Here is the original message. It has the comment "/* Attempt to set Vendor ID to 0 */". So yeah, they are intentionally fucking with a chip when it fails to validate. And in addition to fucking over buyers of equipment where the manufacturer may have unknowingly been given counterfeit parts, they've also told the cloners exactly what to change for their next run of chips.

Wow, just WTF. It's one thing for them to claim some loss, no matter how slight, from people leeching off of their Windows driver. But considering that the clones do not copy FTDI silicon (have ANY of them been found to do so?), and they have absolutely no claim to ownership of the Linux kernel driver, this is just greed at its worst. Also, not all clones have counterfeit labeling on the chip and can thus be considered fair competition. I wouldn't be surprised if some are even in package types that FTDI doesn't sell. Their driver may see their 16-bit VID number on the chip (you can't trademark a number, that's why Intel renamed the 586 as "Pentium"), but it can't see whether FTDI is etched on the chip or not.

Or maybe someone can point me to something that says you can patent a register layout and chip pinout. (essentially the hardware equivalent of software APIs) Except again, there is no way that the driver can even know that the chip uses the same pinout.

Now maybe if they had the chip return the text "FTDI" (aka actual trademark-able text) and checked for that along with some other kind of "real chip" test... but that still won't justify fucking with the chip. Just refuse to run is all you need.

Except for the little fact that the "IP" probably isn't theirs. While there may be trademark violations, both with the VID, and some (but not all) chips having counterfeit labeling, there isn't much else. Sure, they're register compatible (no problem there, that's just like a software API), and illicitly using the FTDI drivers, but the silicon itself generally contains no FTDI IP. And there is no way for the driver to see what is engraved on the chip, just that it isn't 100% bug-compatible.

However, I expect to see this cause some changes in the clone chips. The most obvious one is "don't allow the PID to be changed to zero". Or they may move to cloning another type of chip other than FTDI and Prolific. All it takes is for one smart guy in China to come up with a new design, and suddenly everyone is using it.

This already happens more or less with pay-at-the-pump gas. Do you want a car wash? (Y/N) Do you want the special engine cleaning additive? [PLAY JINGLE] (Y/N) Do you want a receipt? (Y/N) (That last one is always fun when you DO want the receipt, but have been hitting "no" to all the stupid upsell options.)

I'm guessing you've never had the pleasure of repairing an "aluminum"-era MBP? The case design that started back in the PPC era was flimsy as shit. Compared to that, the current models are built like tanks. And I also had a Pismo-era PowerBook, which was flimsier than that.

One major problem was that the optical drive would get out of alignment with the slot in front, and it would be unable to eject discs. Another problem was that the latch wouldn't close because dust or something clogged the little latch thingies. And then there was the surface treatment of the aluminum. There was such a wonderful pattern of pitting where the palms of my hands rested on the case. I had one PPC and two Intel of those, and they were all bad in the same way. I've had a unibody MBP for over two years now (the last of the 17" models, which I promptly downgraded to 10.6, and that isn't easy), and it's just fine, no pitting. And the lid never fails to close. The only place where the case is deformed is over the expresscard slot, which you can tell is a bit sagged if you run your finger along the edge.

I've also fixed my cousin's MBP (I think it's a 2011 15"), which somehow got dropped on the corner with the battery, so a couple of the battery cells popped up. I was able to replace it with a 3rd-party battery and a new hard drive (to replace the aging hard drive that had started to fail, the reason he needed me to fix it). Still built like a tank compared to the older aluminum generation.

But I'll agree with you about the current iMac generation. Not that I would touch an iMac (I always want a separate monitor with a desktop system), but holy crap the display is actually less replaceable than a laptop.

The "specs" on RAM limits usually under-represent the maximum possible. The reason is that when the specs are released, the chip sizes needed for that maximum likely do not exist, and Apple doesn't want to advertise something it can't test. If you check the lowendmac page (assuming I found the right one), it says there's a 16GB limit.

I vaguely recall that the reason Apple is pretty strict about this is because of the Mac SE/30, which didn't have 32-bit clean ROMs, limiting it to 8MB. The physical limitation was 32MB with 4MB simms, or 128MB with the very rare 16MB simms. (16MB simms were very expensive when new, the most memory you can put on a 30-pin simm, and only came out right before everyone switched to dimms.) They bought out the Mode32 product from Connectix, rather than produce a new ROM module, to avoid a class-action lawsuit. (FWIW, you can install a Mac IIfx ROM into an SE/30, removing this limit. If you ever find a IIfx, putting its ROM in an SE/30 is a better idea than trying to upgrade IIfx's unusual RAM.)

But... but... but... you can't get Macs for $100 at the place that sells used off-lease Dells!

The best way to counteract the Slashdot Effect is to do what Slashdot did and make a horribly unusable beta.

And I'm in northwest Austin near 183/620, so I share your pain. There's quite a few tech companies along Parmer in general. The dots on the map (with a few exceptions that may be the "special" places they wired like libraries, etc.) stop at 38th and Mopac, except for the Mueller area. I'm actually kind of surprised how they put so many blue dots east of 35, but then I guess the hipsters have been gentrifying the poor out of the east side lately. Lots of blue dots in downtown, too.

Since Google does a "fiberhood" at a time, it may be behind your house, but not behind every house in your neighborhood. Also, if the fiber is a backbone, you don't connect a customer to that. You can't just stick a drop in the middle of a fiber cable. As I understand it, they put priority on neighborhoods with the highest demand and the lowest construction issues.

Also, if you don't get it when they come around, they aren't bringing the backhoe back when you change your mind later. I think it's worth it to pay $300 for them to wire up your house. By the time they get to my house, I'll probably have moved back to San Antonio and be renting it out, and I'd pay the $300 just to be sure it gets installed to the house.

Cherry picking. Partly they're going for neighborhoods with higher anticipated demand (and higher anticipated density), partly they're going for neighborhoods with physically easier installs. Just remember that when they choose your neighborhood (aka "fiberhood"), if you don't get the install done then, you're pretty much fucked until they pick your neighborhood for again a second chance. My understanding is they generally won't go back for the stragglers and wafflers who didn't sign up when they were doing that neighborhood, at least not if it means a backhoe roll would be involved.

So he didn't just want to be on the bleeding edge, he also wanted to be the first guinea pig to do it with Canadian plastic money? Sounds like we have a diligent idiot here.

let's replace it all with emacs