Open Letter to Steve Jobs (and a reply) | Hacker News Reader

Open Letter to Steve Jobs (and a reply)

blogs.fsfe.org

169 pointsZeroGravitas115 comments

Open articleView on HN

"A patent pool is being assembled to go after Theora and other “open source” codecs now."

That's kind of scary.

Is it real? I am no fan of Apple or Jobs, but it almost sounds too evil.

Edit: in reply to a "tweet" of mine, he posted the headers:

http://hugoroy.eu/jobs-os.php

It's really hard to believe they'd be such arrogant jerks as to go after something open source like that.

And on top of that, I'm pretty sure webkit infringes at least some patents too, which is used as an example of Apple supporting and contributing to open source. So he's using double standards again.

Do you have any proof that the "arrogant jerks" at Apple go after Theora themselves ? Steve Jobs did not state it anywhere. Way to get worked up over the wrong guys.

This isn't "Apple" or "Jobs" going after anyone. Jobs is stating that members of MPEG-LA plan to do so. If Steve Jobs mentioning that other people plan to sue is an example of Jobs being "evil," then logically you must be evil too since you have also mentioned the potential lawsuit.

Presumably it would be the holders of the putatively infringed patents that would be assembling the "pool". I don't think Apple is in the Video Codec patent-holding business. As a major licensee of H.264, you'd have to presume Apple be in a position to know if someone else was rounding up the usual suspects.

Apple is not only licensee, but also licensor.

Every time you license H.264 (or something other from MPEG-4 pool), a cut goes to Apple (also to Microsoft and Nokia).

I stand corrected.

Apple has one patent in the H.264 pool, US 7292636, and one in the MPEG-4 Systems pool, US 6134243. I have no idea if either of these would be relevant to Theora.

For what it's worth, there are hundreds of patents in the total MPEG-LA pool.

MPEG-4 Systems is not offered any more (since it's effectively dead).

"Systems" is mp4 container.

Did you mean MPEG-4 Part 2 (Advanced Simple Profile)?

The things which are patented generally apply to parts of the spec that have never been deployed.

> For what it's worth, there are hundreds of patents in the total MPEG-LA pool.

Furthermore, it's not clear what Apple would stand to gain by enforcing its two patents that may or may not be relevant to Theora. Maybe I'm wrong, but media patent enforcement doesn't seem to be one of Apple's core businesses (HTC lawsuit notwithstanding). There are plenty of other companies, though, that likely have a vested interest in protecting their much-larger video patent portfolios.

I think it would be wrong to assume at this point that Apple is instigating patent litigation against Theora. Occam's razor would encourage the view that Apple knows the people who want to take down Theora and is rationally fearful of them.

Apple stands to gain more than just licensing fees by threatening any non-H.264 codec. As Steve Jobs points out in his "Thoughts on Flash", Apple has millions of devices in the wild whose hardware is optimized exclusively for H.264. If any other video format gains dominance, the devices will not work nearly as well (cutting the battery life by half just to play an older codec, in his example).

No, his example referenced a little known fact about H.264, comparing hardware decodable H.264 versus non-hardware-decodable H.264. The iPhone isn't playing some other codec.

H.264 can be encoded, using the same tools, in a way that can, or cannot, be hardware decoded.

I was pretty sure he was referring to VP6-encoded Flash video, not some version of H.264 that is harder to decode. Could you provide more information on non-hardware-decodable versions?

He probably means H.264 profiles.

iPhone (and Android phones) can decode only certain profiles (Baseline), and up to certain levels (depending on resolution). Even hardware assisted decoding with PC graphics cards can accelerate only up to certain level (usually high@4.1, but some nvidias can decode high@5).

The rest is more complex than hardware assisted decoders are able to decode and pure software codec has to be used.

iPhone supports High profile just fine.

IPhone does not support high profile by any stretch of imagination. Why don't you go and check Apple documentation? There's a technote for that.

Or just try playing high@ video on iphone.

http://120fps.org/l_high3.mp4

I copied this High@3.0 file out of my iTunes library and it successfully syncs to my 3GS.

I can't get it to play over Wifi, while I can get a Main@3.0 file to play, but I think this has something to do with not enforcing some bitrate limit rather than the profile.

iPhones don't even support Main profile (though some suggest that the 3GS actually has enough power to do so, it's only software limited by Apple to keep the line homogenous).

Search for "Baseline Profile" on this page for the full details:

http://www.apple.com/iphone/specs.html

VP6-encoded Flash video doesn't run on the iPhone, so that's not what he was referring to.

I believe Jobs is talking about comparing two videos on the iPhone, with hardware accelerated video lasting 10 hours and non-accelerated video lasting 5 hours. You may not notice if you're not paying attention to battery while you watch video.

To the sibling comment here, I'm not referring to profiles. I'm referring to the fact that various implementations of H.264 encoders produce content that, depending on the implementation, may or may not be accelerated by the iPhone's PowerVR SGX 535.

One can produce a pair of H.264 encoded in .mp4 container movies, encoded from the same mpeg2ts original, one of which can use hardware acceleration, and the other of which cannot.

(I haven't had the chance to check yet if Adobe's 'Gala' responds the same way to such files on the desktop Macs Gala supports.)

It's a good point that Apple's mobile devices are invested in H.264, but do you really think that -- were Theora successful and safe from litigation -- Apple wouldn't put a hypothetical Theora decoder in the iPhone? It doesn't seem rational to me to oppose Theora just because current hardware doesn't support it. Consider, for example, how Apple has largely left FireWire (its own technology) behind in favor of USB. Apple is a very "out with the old" company.

Google purchased the creators (On2 Technologies) of the basis (VP3) of Theora. http://googlesystem.blogspot.com/2010/04/google-will-open-so...

Now, just because you create something doesn't mean it doesn't infringe on someone's patent (read: it could infringe a patent). However, google presumably did some serious due-diligence before forking over their measly $133M, especially on the patent position. That doesn't make mean they're right, but it does mean that they think the patent position is pretty strong. Of course, they would have focused on VP8, the one they bought, not Theora's VP3 (and which might infringe due to subsequent open-source additions.)

Google might open source VP8 on May 19-20 http://www.reelseo.com/html5-vp8-google/

At which point Google gets to find out what patents VP8 infringes.

Ballsy move.

Actually, theora devs have been asking for the patent pool for a long time so they can work around patents once and for all http://news.ycombinator.com/item?id=1309412

So this is excellent news for Theora, as it could be an end to the "submarine patent" FUD.

http://www.osnews.com/story/23058/Theora_More_of_a_Patent_Th...

  For 15 years, Xiph.Org has carefully "played by the
  rules," fully within the bounds, intent, and letter of
  intellectual property and patent law. For the past ten
  years we've informed the entire world, including MPEG LA,
  of our specifications and algorithms in detail. We've
  requested in open letters that any group believing we
  are infringing to inform us so that we make take
  immediate corrective action.

  I predict that MPEG LA may counter that they know
  groups have been pressured into licensing patents in order
  to use Theora. This has been a recent back-room assertion.
  You might want to ask point blank if MPEG LA itself or any
  of its constituent members has engaged in this practice,
  thus manufacturing the evidence that "vindicates" their
  patent allegations. I beg you - tell me immediately if you
  get a straight answer (or good video of any squirming)!

  I'm sure you can tell I'm a bit peeved; this has been going
  on for over a decade. It's amazing they've never been
  called out on it.

While I agree with the spirit of the message to Steve, the attack on Apple not sharing back open source software modifications is out of place. Darwin is open source. WebKit is open source and actively developed by Apple, giving competitors the ability to quickly catch up with one of the iPhone's biggest advantages. Apple is also the chief contributor to LLVM.

Webkit is open-source because the people who wrote it (KDE devs) put it under the LGPL. Apple was playing catch-up, and used the best code base they could to get started. They aren't philanthropists, they are just (grudgingly) abiding by the license of the software they are building on.

It certainly used to be grudging -- originally Apple just threw a huge clump of code back to the KHTML guys, technically satisfying the LGPL, but not doing anyone much good. That's a far cry from the community they've built around the project now. I wouldn't call Apple philanthropists, but they're not simply following the license either.

They also could have rewritten that. The original KHTML code is a pretty small part (about 140k loc initially) compared to the WebKit project (about 3M loc).

Also, they only really needed to have WebCore and JavaScriptCore open but they decided to make all of WebKit open later on (all the rest under a BSD licence).

That's true of all companies that contribute to open source.

How do you come to that conclusion?

I thought it was obvious. Perhaps I should rephrase it: All for-profit companies contribute to open-source because it benefits them.

Yes, at least the companies should aim for that.

I asked because some companies see the embrace the benefit and do not grudge.

Yes, at least the companies should aim for that.

I asked because some companies embrace the benefit and do not grudge.

What grudge? How are you measuring grudge-ness? How do I know with confidence whether a company is grudging or not? And furthermore, why would I care?

What do I know. One of the ancestor comments introduced the concept: "They aren't philanthropists, they are just (grudgingly) abiding by the license of the software they are building on."

As an attempt to answer: They can just satisfy the letter of the GPL by a code dump--which I would count as grudgingly, or they can embrace open development.

Everyone (businesses and individuals alike) who contributes to open-source does so because it benefits them.

Or because they think it benefits them. Might be a tautology, though.

Macruby is open source. Apple was under no obligation to open source it, yet they have...

Grand Central Dispatch also comes to mind.

All video codecs are covered by patents.

All the popular ones, maybe. But MPEG-1 is over twenty years old, so unless there is a submarine patent out there, all the patents on it have already expired. Likewise, most aspects of MPEG-2 should be freed up within a decade. Even H.264 will eventually be unencumbered, given enough time.

On2 had a pile of active patents covering VP3, which they granted irrevocable royalty-free licenses to when they abandoned it to become Theora.

Nobody's ever done codec development without either filing a ton of patents or at least publishing rigorous documentation to establish prior art (Dirac). Some of those patents may no longer be in force, but they still exist.

The On2 patents on VP3 only covered some optimization techniques and are not required to implement VP3 or Theora.

Steve has been quite busy with his email lately. I don't remember him ever doing this much correspondence via email.

Very interesting.

I had the same thought. I think its because he wants to get "Sent from my iPad" on the internet on as many web pages and into as many minds as possible. Its a marketing ploy.

Ah, that's why they're so short!

I figured it was the soft keyboard on the iPad, forcing the author to be as concise as possible since it is such a pain to type on.

He's been doing this for years. I know several people that have gotten terse, 3 or 4 word responses to him. It's only recently that the press and the bloggers have noticed this.

He's been doing short emails for years, and the occasional open letter, but it seems that he's doing it a lot more as of late.

Presumably in response to all the recent Apple controversy.

Either that or more people are posting/linking them.

Yeah, 5-10 years ago no one except the niche community of Mac users and developers was following Apple this closely. It's almost hard to remember how small Apple's influence had gotten by the time of Steve's return - before the iMac, iPod, iPhone, or iPad. How many people even remember products like "Performa" or "eMate"?

I really wish people would stop redefining the word "open" to mean whatever makes them feel good. (Vain hope.) Patents are totally orthogonal to copyrights and standards. You can put the source under whatever terms you want, and get whatever international standards body you want to bless the protocol, and that has nothing whatever to do with the patent troll who will show up the next day and start charging license fees.

you like Openness so much that when you use “Open Source” Software to build your Mac operating system

Only some parts of it. The bulk of OSX is closed source.

May I remind you that H.264 is not an open standard?

Jobs said H.264 is an industry standard not an open standard

But the whole point in this debate is that Flash is also, for better or worse, an industry standard.

Is there a specification?

I think "de facto standard" is a more appropriate phrase for Flash.

Yes, there is (http://www.adobe.com/devnet/swf/).

It is even available gratis (H.264 specification is not).

H.264 specification is available gratis: http://www.itu.int/rec/T-REC-H.264-200903-S/en

I stand corrected, thanks.

Interesting, but I don't see how one could create a complete implementation without a spec for the Sorensen Spark codec.

FFmpeg has reverse engineered it and there is an open source decoder and some documentation.

Sorenson Spark == H.263 (not a typo), which is documented. The problem with Flash is VP6, although FFMPEG reverse engineered it.

Actually, it is based on a draft h.263, not the final h.263.

One of the older flash specification did document it. The svq3 specification was removed from swf specs around time it was made available under more liberal license (Sorenson probably didn't agree with wider specs availability).

Flash isn't an open standard by any useful definition.

Sure they specified the container format and data structures (just like with the MS Office formats), but everybody had figured that part out a long time ago because it's obvious and easy to capitalize on by developing an extractor. The runtime APIs are incredibly hard to reverse-engineer, comprise the vast bulk of the Flash implementation, and are the entire reason for the plugin's instability.

That's a fair point. I think that's more or less how Google approaches it even though they are not fans of closed web formats.

I do think there is a world of difference between a big complex closed platform exclusively available from Adobe and H.264 which has a variety of encoders/decoders available from multiple license holders. I seem to recall in the past Adobe has licensed in Flash to Nokia in almost the same way. Nokia ported & optimized the code for some of their handsets (N900 and others) It would be interesting to know if Adobe offered a similar arrangement to Apple. That would completely debunk everything Apple has been saying if they had the flexibility to control the Flash implementation on the iPhone themselves.

Was an industry standard.

And that's the point, flash has some real drawbacks and since it's not an open standard with multiple implementations, it stagnated while Adobe thought it had no competition.

To add to jsz0's correct observation, H.264 does not show up at all under the first section discussing proprietary vs. open standards.

H.264 shows up once under the heading Second, there’s the "full web" but not really a primary point.

The main H.264 is in the section Fourth, there’s battery life. Jobs wrote that the benefit of H.264 vs. flash was that there was hardware decryption support for it. The result of hardware decryption support was a doubling of battery life vs. software decryption (e.g. flash).

I see a lot of jumping from Job's discussion of open vs. proprietary to his advocacy of H.264. While there are serious problems with the openness of H.264 (from an OSS POV), Jobs tied battery life to H.264 in his objection of flash codecs, not openness.

People are making this jump because Job's is committing the fallacy of special pleading when he applies his rule of openness to Flash but not to H.264.

Hopefully Bilski renders these patent issues moot. I really, really hope that.

And if it does, I expect multi-media to improve quite a bit in the next 2 years.

If you don't know what "Bilski" is (like me). I founds us a link:

http://en.swpat.org/wiki/Bilski_overview

I doubt it. Bilski will be decided by SCOTUS in exactly the same way they "decided" Eldred... by punting it back to a very well-bribed Congress.

And that won't even be the wrong decision, really. There is no constitutional basis for a ruling on software and/or business-method patentability, one way or the other. It's clear that Congress is supposed to make those calls.

"A lot of commercial software comes with H.264 encoders and decoders, and some computers arrive with this software preinstalled. This leads a lot of people to believe that they can legally view and create H.264 videos for whatever purpose they like. Unfortunately for them, it ain’t so. Maybe the best example comes from the Final Cut Pro license:"

http://bemasc.net/wordpress/2010/02/02/no-you-cant-do-that-w...

I think that, ultimately, Apple have as much ability to define what an Open Standard means as the FSF within the context of what they are saying.

As it happens I think I agree with where Jobs is coming from... or rather I think this letter nit picks at a specific phrase (Open Standard) that SJ's was just using in a different way to how the FSF would.

Ultimately what Jobs meant by Open Standard was reasonably apparent (for the reasons laid out in this letter!) so nit picking over it seems a meta point.

h.264 is far, far more open than Flash. What matters is that the spec is public and that you can choose from many implementations. There are no usable competing implementations of Flash, but dozens for h.264 including excellent open-source ones.

The h.264 terms aren't onerous. If you ship less than 100,000 devices you pay nothing. At large scale, it's $0.10 / unit. When selling hardware, that's a pretty light tax.

As Jobs points out, Theora is not immune to patents. On2's many patents are irrevocably licensed for free, but there are thousands more patents that may apply held by trolls. It sucks to ship devices with unknown patent liabilities. It almost destroyed RIM, for example.

Here's a summary of the license terms:

http://www.mpegla.com/main/programs/AVC/Documents/AVC_TermsS...

Let's say you make videos and sell access to them on your web site on a title-by-title basis :) Videos <= 12 min, no charge. Videos > 12 min, USD $0.02 per title or 2% of access price, whichever is less.

Oh, maybe that's the basis of the general YouTube 10-minute limit...

Perhaps, but that policy would be forward-looking. Current there's no charge for web delivered content if you don't charge for access.

It pays to be forward looking when your supplier reserves the rights to re-negotiate your terms.

Right, 2% is unlikely to change anyone's business model.

http://lists.xiph.org/pipermail/theora/2010-April/003766.htm...

also

http://lists.xiph.org/pipermail/theora/2010-April/003769.htm...

I don't get it.

In his open letter, Jobs explains why he chose HTML5 and H.264 instead of Flash, and the author here doesn't not answer about Flash but H.264.

Even though Theora looks more open than H.264 for video, I think HTML5/H.264 still stands more open than Flash for the entire web.

So, yes, Theora sounds cool, but that's not Jobs point. The point is, get rid of Flash.

The writer seems to think that an open standard must be free of cost but is that actually true? Or is it just what many would like to be true?

Most standards, up until the turn of the century, were required by governments to have RAND licensing terms in order to prevent abuse of market power. That's Reasonable And Non-Discriminatory, so you could charge fees but couldn't mess with competitors.

Governments are now pushing for Open Standards which must be RAND-Z or RAND-RF. Z stands for Zero cost, RF for Royalty Free. These Open Standards make much more sense for software or services delivered at low or no cost via the web, or as Free or Open Source software.

Could someone please explain to me how Flash isn't open? To my knowledge, everything about flash is open other than Adobe's IDE. The spec has been released, the sdk is OSS and anyone is free to compile to SWF... Am I mistaken here?

The player certainly very much isn't open, aside from the Tamarin runtime.

Can you qualify that statement? edit: You are correct, the player provided by Adobe is not open.

However, anyone is free to make their own player, one alternative player is Gnash: http://en.wikipedia.org/wiki/Gnash

I can. "License: Proprietary freeware EULA" on http://en.wikipedia.org/wiki/Flash_Player

And the argument that anyone is free to re-implement is just as valid for anything. The fact of the matter is that thousands of man-years have gone into building Flash Player, and replicating them is far from a straightforward task. The spec is huge, and there are millions of .swf files to which an alternative implementation needs to be compatible. It's a bit like suggesting someone re-implement Windows if they don't like Microsoft's version of it. It's open, right? Anyone can look up the ABI for user space executables, kernel drivers, etc., so how hard could it be? ReactOS (or Wine) hard, as it turns out.

Note also that for the supposedly open components, Adobe controls them so tightly, they are only "look but don't touch" open. Openness is only useful insofar as you can actually do something with the information or code. They maintain this control precisely because they have an iron grip on the player. There is no point for a third party to extend the Flash spec or compiler, because they can't realistically build an implementation that actually runs code obeying the forked spec. It's as ridiculous as creating a driver to run on a hypothetical version of Windows. Yes, you can do it, but it's also completely pointless.

Thanks for the link, you are certainly correct that adobes player is proprietary, but I provided you with a link to an alternate player... Just because something is hard to reproduce doesn't make it closed. Do you consider java closed because Sun/Oracle makes the best applet viewer?

I am not trying to say Adobe is guilt free or anything, but it looks to me like they are making it as open as they can, unless I am missing something.

Java is licensed under the GPL.

No, I consider Java to be open because it is licensed under the GPL, an Open Source/Free Software license.

but it looks to me like they are making it as open as they can

With all due respect, you're either trolling or extremely ignorant for someone reading HN. In case of the latter, please educate yourself and read up on the difference between proprietary and open source software. Also take note that "freeware" does not fall under the definition of "open" software.

FWIW I will flag any further trollish comments.

I am sorry if you are offended by my ignorance. I am not currently arguing that the flash player provided by adobe is open. I conceded a while ago that the flash player provided by adobe is proprietary. However, an alternate flash player exists and you can program SWF without ever touching an Adobe product. There is a spec and open tools for doing so. I was simply under the impression that Flash was open minus Adobes implementation and am trying to see why everyone is saying it is not.

edit: I don't understand why me disagreeing with you automatically makes me a troll. People are quick to jump on people who don't agree with their viewpoints.

Adobe Uses DMCA On Protocol It Promised To Open http://yro.slashdot.org/article.pl?sid=09/05/22/1254246

How do Gnash developers work with the Adobe/Macromedia EULA? http://www.gnashdev.org/?q=node/25#eula

Read this two links and come back to discuss about flash openness.

edit: apparently adobe did open the RTMP spec? http://www.adobe.com/devnet/rtmp/

How does Gnash being built without Adobe tools do anything but further proving that you don't have to use adobe products to use/work with flash?

I am not claiming with authority that Flash is open, I am just trying to see how it is closed. So far it looks open to me.

I'm not even a Flash developer, so please forgive me if I am not understanding properly.

Did you really read the links? I can't do your homework for you.

"I'm not even a Flash developer, so please forgive me if I am not understanding properly."

Me neither I am just an average guy that likes computers. I am not even a developer of anything. But it is very clear to me that flash is not open and Adobe does not intent for it to be other than for PR. So far I haven't seen evidence to the contrary.

If Adobe would open source (GPLv2/BSD/Apache, etc) flash and would give the future development and implementation to an independent party (W3C?) made of various corporations, organizations and individuals that have stakes at flash I'd be all for flash.

P.S. I think a BSD/MIT/Apache license would be better.

Consider the "definition" shared in the letter.

Compare that definition to WebKit, then to Flash.

Which part of the definition does flash violate? Also, Apple doesn't give much credit to KDE with WebKit, they certainly like to imply that they are very open with it without mentioning its origins...

Do you work for Adobe? This is your first day commenting and you seem to defend flash as if your life depended on it.

I don't work for adobe, nor do I even code in flash. I am just ignorant on the subject and wish to be enlightened.

edit: it's also not my first day commenting, I just forgot my old account and had to make a new one a little while ago.

Props to Steve to actually answer to this open letter, but i think he could have put a little more effort into it, sounds like he has no confidence in open source. I wish apple would be a little more involved in open source.

He did put effort into it - there's a full paragraph there, Steve's emails rarely go beyond a sentence fragment; rightly so, as I doubt he has the time to research and write out a fully-formed response.

I think Steve has plenty of confidence in open source. He's just a realist and understands that current software patent law doesn't really allow for completely open video codecs. Even if you create an open implementation of them, you're going to be subject to patent lawsuits, and as we know, most open source projects don't have the funding to withstand the legal battles.

I wish apple would be a little more involved in open source.

hopefully they keep the LLVM team well-funded and give them good freedoms.

Why doesn't he at least say something about Apple's closed ecosystem?

may I add that software patents are non-existent here in Europe. So a codec made here in Europe should not have to worry about infringing anything..

I asked on the Theora mailing list their thoughts on SJ's answer and this is what I got from one of the dev:

> (...) A patent pool is being assembled to go after Theora and other “open source” codecs now. >

I think that's great news... If the "industry" finally compiles a list of patents that they think theora voilates we can:

1. find prior art. 2. work around the possible valid patents.

Afterwards all the submarine patents are gone, and we can prove it...

Afterwards all the submarine patents are gone, and we can prove it...

Well, no, only the submarine patents from MPEG-LA signatories are gone.

Oh, who cares about H264 compared to 3.3.1..

Exactly. This whole video-codec business is a nice bit of deflection on Jobs's part.

Reply on news.ycombinator.com