One of the most incredibly standardized things in the history of computing is the Plain Old Serial Port. These little buggers haven’t changed detectably since 1990 or earlier.
I recently picked up a serial port card on the grounds that it should “just work” on Mac OS X. It almost does. It’s a standard part which shows up as a normal serial card. One slight quirk; it claims to support 115,200 bps (the measure of speed for a serial port: bits per second), but the documentation all talks about 920kbps – about 8 times as fast.
Well, that’s the thing. Obviously, the vendor (siig; www.siig.com) has done something weird to make this card run at different baud rates. The closest I’ve gotten to a successful match is that, if I configure it for 57,600bps, no parity, it comes up at something very close to 76,800bps, with space parity. Which makes no sense at all.
I had once written them to ask whether they had documentation available for what is, after all, someone else’s chip with a sticker on it. They said no. Now I think I see why; this thing has been carefully made subtly incompatible, so it can ONLY be used with their proprietary Windows-only drivers on Windows machines only. No Unix, no Mac OS X, no NetBSD, no Linux. In short, none of the people who have any reason at all to buy a 4-port serial card can actually use this one without documentation Siig has, thus far, refused to provide.
I don’t get it. They’re a hardware company. Why, I wonder, would a hardware company try so hard to make people use their software? Do they not understand that following the specs so that regular users can just use the hardware without hassle is cheaper and better for everybody? Why do they want to treat as a proprietary trade secret the fact that they used a perfectly standard chip off the shelf and configured it funny? Why not just let people know how it’s configured, and use it?
Date: 2005-06-20 17:01:43 -0500
A wise guy once said: Almost every question that starts with
why' can be answered withBecause people are stupid’.