it will only be able to plug into something that has no pin in that location, and drives generally have pins in all locations. For older cables that can be plugged in that way, it doesn't do anything. remember that regardless of wether you plug it in "forwards" or "backwards" the transmission length is the same.
The main, and deciding factor of why it's not done this way is simply because you cannot reach form the motherboard connection to one of the drives with the short length between the two end connectors.