[qcc] TODO?

Rob Landley rob at landley.net
Wed Oct 21 18:53:34 PDT 2015



On 10/21/2015 08:39 AM, Sean Lynch wrote:
> 
> On Wed, Oct 21, 2015, 04:20 Рысь <lynx at sibserver.ru
> <mailto:lynx at sibserver.ru>> wrote:
> 
>     You're talking about android like it's The Only Future and we can't
>     escape. I wonder why such a hype to a corporation-influenced binary OS.
> 
> Rob is talking about extracting open-source code from an open-source
> distribution of Android.

The use case I keep thinking of is an 8 year old girl in rural india who
inherits her mother's old phone and an old HDTV, and fishes a cheap
chinese USB hub, keyboard, and mouse out of the trash. It should be
possible for her to learn programming using that phone as a development
environment.

The thing is, it already _is_... if you call javascript programming. If
you accept that system programmers are magic and special and userspace
programming is a different category, then it's easier to learn
programming than ever. If you accept that the billions of devices that
will wind up in drawers and landfills in 3-5 years and thus accessable
to poor people and children as "their device" lock you out of systems
programming and only let you make apps...

Me, I think we should still be able to do systems programming on these
things. It should be possible to use a phone as a development
environment. This is not postulating a phone is an ideal development
environment, this is acknowledging that these are the devices that have
the unit volume and thus have the best price/performance ratio, by
multiple orders of magnitude.

(And then have the largest deployment opportunity, who cares if you
write a program for a platform that has 10 machines on the entire
planet? If you write a program that can run on a billion machines,
that's a lot more interesting.)

> So "binary" is false. And
> "corporation-influenced" is an ad hominem attack.

Hands up everybody who's using a process that _wasn't_ produced by a
corporation.

> You have the code.
> Call out what you don't like in the code and point to some thing better
> that's not "corporation-influenced."
> 
> I work for Google, but I have no particular love of Android; my only
> particular interest in it is that it runs on small devices. I run
> Cyanogenmod on my Nexus 7. I use F-Droid on my corporate phone to
> install open source alternatives whenever I can.
> 
> My biggest gripe is against the Play store. It has convinced potential
> and past open source developers that they can get rich on their, when
> most of the time they'll get a pittance while the community loses far
> more value through the closing of their code.

My current biggest gripe about Google is that Youtube is going full-bore
evil at the moment:

http://zoekeating.tumblr.com/post/108898194009/what-should-i-do-about-youtube
http://techcrunch.com/2015/10/21/an-offer-creators-cant-refuse/

Google is a large entity with lots of different groups working at cross
purposes. The chromeos and android guys weren't exactly cooperating when
I was paying attention to chromeos 5 years back. (The chromeos guys
eventually added an android emulation layer. Gee, I wonder why?)

> And more often, they keep
> it closed just to put ads in it, though many of the projects in F-Droid
> are open source apps that have ads in their Play Store version. That I
> can live with; if you are too dumb to install F-Droid you can live with
> the ads.

I've never bothered to cyanogenmod my phone because I use it as a phone.
I have like 5 old phones lying around I keep meaning to turn into
development thingies...

> The worst thing is that AdMob is one of the products I support. I am an
> SRE so I keep it running rather than adding features to it, but still. I
> do complain about it a lot though.

I applied to become an SRE at Google once, back in 2009. They strung me
along for 8 months, visited The Googleplex, my resume apparently made it
to the desk of whatever co-founder personally approved all hires... and
then they told me the position I was interviewing for was filled and I'd
have to start over re-applying form the beginning and I told them I'd
rather get on with my life.

Since then the annual google recruiter randomly contacting me (still!)
has always been CONFUSED about my status in their system...

(I should write up why this "alphabet" thing is not only healthy for
Google but at least five years overdue...)

> I wish the Play Store had a way to search for open source projects. Even
> searching for "open source" turns up mostly closed source apps. Which
> reminds me, I need to submit that as a feature request.

The play store's a walled garden, trying to "fix" it really doesn't
interest me, I'm waiting for it to collapse and be replaced.

Technologies tend to start out as a monopoly and then get commoditized
as they mature. Various bastards try to corner the market and
de-commoditize things, and sometimes they succeed for a while, but they
invitably re-commoditize again. (Although "inevitably" can be 30 years.)

I'm not trying to predict how this will commoditize, just that it
eventually will. The "install from curated repository" model does seem
to be what internet users want. Multiple competing Linux distros each
with their own repository was our "whitebox pc" model on the software
side. While we do have multiple android distros, they're tired hardware
provided by phone vendors bundled with service contracts. Google's "play
store" strategy for the past few years has been to shove a single
unified repo down all their throats (like it or not) and use bundling to
force this, with specific killer apps like youtube only available from
that repo and mandatory package installs as baseline prerequisites of
anything else in that repo.

Oddly Samsung is the big killer hardware vendor in the US. (Not so much
in India, which is itself a billion people, that's where Tizen's doing
well.) Having multiple players each refusing to be subordinated to each
other is usually an important part of commoditization...

*shrug* I'm not trying to steer at that level, at best I'm white-water
rafting.

Rob

 1445478814.0


More information about the qcc mailing list