Having to deal with this issue, again, lately gave me the idea for this post. Ever buy a radio programming cable, or a cheap USB-Serial adapter, only to plug it in and have Windows tell you that there was an error installing the device? Typically, when you go into the device properties it tells you that it "Cannot start device (code 10)".
Prolific is the original manufacturers of the PL-2303 chip. It is a single chip device that plugs into a usb port on one end and gives you a normal RS-232C serial port on the other. The chip got popular enough that other Chinese manufacturers cloned it. I'm not sure if they cloned the exact chip, or if they made one that does the same function but uses the same protocol as the PL-2303 chip and uses the same drivers. At first, when these chips came out the version of drivers which existed would work 100% with these clone chips.
Prolific didn't like this however. They have a good reason I'm sure. Since clone makers tend to ride on the coattails of the authentic chips, the authentic manufacturer gets to deal with people coming to them when things don't work. So it's an issue of supporting chips that you didn't make and you didn't profit off of. What they did, I've read, is add a check into the newer versions of the drivers so that they check for a specific response from the chips. If it responds with 1 code then it's authentic and the drivers work, if it responds with another code then they know it's not authentic and the drivers give you the oblique error message.
It's been awhile since they changed the drivers to add this check in, so it's rather hard to find older drivers that work in Windows for your new serial adapter.
Prolific has a good reason for doing this, but it really only hurts the end user who buys a device that doesn't have an authentic chip in it. It's not like the end user can check to make sure it does, or even knows that it might not have an authentic chip before they buy it. Some locations to buy these are also not receptive to returns either. It's possible that someone could lose some money buying a cable that they don't have the expertise to fix. Moving up the chain, it's even possible for the people who are designing the cables to errantly use unauthentic chips even if they think they are buying the real deal. Some of the older drivers do still exist that work with the clone chips, and I'm willing to bet that the Linux drivers are not hobbled in the same manner.
Given all of the difficulties with these chips, why can't an alternative show up to replace them in designs? I think even if someone could produce an open-source adapter design using an inexpensive PIC chip or similar with USB support then that would benefit a lot of people. If I find a list of alternative chips later, I will append it to this post. Maybe an open-source project has already been accomplished if this page about LUFA is any indication.
Shameless plea to radio manufacturers: Please build USB support into radios. Please? It would be beyond great to have a micro-usb port on one side of the radio that could accept a standard usb cable for programming. (Trickle charging would be a good benefit as well, but I won't push it.) Seriously, I have separate, incompatible, cables for my new Baofeng, my Icom R3, my Yaesu FT-50R, a Puxing PX-2R and I think one or two older radios floating around. In the future, I should be able to get all new radios with a micro-usb port and just use the same cable as I do for my phone to plug into my computer for programming.