Browser compatibility: will Polymer be usable in the next 3 years?

[fab...@youfoot.com]

Polymer is amazing! (but you probably know this)
Polymer and Material Design was one of the most exciting things I have since in a while!

I really liked the idea you could build apps with UI standards for Native device also across the web -- these days so many people think that the web is dead: because Apps are dominating in Mobile and Mobile is quickly dominating the web.
I also especially liked transitions/animations which I think is the next paradigm and really can improve UX. The whole Material Design is amazing.

Really really really amazing! I found myself watching as many videos as possible on Polymer and Material Design and there are now quite a few!

Browser compatibility: will Polymer be usable in the next 3 years?
But the one big BIG disappointment is browser compatibility. I was disappointed when I say the compatibility guidelines: http://www.polymer-project.org/resources/compatibility.html
But somehow I couldn't beleive it and I just hoped it somehow degraded nicely on older browser. I was very disappointed when I found it really doesn't degrade beautifully at all on things like Safari 5 or IE9. It just completely falls appart. Apparently even Android web-view in some cases Polymer will completely break.

So everything that was exciting about it especially being cross-device suddenly looked very over-stated at best. I mean we are all still hoping we can finally put IE6 behind us... So Safari 5, IE10, or older versions of Chrome & Firefox: that's at least 3 to 5 years at best!

I am surprised more efforts were not put to help adopting Polymer by creating a smoother transition by having ways to degrade Polymer beautifully on older browsers. Even if that meant loosing the benefits of Polymer for any of these Browsers.
Have I missed something? Is there a way to degrade beautiful Polymer to support the majority of browsers? Will browser compatibility improve: is that somewhere on the near future of the product roadmap?

Note: Attaching screenshot of Polymer demo running on Safari

[Rob Dodson]

Hey fabrice, as I mentioned over on Youtube, we only support the last two version of each browser. Polymer is a future facing library and we would need to significantly increase the size and complexity of our polyfill layer if we started reaching back to support extremely old legacy browsers.

Btw, you mentioned IE10, we actually do support that one. And there aren't many users on older versions of Chrome and Firefox as those browsers auto-update.

Follow Polymer on Google+: plus.google.com/107187849809354688692
---
You received this message because you are subscribed to the Google Groups "Polymer" group.
To unsubscribe from this group and stop receiving emails from it, send an email to polymer-dev...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/polymer-dev/b7192201-3fd6-48fe-ac20-a042e6f6ee41%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[fab...@youfoot.com]

Thanks Rob and I understand that the Polymer is advancing the future of the web. I am just wondering if the best way to reach the future faster is not to find a somewhat elegant way to deal with the past (and that's what I understand has worked best in the past to advance the web...)

When HTML5 new input tags was introduced I thought it was great how older browser would just treat these fields as text input. It really helped adopt HTML5's new inputs. Couldn't there be a similar way to achieve this with Polymer?

We still get around over 15% of our traffic coming from older versions of major browsers, 2,5% from Opera and Opera Mini, and a little under 1% from browsers that I don't know how they would behave (Opera Mini, Opera, Ovi Browser, Blackberry, Maxthon, Amazon Silk, Dolfin, PS3, IE with Google Frame,...).

So despite my huge enthusiasm to discuss with my team adopting Polymer for YouFoot, we can't just cross out 15% of our users and it will probably take 3 years for this 15% to become less than 5%, and 5 years to become less than 1%.

It would have been particularly great to adopt it to create a more consistant look across web and mobile devices, make interfaces more beautiful, code leaner...

---------------------------

Internet Explorer:
Speaking of IE which amount for 15% of our sessions: 33% of these are from IE 9 or previous (almost half is from IE9 and and half from IE8, with a small amount from IE7). Fun fact: we even have a few users connecting from IE5 and IE4 apparently!

Firefox:
For us Firefox represent 22% of browser usage. While you are right that it's generally better, you'd be surprised to see we get 205 different versions of Firefox. While 80% of it are from Firefox 30 and 29 we also have almost 1% on version 12, 0,5% on version 11, etc... There might be 5% of Firefox users on versions below version 10 (with version 3 and 4 being pretty popular still).

Chrome

Chrome sees the most usage with 48,8% but while 79% are using version 35, the next version that is most used is version 1,5 with 2,4% of Chrome users. There are more than 1000 different versions being used although 990 versions probably amount to less than 2%

Safari:
Safari is only 2,9% of our usage. Among that 7% are using Safari 5 or older.

[Joern Turner]

That are astonishing and interesting statistics and i very well understand that you must be cautious which technologies you adapt. On the other hand - being an open source developer for more than 10 years - i very well understand Robs' argument. Especially if you try to move things forward it's hard to get everybody on the train without loosing a lot of speed and momentum.

However i support your statement that we should look for 'cheap' fallbacks so old browsers degrade to something meaningful. E.g. take the simple 'no script' of HTML - if there are users no having javascript active for some reason i at least can tell them that the site won't offer its whole functionality (if any at all). This already makes a big difference as most users today have an alternative - they don't need necessarily connect with their aged smartphone browser but can go to their desktop.

On the other hand i think developers today need to put some pressure on users to upgrade if they want to offer the latest and greatest. Backward compat can become a too big burden sometimes and hinder innovation. Maybe you should probably have a look into how much profit these 'minorities' actually bring. Wouldn't it be possible that you can raise your profits by offering a much better (and more fun) interface for all the others?

Just my 2 cents,

Joern

[Rob Dodson]

However i support your statement that we should look for 'cheap' fallbacks so old browsers degrade to something meaningful. E.g. take the simple 'no script' of HTML - if there are users no having javascript active for some reason i at least can tell them that the site won't offer its whole functionality (if any at all).

We do have an `unresolved` attribute which can be placed on any custom element you use. You can style against this attribute (maybe displaying a background image that says 'loading' or 'web component support required'). When Polymer upgrades your element, it will remove this attribute. You can read more about it here: http://www.polymer-project.org/docs/polymer/styling.html#fouc-prevention

Follow Polymer on Google+: plus.google.com/107187849809354688692

---
You received this message because you are subscribed to the Google Groups "Polymer" group.
To unsubscribe from this group and stop receiving emails from it, send an email to polymer-dev...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/polymer-dev/fc229dc8-7f22-4c1a-8154-9b798517d2a5%40googlegroups.com.

[Joern Turner]

Thanks Rob,

i already wondered about that attribute ;) Thanks that's an interesting information.

[fab...@youfoot.com]

Thanks Rob! This is interesting! I am all for promoting the web's new standards and that's what graceful fallback indeed can help achieve. It's saying "please update your browser to get the full experience" instead of saying sorry "this new standard just broke your website." I like that.

[Marco Jakob]

Thank you for the amazing work, Rob (and others involved with Polymer)! 

I'd like to make a statement for your strategy of supporting newest browser only: Please do NOT support IE9 and before!

This is why: We really should move on. Developing for older browsers makes our lives as developers really hard. If we keep supporting old browsers the users and companies will never feel like they should update. If we collectively make a bold step into the future (or rather present), users will upgrade and our lives get much easier :-)!

I'd rather have very good support for the newest browsers in Polymer than having to wait until every possible polyfill is ready for old browsers.  

[fab...@youfoot.com]

@Marco I am sure some developers at Microsoft also said the same about staying compatible with older versions of Office, the same developers from Playstation about supporting PS3 games in PS4... "Let's make our life easier! Who cares about supporting legacy stuff!". Things that users hate and only suffer because they don't have much say about it.

Its a fundamental question as we use digital for more and more things in our lives : do we want our digital products to only have a 3 to 5 year-life span? Imagine tomorrow if a new version of JPEG came and none of your previous photos could be read? Sorry Facebook doesn't support this old thing called JPEG please upload everything again in JPEG2000 :)

Not everyone lives in a 'rich' country and has access to computers that would support the latest 2 versions of major browsers. XP doesn't support IE 10 and XP is still 25.27% of OS marketshare worldwide! It's going to take 3 years at least for that to start fading to an insignificant amount.

We work with professional and amateur football clubs: a lot of them are associations. They don't do it for the money. A lot of these associations got computers donated to them from companies; Computers that are 4 or 5 years old but still working fine. And guess what they run typically: XP. I am happy when at least they don't have IE6!

Twitter was knows better: it recently is started supporting Gif! Yes Gif this old and crappy format! Talk about old stuff! But for performance purpose Twitter converts seemlessly for users Gifs into H264 videos. That's a much better user-friendly approach.

I think it's a bad long term calculation. You should also have some respect for your users. Not everyone is a geek who cares or thinks about updating his browser and what not. But if anything, when a user goes to a website if it uses new standards it should not just break the website and expect the user to figure out that the reason the wbeiste is broken is because his browser doesn't support "web components" or what not. It should tell him "Please update your browser to get the full experience" or "Sorry you need to update you browser to visit this website" and offer links to all browser vendors so it's one click away from being a problem solved. (You'd be surprised how many people will download by mistake a version of a browser filled with useless toolbars, spamware and what not from a third party download website.)

Here is a the reality check of what people know about Browsers: https://www.youtube.com/watch?v=o4MwTvtyrUQ

I think standards have more chance evolving at a fast pace when they also address the previous standards instead of just ignoring the past. It might be fine to get a project going but if Polymer wants to become used for mainstream websites within 3 years it will have to address older browser compatibility gracefully -- in my humble opinion.

[fab...@youfoot.com]

Rob,

I am also following the progress of famo.us which is"JavaScript framework that helps you create smooth, complex UIs for any screen" that was recently created after many years of work; 

They just recently shared some thoughts with the community which I think are relevant to this "Browser compatibility" topic regarding Polymer:

Why does performance on Android still suck? (they are talking about famo.us)
In the Android ecosystem, users are slow to adopt new versions of operating systems and consequently new versions of webviews on those systems. In many—or perhaps more accurately most—situations, users don't upgrade until they buy a new phone. This means that the majority of users lag behind the current OS and webview, sometimes for years.

As of today, 65% of Android users are running the rough equivalent of an S3 device running JellyBean, which has an Android webview that is roughly equivalent to Chrome 9 (note that the current version is Chrome 35). Only 13% of the population is running the current OS (KitKat), which has an Android webview roughly equivalent of Chrome 30.

So, the problem is that to reach most of the market on Android, you are effectively building for Chrome 9 and must work within its capabilities and limitations

What was interesting is that as a result:

The Famo.us Wrapper
The wrapper will enable developers to create a pluggable webview wrapped inside their Android apps. The good news is that you can provide a Chrome 35 webview with this method. The bad news is that it comes with a 10-20MB payload. We are collaborating with another company on this project—sorry, but we can't say their name yet. This solution is in the labs and is not ready for production. We will let you know when it is available. If you'd like to participate in our beta, please contact f....@famo.us. 

Wouldn't such a wrapper be an option for Polymer as well?

[Eric Bidelman]

This seems like a great project for the community to tackle.

 

Follow Polymer on Google+: plus.google.com/107187849809354688692
---
You received this message because you are subscribed to the Google Groups "Polymer" group.
To unsubscribe from this group and stop receiving emails from it, send an email to polymer-dev...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/polymer-dev/0a278b57-26cc-40fb-8f1c-bdb316e4ad84%40googlegroups.com.

[Scott Miles]

This was mentioned in another thread: https://crosswalk-project.org/#documentation/cordova

To view this discussion on the web visit https://groups.google.com/d/msgid/polymer-dev/CACGqRCBc5FGiZGmNtQvFk_gp3eUBYefQTzha2k5s2D6LsxZYtQ%40mail.gmail.com.

[fab...@youfoot.com]

This is very interesting! Thanks for sharing.

[Jan Paul Posma]

In order to use web components in production, we have created a set of polyfills that is mostly a subset of Polymer's polyfills: https://github.com/Versal/component-runtime We intend to keep it as up to date with the moving standards as possible, while maintaining compatibility with more browsers and other libraries. For example, we haven't yet included the ShadowDOM polyfill, as it is still too buggy for us.

I hope this can be useful to some of you.

[Joern Turner]

Thanks for sharing. This can indeed be very useful for the adaption of web components.

[Johnny Larue]

Hi Rob,

Thank you and team for Polymer, it's a wonderful thing, as is DART and Angular! 

I am just leaning Polymer Dart and I plan to adopt these frameworks/languages for all future HTML 5 development. 

What would help me a lot, and likely a ton of others is to understand when will the Polymer team complete full Polyfill support for IE10+, Safari 6+, Mobile Safari?  

Currently IE10+ shows Limited (Useable) on Template, Mutation Observer, Custom Elements. 

This may be the reason that whenever I load up a Polymer based application it rarely renders in IE11.09 (Desk Top) and almost never in the Windows 8.1 (Tile View) IE Browser.

Do you see IE10+ getting full PolyFill support anytime soon (perhaps before end of 2014)?

Also, what's up with IE, why hasn't MS provided any native support of Web Components. I thought MS was part of the group pushing the HTML 5 spec?

Cheers

John

[Rob Dodson]

Hi Johnny,

My understanding is everything should be working in IE 10+, Safari 6+ and mobile safari. If something isn't working then it's probably a bug. We're doing a big push internally to upgrade our testing infrastructure so we can do a better job on cross platform/device testing. 

To view this discussion on the web visit https://groups.google.com/d/msgid/polymer-dev/3a70a298-1f9c-4440-a503-2cf981b57d09%40googlegroups.com.

[Karl Waclawek]

The old browsers are holding back progress and supporting them consumes enormous amounts of time and effort better spent on new development. Please let them die, just put a Chrome install link on your web site.

For internal/Intranet type of web applications it is even easier, as you can just say that your application platform is Chrome. One of our customers went all "nervous" on us about not supporting IE8 as this is their "browser standard". To that we replied with two facts:

• We could have developed it as a Winforms application and we know that they would have accepted that, so what's the difference?
  
• Chrome can coexist with another browser and also supports Enterprise environments.

That was the end of that discussion.

[Fabrice Lorenceau]

@Karl: Just wishing them away is not going to happen. People hang on to phone, tablets and computers they bought for as long as possible and most won't (and sometimes cannot) update the browser. There are actually plenty cases where people cant update browsers: when it's a company device, or because they bought an Android that was modified by the OEM or Network provider and will never get latest releases of Android (and thus Webview that supports Polymer hybrid Mobile Apps), or because it would perform too slowly if they updated to the latest version of the OS...

So old browsers will fade away only as users stop using their devices. Developers will rarely adopt a product that only fits less than the vast majority of their users. This is why Polymer would only -- in my opinion -- get adopted in 3 to 5 years if it only supports evergreen/modern browsers...

I really like the suggestion about using Crosswalk along with Polymer (https://crosswalk-project.org/#documentation/cordova). If I ever decide to use Polymer, that's the first thing I would look into to see if using Crosswalk we can achieve decent browser support for our users.

[Ron Truex]

@Fabrice

I disagree, companies have already started dropping support for IE8, (Google being one of them).

http://www.afterdawn.com/news/article.cfm/2014/08/08/microsoft_sets_end-of-life_for_all_internet_explorer_versions_before_11

Which means 2 years before they are cutting the browsers before IE11, but then again depending on your market Chrome is the most popular browser.  Which is the reason you have to know what your traffic is like.  If you are a new site, what do you care if you tell people to download chrome if they are running IE8 or IE9 it is for their benefit.  One of the Best parts about chrome is typically any user can install it even with out admin.  

Then depending on what stats you get your information IE8 is under 10% worldwide

http://matthewmorek.com/blog/dropping-ie8-support

And according to most of polymer's documentation IE9 should work for most of the items but will only be supported for IE10+

So if you are developing a project that will not be launched for 6 months to a year why not start looking into polymer, I think what polymer brings to the table will in the long run help make developing web sites easier and quicker for those who know how.

Ron

[Fabrice Lorenceau]

@Ron

Google is scheduled to drop support for IE8 in november (future) and that doesn't means none of its major apps won't work on IE 8 right away. There's a good chance Gmail, Google Search and more will still work. People will abandon XP now that support has ended but it will still be a few years before XP market shares drown to a few percent.

I don't think a new visitor that comes to a new site would go "Gee! They are right, I need a new browser if I wanna visit this website!"
Obviously big players have more power to force users to update. Still it sucks. Just last weekend I had to update Skype on my grandpa's ipad 2 (released in 2011), cause the new Skype App didn't support iOS 4 (cause you know it's 3 years old!). He had a PC with XP (if 3 years is 'old' surely XP was archaeology right there!) and I couldn't update iOS using his iTunes (and I couldn't get iTunes to update). So I had to back up the iPad, make the iTunes update on my laptop, upgrade iOS and then reinstall all his stuff from his computer.

So let recap: because his iPad is barely 3 years old, he was expected to buy a new PC to update his iOS in order to update Skype (If I had I not been there with my laptop). Oh and two other grandson had tried to solve his problem before me without success. And me it only took me five hours! And this is the new great paradigm?!

===

Our stats (reposting)

Internet Explorer:
Speaking of IE which amount for 15% of our sessions: 33% of these are from IE 9 or previous (almost half is from IE9 and and half from IE8, with a small amount from IE7). Fun fact: we even have a few users connecting from IE5 and IE4 apparently!

Firefox:
For us Firefox represent 22% of browser usage. While you are right that it's generally better, you'd be surprised to see we get 205 different versions of Firefox. While 80% of it are from Firefox 30 and 29 we also have almost 1% on version 12, 0,5% on version 11, etc... There might be 5% of Firefox users on versions below version 10 (with version 3 and 4 being pretty popular still).

Chrome

Chrome sees the most usage with 48,8% but while 79% are using version 35, the next version that is most used is version 1,5 with 2,4% of Chrome users. There are more than 1000 different versions being used although 990 versions probably amount to less than 2%

Safari:
Safari is only 2,9% of our usage. Among that 7% are using Safari 5 or older.

So in all it's about 10% or more that couldn't use Polymer. If you added all the people with mobilephone which webview are also lagging being and that would need an OS update to use Polymer, that would be a lot more.

[Fabrice Lorenceau]

@Ron

"Giving up on IE8, however, is markedly different than dumping IE7.
Last year, when Google said it would stop supporting IE7, that edition accounted for just 7% of all browsers used worldwide, according to Web analytics firm Net Applications.

IE8, on the other hand, was the most widely-used browser edition in the world last month, with a usage share of 25%. Of those who ran one version or another of IE, nearly half, or 47%, ran IE8 in August.

Windows XP faces its own end-of-life cutoff; Microsoft will serve users with that operating system's final security update in April 2014. But like IE8, Windows XP remains a major presence. Last month, Net Applications measured XP's global usage share at 42.5%, just behind the three-year-old Windows 7's 42.8%.

Google is the first major online software maker to drop 2009's IE8 from a support list. Microsoft, for instance, has committed to supporting IE8 on Windows 7 until 2020."
Source: http://www.computerworld.com/s/article/9231316/Google_to_drop_support_for_IE8_on_Nov._15

Some conclusions:
1. Sure the article dates from Sept 2012 but I think Google is likely to at least "go easy" on "not supporting IE8" if not postpone this completely because XP marketshare has not dropped enough since they made this announcement. As I doubt Google would skip on the relatively high number of users worldwide still using XP. Hwoever, they might be ready to fight this but you forget that they have a much higher stake in this: they are the only company which OS is the web. Educating users and pushing the industry forward is mostly helping Google. And this is fine...

2. Note however that when Google bought On2 and tried to promote WebM over H264... Still to this day Google wasn't as bold as to remove H264 support from Chrome (despite their announcement to stop supporting that dates back from in 2011), YouTube, or Android even years after the acquisition. I am not 'daring' them... I am saying it's perhaps a good thing. Shortening too much the life span of digital products that we rely on will backlash. A ton of people use H264... few products/video players can probably read WebM. Few cameras -- if any-- can record in WebM. The whole industry of video/film is not at all WebM aware or ready (at first editing software won't provide WebM, hardware and software encoders don't support it...)
3. I think using Crosswalk with Polymer (https://crosswalk-project.org/#documentation/cordova) is much better / realistic strategy to move the web and mobile forward. Recognize the past and help users with legacy hardware move forward without antagonizing them. The only other choice is that Polymer will first be a niche product (read the bestseller: "Crossing the Chasm") for the next 3 to 5 years before it can become mainstream. But it could kill it because in the meantime other techs will likely make smarter moves.

[Ron Truex]

@Fab everything I read google stop support fr ie8 at the end of 2012
http://www.computerworld.com/s/article/9231316/Google_to_drop_support_for_IE8_on_Nov._15

That article is from 2012.

I stand by what I said. And I'd your using crosswalk y not just use modernized which does some of the same things. But in my statement you have to look at your audience. By that if you already have a site running you need to look at your stats. If you are running an site and 10 percent or less is ie8 you could make a case to them to upgrade their browser. Ie8 was realized 5 years ago. So what you as a website developer are expected to support it till its 10 years old? Again if that is your clients then yes. But again developing for mobile, ie8, and modern browsers adds quite a bit extra dev depending on what you need to do.

It's is not a good item to use since if you are unable to upgrade it to ios7 sounds like an issue with either the computer or the ipad. Since apple's update goes to all devices if it doesn't work to me it sounds like there is an issue. Now if if you said android that would be different since I think about half of android devices are older then v4.

But even then if that is true. Your stats and audience would state if that matters or not. If your stats say you have a high percentage of users that doesn't support or if your worried your target audience will be on ie8 of course you won't use polymer.