Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Excellent. It's always been weird having a mobile device run YouTube more responsively than an GBP1500 laptop because the latter had to run the Flash plugin.


Now only if Amazon/Netflix would drop Silverlight all would be right with the world.


Netflix has HTML5 support for Safari and iirc IE.


The "HTML5" support requires proprietary binary blob that has no standard API (besides an API to launch it with opaque data from a proprietary server) and is licensed only to selected companies.

Basically it's as much HTML5 as <object data="swf"> was HTML4.

It's supported only by DRM vendors: Google (WideVine), Safari (FairPlay) and Microsoft (PlayReady) and doesn't work in open-source browsers (not even in Chromium or custom WebKit builds).


Mozilla is working hard on this, but I don't believe there is a published timeline for EME (using Adobe Access DRM) that supports Netflix yet. Getting closer, though: https://bugzilla.mozilla.org/showdependencytree.cgi?id=10158...


Netflix works fine in Chrome on Linux. What do you mean by "licensed only to selected companies?"


Works, because Google is a DRM vendor (WideVine).

You can't make a web browser that plays Netflix without signing contract with one of the DRM vendors that Netflix supports, or rolling your own DRM and convincing Netflix (and distributors that pull their strings) to adopt it.

For example if you compile Chromium from source it won't work (at best Netflix will send you encrypted blobs that you won't know how to decrypt and there is no public spec for it anywhere).

From video publisher's perspective it's locked-up as well. To have your own video DRM-protected with the same tech as Netflix you'd have to get a license from Google (for Chrome WideVine DRM support), Apple (for Safari FairPlay) and Microsoft (for IE PlayReady DRM).

Netflix's playback without "without plugins" is achieved not by removing plugins and having open standards, but merely by convincing browser vendors to bundle closed DRM code with their browsers. Technically it's more like Chrome shipping bundled Flash, but fortunately nobody calls that a "HTML5 native Flash without plugins!"


Where do you cut the line? It's a plugin because it uses the plugin api? And were it integrated in the main binary, would you call it a "plugin-less proprietary HTML extension"? I think the main differentiation point is that it's made by the browser authors, shipped with the browser, updated with the browser. That, to me, means that it's not a "plugin", even though it uses the plugin interface; it's an implementation detail.


"Plug-in" in the web browser context also had a wider meaning of "3rd party proprietary binary blob, that has limited portability, and adds functionality and APIs that are not part of Web standards."

You're right, technically EME's integration and distribution are different, so it's not a plug-in in that sense. However, the other downsides that were associated with plug-ins have remained the same in the built-in ex-plugins.

Browser plug-ins get so much hate not only because they're pain to install/upgrade, but also because they're usually not as open and portable as the rest of the platform.

(a bit meta, but term for this conflation of meanings is disguised query: http://lesswrong.com/lw/no/how_an_algorithm_feels_from_insid... )


What's the point of this DRM? It's not like you can't decode and record.


All DRM has [analog holes][1]. Its purpose is [to give content providers control over the features offered by playback software/devices][2].

[1]: http://en.wikipedia.org/wiki/Analog_hole

[2]: https://plus.google.com/+IanHickson/posts/iPmatxBYuj2


Chrome isn't an open source browser. Google is "one of the selected companies".


Correct. Netflix does NOT work in Chromium on linux (just tested).


The youtube blog post linked here says they're using EME just the same as netflix is...


Maybe for their paid videos. They can't be using it for regular videos because I have browser extensions that simply download the videos.


And Chrome.


Safari only works on certain 2011 Macs and all 2012 and later Macs.

There are a fair number of people who can't skip Flash quite yet for Netflix.

Good start though.


Also Chrome on Linux


And my brokerage account.. Their ActiveTrader platform on Silverlight is a piece of crap. It crashes exactly when you are trying to pull the trigger on a trade. I despise Silverlight with every ounce of my being.


Honestly, consider switching banks. Not sure the situation in France, but in the US there's loads of brokerages. It's a very competitive space, and there's no reason to put up with a horrid UI (one of the most important parts of a brokerage!) if you don't have to.


Microsoft just outright pays companies to use it.

"Here, if we give you 10 million USD for free, can you just get your devs to use this untested pile of DRM?"

Really.


This is where we get back to the Firefox DRM support, and kind of vindicates them. Sure, we'd love to have no DRM, but we can't get that so at least let's provide native support and axe Silverlight.


Amazon also supports Flash, but that isn't really a big improvement.


I've had a pretty poor experience with HTML 5 video. It stutters far more than flash and in particular it does very weird things when I try to skip around in videos. I'll miss flash.


The thing I find is that ever time the HTML5 window changes size is that it appears to drop all data and start loading again.

So my old tactic of open 5 tabs, maximise and watch one by one is completely nerfed


The Flash-based player would do exactly the same (with a slight delay) even if it wasn't as obvious from the UI.

To prevent YouTube from switching sources when resizing the content, select a specific quality first (it defaults to “Auto”). I wouldn't be surprised if there were browser extensions that automatically do this (e.g. always choose the highest quality).


I think the flash player doesn't change quality on resize and thus was able to keep using buffered quality?

I switched to forcing HTML a while ago so can't test. Either way, I'll try that - thank you :)


Same, Flash serves h264 to my old core2 laptop, html5 forces shitty VP8. Result is smooth 720p of mpeg4 versus playable 640x480 on googles codec.

Of course computer itself plays 1080p h264 perfectly fine in mplayer, so I might end up writing javascript that fires mplayer with mp4 link (to stream rather than downloading whole clip) instead of displaying video in the browser :(

There used to be mplayer plugin for firefox a loong time ago, nowadays all browsers use building unoptimized codecs.


Yes, one of my friend has an old computer and I had to install a "force flash on Youtube" extension for him to make the videos stutter less.


For me it's the other way around, in a 4 years old netbook I still use, the HTML5 video usually performs better than flash.


I don't think your experience has anything to do with my experience. That fact that on your hardware you felt html 5 video works better has no bearing on whether it work for me, on my hardware.

The fact that somebody down-voted me for adding a relevant anecdote, but happens to go against the tide of this threads opinion(ie: FU Flash), it's pretty indicative of the extremely poor voting etiquette around here.


I find it ironic that you are berating JustGotHere for sharing his experiences, but feel that it is inappropriate that you were downvoted for sharing yours.

Your anecdote isn't relevant, because I don't know what hardware you run, and it's likely that most other readers aren't using the same hardware. Your time would be better spent filing a bug report.


The thread is full of "flash sucks, good riddance, poor performance, etc", so he was not at all adding anything new to the equation and his response was in direct reply to mine, as if any repose other than one which was in lock-step with the thread's general opinion required instant rebuttal.


Youtube and Twitch are the only reasons I keep Chrome around - I much rather supporting the work that Mozilla are doing with Firefox (ie a proper open source browser) than Chrome.


I'm finding that Chrome hit sort of a decline from a personal peak in that current versions seem to be pigs on resource constrained systems, and has started to move backwards in terms of UI friendlyness (e.g. frequent requests for login to google accounts, rollout of more difficult user profile switching). IIRC the integrated flash plugin in Chrome is proprietary and at times it seems to be at the root of hanging processes or runaway resource consumption with Chrome. Ironically flash compatibility is one of the reasons I keep Chrome around - I've dropped flash plugins for all my other browsers...

I'm hoping it's a local max/temporary decline for Chrome, but Firefox now feels like the lighter of the two browsers in my use.


I've recently switched to Opera. It's ludicrous in my mind because it's Opera and I'm an older guy... history... but now it's webkit and it's essentially a polished version of webkit without the guest crapware of chrome. It has ublock too.


What is funny is there are users that prefer the old Opera (not based on Chromium) too.


To improve the new user profile switching, try enabling this flag: chrome://flags/#enable-fast-user-switching


Out of interest, why is Firefox a 'proper' open-source browser while Chrome is not? Isn't Chrome pretty much just Chromium with a little bit of extra Google integration on top?


Chrome is open source in the same way Android is open source - built on open source components, with some proprietary sprinkles on top, but good luck having any say in the development of Chrome or Android, it's pretty much Google's way or the highway.

With Mozilla, all discussions are open, choices are justified, and literally anyone (assuming your reasoning and technical abilities are sound) can contribute.

(I'll note that I don't actively contribute much to Mozilla, but if nobody uses Firefox then it's not going to keep getting better).


Open source and open development model are two different things (the latter almost always includes the former, but the former may or may not use the latter.)


Mozilla is still feeling the effects of poor choices made by Netscape. If you want a real open source browser, try Konqueror - with (IIRC) <1/3 as many developers as Mozilla they wrote a complete open source browser from the ground up, with a rendering engine so good that Apple and Google adopted it to build their browsers on top of.


Beyond open-sourceness it is also worth noting that while all browsers are behaving well on the internet right now (ever since Microsoft failed with the IE monopoly), Mozilla just does not have the permanent lingering need to make a profit, collect user information or to use their browser strategically in any way.

Occasionally there are already Chrome-only websites, there is a Chrome web store that could potentially become a gatekeeper for software similar to what Apple is doing in the iOS ecosystem.

I'm using Firefox even if that means that I don't get some cutting edge feature that I don't really need until a few weeks later. Hope that makes sense.


Last I checked, the bits of Chrome that aren't in Chromium are:

1) PDF viewer. 2) Flash. 3) Some audio/video codec support (e.g. Chromium has no support for MP3, AAC, H.264, or the MP4 container format; see <http://www.chromium.org/audio-video>). 4) Crash reporting, metrics, that sort of thing.

#1 and #2 are certainly closed-source. I can't speak to #3. [EDIT: as surrealize points out, #1 is no longer closed-source.]

From an end-user perspective, there's a significant difference between "browser that will play the music and the videos and show me the PDFs" and "browser that will not do those things". Chromium, while open source, is really not a viable end-user browser on today's internet.


> From an end-user perspective, there's a significant difference between "browser that will play the music and the videos and show me the PDFs" and "browser that will not do those things". Chromium, while open source, is really not a viable end-user browser on today's internet.

I don't think so, because really it's a difference between "will show the pdf..." and "will launch a program to show the pdf."

As a full time Chromium user, I've had frustrations twice--once trying to figure out why Netflix didn't work after I had heard it "now worked on linux," and the other with an app called mightytext that was doing a very stupid redirect loop forever in Chromium, but not Chrome.

That's over the course of a year or so.


I've done some user support, and browsers showing PDFs in an external app instead of inline is an extremely common user complaint.

Not only that, but PDFs embedded via <object> don't get shown in external apps in any browser. And this does in fact happen. For example, if you want to get your pay stubs or W2s via https://portal.adp.com you end up dealing with just such a setup.

I realize anecdotes aren't data, but this also applies to your anecdote; you're very much not a typical end user (starting with the fact that you're using Linux!).

I should note that "Netflix doesn't work" is probably a deal-breaker for the average user all on its own.


I have half a dozen devices that play netflix. Wii, chromecast, phone, tablet--and it's really not hard to switch to chrome to watch netflix.

I'll agree I'm not a typical end-user, but I don't remember having to "work hard" to get anything I wanted out of Chromium.

I troubleshoot every little problem for every member of my mostly tech-illiterate family, and I don't remember ever having this problem (like before PDFs would open in browsers, say).

Anyway, just more anecdotes.


It looks like they open-sourced the PDF viewer:

http://www.infoq.com/news/2014/06/google-chrome-pdf-engine-f...


Oh, I'd missed that happening. Thank you for pointing that out!


Also, Widevine.


On Android Chrome isn't open source at all. On Windows/Linux several key parts are closed source, that's quite a bit more "than a bit of Google integration".


What key part is closed-source on Windows/Linux?


Maybe "key part" is overstating it a bit, but the auto-update, crash reporting and usage tracking systems are all closed-source, as are the Flash plugin and some non-free audio/video codecs that are bundled with Chrome.

https://code.google.com/p/chromium/wiki/ChromiumBrowserVsGoo...


I think the auto-update, crash reporting and usage tracking code as well as the codecs are open source, they're just not compiled in by default in chromium builds.


Everything. It's only distributed as a binary.


What does this reply even mean? Of course it's distributed as a binary -- should regular users compile it themselves?

Advanced users can compile chromium if they want; from what I understand it's virtually identical to the chrome binary.


If the source code is not available and it's distributed as a binary, by definition it can't be open source. (Plus some might argue about OSS licensing too)

Chromium may be OSS, but Chrome is not.


Yes. Chrome is not Chromium, nor it Opera.


That's what I'm asking: For which part of Chrome is the source code not available?


Are you not answering your own question? The "extra" little bit of Google-specific code tries to tie-in the web to Google.


Well Chrome is not open source at all, all that Google integration is proprietary.

Chromium is fine, and I use it next to FF all day every day. I have never tried committing to it so I have no idea, but the general mindshare seems to be Chromium contributions are Google dominated.


Are you saying that the Flash plugin bundled with Chrome is better than the Flash plugin you can install for Firefox? If so, do you have any idea why?


On Linux, it is. The Flash plugin on Linux for Chrome is v16. The version for Linux Firefox is 11.2. Adobe stopped supporting Flash on Linux, so Google stepped up to the plate and started bundling Flash with Chrome and supporting it themselves. Flash does still work on Firefox, but at least on my install Click to Run was enabled by default for it (that might be a corporate setting though, I'm not sure".

Regardless, yeah Flash on Firefox on Linux is several years behind.


On linux it's because the NPAPI flash plugin hasn't been updated in a few years. It went Chrome/PPAPI a while back and firefox doesn't support it yet. I think it's supposed to get security fixes but performance issues are never going to get fixed in it.


Also, being on PPAPI means that it's fully sandboxed; add to that that google auto-updates it without nah screens and installer downloads, and you get a far higher level of security


If your browser supports VP9, that will be forced whenever possible.

I'm not sure if switching to CPU-only video decode will be a net gain for most.


Well, certainly in developing countries where bandwidth is limited, you're going to be watching a small resolution, and any amount of CPU is fine so long as you can cut the bitrate.

There are a lot of people in places like that.

Also, Allwinner, Intel, Mediatek, and Rockchip all have SoCs announced with hardware VP9 decode: http://wiki.webmproject.org/hardware/socs (not sure how many of those are already shipping).


There's a talk online by a Youtube engineer that covers them rolling out VP9.

Apparently the stats on user engagement show improvement. It seems that initially they showed a lot of improvement, but then they realised they were comparing the average video against only the most popular videos in VP9 (since those made the most sense to transcode first) and obviously people are more likely to keep watching something really popular.

Once they took that into account they did identify that some older machines would have a degraded performance and therefore lower engagement with VP9, but the heuristic for avoiding them is "Running XP" so it seems that most modern machines have a better experience with CPU decoding of VP9.

I believe they were up to 60% of all desktop views in VP9 by the end of last summer, so I'd guess if it was going to melt anyones's computer it would have happened by now.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: