RE: Netflix HTML5 player in IE 11 on Windows 8.1

+1.  Mr. Cornelis, a certain amount of decorum is called for please. I have
thick skin, but I don't think we need to be using even mild cussing
(bullshit), we are supposed to be gentlemen (and gentlewomen) here.

 

JF

 

From: Mark Watson [mailto:watsonm@netflix.com] 
Sent: Wednesday, July 03, 2013 4:45 PM
To: cobaco
Cc: public-restrictedmedia@w3.org
Subject: Re: Netflix HTML5 player in IE 11 on Windows 8.1

 

I wonder if you can tone down the invective. It's not helpful or
appropriate.

On Wed, Jul 3, 2013 at 3:07 PM, cobaco <cobaco@freemen.be> wrote:

On Wednesday, Wed, 2013/07/03, Mark Watson wrote:

 

> > As a consumer all I see is Big Content making things more annoying and

> > killing good services through litigation or absurdly high licensing

> > fees (yes digital media is a good with a marginal cost of near-zero,

> > that means prices will and should be low, did you guys fail economy

> > 101?)

> 

> The marginal cost of content distribution before digital media was not

> that big a fraction of the overall costs. Production and marketing costs

> dominate, so the transition to digital media, whilst certainly offering

> the possibility of a better user experience, doesn't in itself imply

> dramatically lower prices. 

 

tangentially off topic again, but lets compare...

 

marginal cost of distribution of physical media is made up by:

- cost of sourcing a physical box

- cost of printing the artwork for the content

- cost of pressing the disc

- cost of combining the media, the box and the artwork into 1 package

- cost of a marketing push to convince stores that they should allocate
their limited shelfspace on your product instead of the competition

- cost of getting the box physically to the store

- cost of the logistical layer that ensures enough, but not to many,
physical media are made and distributed to every store and all the tracking
that requires

- cost of employees to man the stores and process the sales

- cost of transaction fee for payment

- cost of getting rid of excess physical copies that haven't sold

 

marginal cost of distribution over the web is made up by:

- the bandwith and cpu usage needed to server the media file and process the
sale, both of which are negligable already and getting lower every year

- cost of transaction fee for payment

 

You're cutting out the following middlemen out of the picture with fully
webbased distribution:

1) the box manufacturer 

2) the artwork printer 

3) the dvd presser 

4) the packager 

5) the sales agent convincing the stores to carry specific content

6) the logistical expert keeping the distribution and manufacturing going as
needed

7) the trucker 

8) the store sales clerk

 

In addition web distribution avoids completely:

- the risk of over or underproduction

- the thorny logistical problem of assuring timely manufacturing and
delivery, and of (trying to) make sure no store has to few or to many copies

 

Next the capital costs of 

- a server or even a complete datacentrum, uplink and website versus

- a physical distribution network with tens of thousands of stores (to reach
the same global audience) and all associated costs (manufacturing, trucks,
stores, physical inventory) 

... are again easily an order of magnitude different.

 

Quite obviously that easily allows a serious price differential

 

No, this is not obvious at all. You're ignoring the costs of producing the
content in the first place, which is the dominating factor. 

 

 

But lets look at a practical experiment with video:

- Louis CK sold his show "Live at the Beacon Theater" on his site for 5$ and
DRM free

- the site was new no so he had all the startup overhead included in the
production costs (and he didn't skimp as the site survived the unexpectedly
rapid and massive success)

- after just 12 days he stated publicly that he was completely amazed to
have made a million dollars (that's 200k sales) [1]

- cost breakdown as stated on the above link:

* 250k production costs for the show and website combined

* given how much he was (unexpectedly) making he voluntarily gave another

250k in bonuses to those involved with the production

* he gave 280k to various charities

* kept 220k for himself

- To compare: amazon lists a DVD of his previous shows as 10-15$ a piece
[2]. I have no numbers of how much of that DVD price ends in royalties for
Louis, but given how amazed he was at how much income the digital
distribution was bringing in, it's obviously a lot less then 5$

 

=> In other words cutting the price to a 3th or more is easily doable while
increasing profits (for the content producer) massively

 

For a piece of content with a relatively low cost of production.

 

> > Worse not only is Big Content killing good services through litigation,

> > they're refusing to then step up and offer a similar (level of)

> > service.

> 

> I'm not sure how this claim is relevant here.

 

Ok, answer me this:

Why is Netflix's streaming service not globally available right now?

 

I answered this in another mail.

 

> That's what we'd like to do, rather than 'click play, read annoying

> install plugin dialog, wonder if installing the plugin is safe, check if

> your system meets plugin system requirements, wait for plugin to download

> etc. etc.

 

to meet the new boss (same as the old boss) do 

s/plugin/CDM/g

 

the CDM is still a black box of which I need to wonder:

- what it does and consequently if it's save to run

- if my system meets its requirements (probably not as I run Debian with all
free software and no hardware protected media pipeline)

- still needs to get to and be approved on my system (you approve and
download only once, same as flash/silverlight/java. Or possibly it might be
included in your browser already, as chrome does with flash)

 

The intention is that browsers explicitly choose which CDMs they will
integrate with and ideally either ship the CDM with the browser or find it
already on the system. So, yes, you need to make the above decisions, but
you make them when you choose and install your browser or operating system
and with the knowledge that the CDM has been explicitly approved by the
supplier of that software. And browsers could choose to provide you with the
ability to switch it off and, I would hope, the ability to clear stored data
etc.

 

It differs from the plugin situation because you have a choice (content
providers will need to support multiple CDMs and the design supports that),
you make that choice once, in advance, not every time you visit a content
site and you have some assurances from a supplier that you hopefully trust
to have your interests (in terms of privacy, security and accessibility)
more at heart than the content provider does.

 

I do think these are real benefits. Of course, they're by no means a
panacea.

 

 

> > Hollywood has a documented history of abusing drm features to force

> > commercials down our troats, and keep us from (legally) playing a DVD

> > on our linux/BSD/... systems, keep us from watching a DVD bought on

> > vacation, and, and, and ...

> 

> The proposed Encrypted Media Extensions don't support forced commercials

> or region restrictions etc. (in respect of streamed/downloaded content,

> or DVDs). If you think the spec could be used that way, I'd welcome

> proposals to change it to stop that.

 

EME itself doesn't, true.

 

But EME is only the API for the browser to talk to the CDM. Given that a CDM
is a black box there's no telling what requirements/conditions the CDM's
will impose. Regarding the specific abuses of region locks and forced
commercials, see below...

 

First regarding forced commercials:

 

If Big Content wants to provides DRM'd files that include a trailer or
commercial before the movie. 

If the CDM is one of the kind that doesn't return decrypted frames but plays
renders them directly (see the graphic in section 1 of the spec [4])

Then how exactly is EME going to stop the CDM from preventing the skipping
of the trailer/commercial? I don't see any such guarantees anywhere in the
spec.

 

IIUC, if you set the playback position on the HTML media element, then
playback will either move to that position or you will get an error and
playback will stop. To support forced commercials, there would need to be a
behavior specified *in HTML itself* where playback continues from the
existing position. Perhaps I'm wrong and that behavior is specified, but if
so it could be removed.

 

Anyway, my point is that by working on this in the W3C we can decide what
accommodations in the workings of the media element we do or do not want to
make for CDMs. Also, the whole model is that it's the existing media element
that performs playback, so its existing behaviors - specified in HTML - are
expected to be maintained with minimal exceptions (for example the ability
to pull pixel data via a Canvas).

 

 

This is not a hypothetical thread:

The draft [4] currently contains the following in section 7.1.1: 

"WebM streams may be partially encrypted, both at the Track level and the
block level." 

So obviously the necessary groundwork for different restrictions on
different parts of the stream are already explicitly present.


I don't see the connection between this and forced commercials.

 

 

WebM is a Google thing, Google is waging the battle of forced adds right now
on youtube and most recently against Microsoft [5]

What did you think Google was going to use WebM+CDM for?

 

Second, this is exactly the pattern that hollywood is using with DVDs and
their DRM, there's no reason to expect them to suddenly change their tune.

 

Since the CDM's are explicitly allowed to depend on specific hardware
access/conditions, what's to stop a CDM from requiring a GPS chip and
checking the location?

 

The specification doesn't define CDMs, but perhaps it can define functional
constraints on CDMs, as can the browsers who integrate with them. This is
much harder to do with plugins and much more likely to happen with W3C
involvement than not.

 

 

> The W3C could also have some influence on the supported use-cases, as

> noted above.

 

indeed, by stamping it 'standards approved' it's gonna encourage DRM
slightly, that's a Bad Thing

> Personally, I agree that a "license" to a piece of content that is

> conditioned on various external factors out of my control is not a product

> I would choose to buy

 

Do you realise you've essentally just admitted that DRM is a ripoff deal. 

How do you reconcile that with activily arguing _for_ support of DRM?

 

(1) Because I don't choose to buy a product does not make it a "rip off".

(2) The "license" model for downloaded content, tied to continued existence
of some service, is just one model which DRM can be used for. There are also
models which don't require server interaction on every playback and there
are models, such as subscription services, where it is rather obvious and
reasonable that the content is only available whilst the service still
exists (and you are still a subscriber)

(3) My argument has consistently been that the situation will be better for
users if this work is done in W3C. I've not taken a position on the more
philosophical aspects, for example whether DRM is fundamentally "good" or
"evil" (or neither) and I don't plan to.

 

 

> (though I would like to have the choice, or rather, I don't want any 3rd

> party telling me what choices I can or cannot have). I would prefer to

> "own" a copy or to rent or access it as part of a subscription service.

> But that is my personal choice.

 

that's a copout:

 

DRM is an obviously abusive technology (you just admitted that it's a bad
deal above)

 

I said I personally don't choose one particular kind of service that uses
DRM. This is a long long way from agreeing that DRM is "abusive". We are in
no position here to legislate as to what products are good and what are bad.
Which will sell and which will not. Consumers should make those decisions.

 

> What hasn't been demonstrated in this discussion is how the W3C not

> working on this, as far as the technology is concerned, represents a win

> for users. 

 

The problem with DRM is the exact same problem as with patents.

The powerbase used is just technological instead of a legal.

But in both cases you're introducing a gatekeeper that needs to actively
approve/support client implementations (of/with CDM's)

 

You're introducing a dependency from the client on the longterm goodwill and
competency of the CDM provider.

 

Again, we're not introducing anything. We're proposing some technology
refactoring to bring it out of plugins and under the control of browsers. 

 

Preventing that is an obvious win (for all the same reasons why W3C demands
a royaltee-free patent grant from standards contributors)

 

W3C not working on this won't prevent anything. W3C working on this is what
could prevent some of the security/privacy/accessibility issues as I
explained above.

 

...Mark

 

 

> W3C's absence will surely make all of the user experience,

> security, privacy and accessibility problems that are raised in this

> context worse.

 

How? Whether you name the black box a plugin or a CDM makes no difference to
security, rivacy or accesibility concerns.

 

And W3C has explicitly told everyone that they won't do a thing to
standarize those CDM's

--

Cheers, Cobaco (aka Bart Cornelis)