Couldn't the same sort of mass-lawsuit strategy recently used by certain copyright holders work here? I mean, if it's really this widespread I could see the FSF getting a court to issue an injunction against these devices being sold, given the quantity of infringement going on (especially given that these are companies who should have the resources to do so.)
It's certainly plausible against the companies that aren't based in China, or resellers of those products located outside of China.
Although only certain OSS rightsholders have been keen to pursue that avenue in the past. Happily, at least two of the products on my list contain busybox, and busybox have generally been keen to defend their rights in court, so there is clear practical scope for this kind of thing.
It strikes me that the most important part is the footnote:
"I have an email from VIA saying they will be proactively releasing the GPL source code used in Wondermedia WM8505-based products like the Eken & Gome tablets."
If all these tablets (and phones) are based on a handful of chipsets then getting that lower level code straight from the chipset manufacturer should be a very important step in getting 3rd-party community distros onto the devices, much like getting Ubuntu etc. onto an Atom based netbook. You'd still have to watch out for SD card readers, touchscreens etc. but it's a great start.
I am more sad about the current state of android tablets. Are there any decent tablets available? When I looked recently, I found a bunch of announcements for stuff that was coming, like the Dell Streak. But the actual tablets available look like knockoff Chinese junk. The Archos tablets seem to be the only moderately respectable devices in the bunch, and those seem to have their share of problems.
The Archos tablet is the same as the Rockchip aPad that is known under many names: iWeb, iRobot, etc. I am doing a sort-of custom rom for the device and it's really hard to do without the sources. I don't really understand the chinese companies, though. Why would they not want people working for them, for free?
Is this true? Because the Archos 5 Internet Tablet source seems to be on their site (I originally didn't find it because the naming is kind of weird, but it's there!)
I have a Dell Streak eval currently. It's an extremely nice and well designed device with a beautiful screen and great battery life. The only knock is that it's running Android 1.6, but that will supposedly be remedied soon.
Really, I can't say enough about this device. It's great and far more portable than my iPad.
The underlying kernel for Android is licensed under GPL, but the rest of Android is licensed with the more more permissive Apache license.
My understanding is that Google explicitly chose the Apache license because it is considered as more friendly to commercial development and proprietary redistribution.
I'm assuming that the manufacturers will state that they have just extended Android non-kernel components or (if kernel changes were required,) perhaps that they've just implemented a new I/O abstraction layer that is independent of the GPLd code
Followup to response: If my comments reflect a misconception, can you please provide an alternate hypothesis for why Google chose the Apache license for Android
This is a common misconception about GPL, but it is false.
To oversimplify Section 3 of the GPLv2: if you release a GPL licensed program then you have to comply by making source code available or showing where source is available, regardless of whether or not you have modified it.
I can assure you that Linux BSPs for most of the underlying systems included in the list are not currently available anywhere (that's what the "source available" column in the table is intending to show.) Although this is actually unrelated to the issue of whether or not the company themselves are violating GPL.
You're correct that (as the article says) quite a lot of functionality is spun off into standalone kernel modules (not necessarily GPL) and the Android layer, but this doesn't excuse the overall GPL violation.
"Followup to response: If my comments reflect a misconception, can you please provide an alternate hypothesis for why Google chose the Apache license for Android"
Of course - you're entirely right that's why Google chose the Apache license for Android. In fact, people from Google have said that and my article says that as well.
That still only applies above kernel level. The GPL violations I am talking about, as described in the article, are failure to comply with GPL for the Linux kernel and for other GPL licensed components that vendors have chosen to include (I identified busybox in at least two of the listed firmwares.)
You are probably right that many modifications are not covered by GPL. However, the vendor still has to comply for the portions that _are_ covered by GPL (even if they made no changes.) In the case of many of these systems, that happens to include the BSP (Board Support Package) that allows Linux to boot on those systems at all. Although that last part isn't the point, at the moment it's a fundamental practical problem.
>if you release a GPL licensed program then you have to comply by making source code available or showing where source is available, regardless of whether or not you have modified it.
This seems pretty arbitrary. If I haven't modified it can't I just point you to the repository and revision I'm using?
Pretty much. GPL 3(c) - I think - says that provided you got a binary version with an attached notice telling you where to go to get the source code, you can redistribute the same binary unmodified with the same notice.
