Queru: Another year of AOSP [LWN.net]

Jean-Baptiste Queru reflects on the ups and downs of the Android Open Source Project in 2011. "Not releasing the Honeycomb source code was catastrophic for the AOSP community. I had never before received so many angry emails, so many threats, to the point where I had to take several weeks off at some point to get away from it. Even today, there's a lot of bitterness left on all sides. From start to finish, Honeycomb probably cost AOSP anywhere from 6 to 12 months."

to post comments

Queru: Another year of AOSP

Posted Dec 30, 2011 18:27 UTC (Fri) by gnu (guest, #65) [Link] (6 responses)

There is no mention of the number of external contributions merged into the AOSP (in the list of positives, where I expected it to be).

Queru: Another year of AOSP

Posted Dec 30, 2011 19:51 UTC (Fri) by Shachar (guest, #67086) [Link] (5 responses)

That's because there hardly were any.

In the beginning of the year, that was because Honeycomb diverged from Gingerbread enough that any attempt to submit a patch to the core OS would result in "conflict with Google's internal tree". Usually that clears up as soon as there is a code drop to AOSP, but for Honeycomb, that never happened. By the time Ice Cream Sandwich was released to AOSP, Gerrit, the patch review system used by Google, was down with the rest of the services that used to reside on kernel.org. To date, it has not been restored.

This means that since about January, and still ongoing, contributing to AOSP became more and more impractical, until the kernel.org fiasco, where it became downright impossible.

Shachar

Queru: Another year of AOSP

Posted Dec 31, 2011 15:00 UTC (Sat) by rsidd (subscriber, #2582) [Link] (4 responses)

Why was Google using kernel.org for critical Android infrastructure? They certainly have the resources to host it themselves. Was it some sort of misguided goodwill gesture? If so, why didn't they at least ensure that everything was sufficiently backed up that they could set up an alternate server immediately if required?

I agree nobody would have foreseen this sort of outage on kernel.org. Still, for a company like Google to use other (free) resources rather than their own seems strange.

Queru: Another year of AOSP

Posted Dec 31, 2011 15:20 UTC (Sat) by janschulz (guest, #82047) [Link] (3 responses)

Because the kernel.org infrastructure and admins are quite good at what they do: https://plus.google.com/112413860260589530492/posts/EJCsV...

Queru: Another year of AOSP

Posted Dec 31, 2011 21:14 UTC (Sat) by drag (guest, #31333) [Link]

That's pretty awesome

Queru: Another year of AOSP

Posted Jan 1, 2012 3:39 UTC (Sun) by rsidd (subscriber, #2582) [Link] (1 responses)

Very interesting. Still, the kernel.org people were volunteers and it seems to me Google had enough expertise to do this on their own. Maybe they did substantially help out kernel.org in the process, so it was a win on both sides?

It will be a pity if the combination of closed honeycomb and kernel.org outage really has reduced outside involvement in AOSP to such a long-term extent as this. On the other hand, hopefully these things - the honeycomb mistake in particular - will not occur again.

I also found it odd that Queru has 'given up' on Motorola, which Google purchased a few months ago.

Queru: Another year of AOSP

Posted Jan 3, 2012 5:40 UTC (Tue) by dtlin (subscriber, #36537) [Link]

Google announced its intent to buy Motorola several months ago. The actual purchase, if it happens, would be several more months in the future – the deal has yet to pass regulatory review in the US and EU.

Queru: Another year of AOSP

Posted Dec 31, 2011 8:51 UTC (Sat) by TRS-80 (guest, #1804) [Link] (6 responses)

In Galaxy Nexus, we also have a wifi/bluetooth chip that can work without requiring any proprietary firmware.

Now that's a big achievement.

Queru: Another year of AOSP

Posted Jan 2, 2012 11:14 UTC (Mon) by dufkaf (guest, #10358) [Link] (5 responses)

Well, does it mean that the wifi/bt firmware blob is simply embedded in the chip (~= not upgradable) or that the firmware code is open?

Queru: Another year of AOSP

Posted Jan 2, 2012 20:53 UTC (Mon) by zlynx (guest, #2285) [Link] (4 responses)

Does anyone actually care?

The hardware design is not modifiable or open source either.

Queru: Another year of AOSP

Posted Jan 3, 2012 0:13 UTC (Tue) by TRS-80 (guest, #1804) [Link] (1 responses)

Presumably the OpenMoko folks would care, as it would mean they could abandon their ridiculous suprerfluous proprietary firmware ROM chip.

Queru: Another year of AOSP

Posted Jan 3, 2012 0:29 UTC (Tue) by neilbrown (subscriber, #359) [Link]

There is nothing to abandon.

Please see: http://lists.goldelico.com/pipermail/gta04-owner/2011-Dec...

> What you are probably referring to was a proposal by RMS/FSF to
> isolate that in hardware for a special variant that FSF could endorse
> (maybe with a different name).
>
> There did no appear a volunteer to build a prototype and demonstrate
> that it is working, useful and improves freedom at all.
>
> Without a prototype, we can't even think about adding such a thing to
> hardware and give it to FSF for promotion.

Apparently it wasn't the "OpenMoko folks" but rather the "FSF folks" that were behind this.

Queru: Another year of AOSP

Posted Jan 3, 2012 17:44 UTC (Tue) by wookey (guest, #5501) [Link] (1 responses)

Of course people care. Firmware is code too, and it's often quite broken. Being able to fix it would be useful.

The graphics situation is much worse than simple firmware blobs. There are great dollops of proprietary code running on the main CPU making kernel upgrades and porting difficult/impossible, and removing the opportunity to fix or improve or re-use stuff.

All the manufacturers have got themselves into a right mess, scared of patents, even though their infringement is exactly the same whatever licence the code uses - either they are infringing or they aren't (they are of course, obviously, because the whole are is rife with idiotic patents).

Some believe that their graphics drivers really are so good that no-one outside could help improve things. Which is an odd thing to believe as we know how hard it is to employ all the smart people in any given area.

They've all got this weird idea that publishing CPU instruction sets and registers so that people can write compilers and optimised code and understand how their device works is good, but publishing GPU instruction sets so that people can do the same things is wrong and bad and must be discouraged. his seems to me to be demonstrable nonsense, but I'm not quite sure how we fix it.

Queru: Another year of AOSP

Posted Jan 3, 2012 20:38 UTC (Tue) by zlynx (guest, #2285) [Link]

Well, one reason the graphics companies hate publishing their details is that they hate to support customers. They'd rather not deal with them at all.

If the CPU and register sets were published, they would be under the gun to support that. If their CPU instructions had a flaw, like the Pentium division problem, they'd have to fix it.

But when everything has to funnel through their drivers they can just release a software update to silently rewrite flawed operations. Its a miracle! No need for expensive hardware recalls!

It also makes it very difficult to reverse engineer their driver code. Are they cheating at a benchmark? Are they silently speeding up a game by only running particular shaders every third frame? Who can tell?

Look at all the crap Intel got because their compiler used the CPUID check for GenuineIntel to enable proper SSE optimizations. No one would have ever known if you couldn't read the instruction set.