Browser support for SilverStripe 4.x

250 views
Skip to first unread message

Ingo Schommer

unread,
Feb 24, 2016, 5:37:47 AM2/24/16
to silverst...@googlegroups.com
Hello everyone,

As you can tell from the various blog posts and many awesome informative inspiring discussions, we're going full steam on SilverStripe 4 development. Since a lot of the new code is related to frontend work (CSS and JS), we need to make a call on browser support very soon.

In short, we're recommending to drop IE8 and IE9 support for CMS users, and want to gather feedback on dropping IE10 support.

Paul Clarke has researched our constraints for this, and we want to give you an overview of the browser landscape in order to have an informed discussion. As always, it's a balance between moving quickly with new innovations and supporting a breadth of CMS authors. To be clear: This doesn't affect how you build your own websites.

A few facts on Microsofts support landscape:
- From January 2016, only the latest IE on a given Windows version is supported. Microsoft has a handy support table for this.
- IE8 is the last version to run on Windows Server 2003 and Windows XP; IE9 works only on Windows Vista and later. "Extended support" ended Jan 2016.
- IE9 mainstream support ended Jan 2016. Microsoft provides "extended support" for enterprise customers, incl. security fixes. This will apply to Windows 7 SP1 and Windows Server 2008 R2 for users who have not upgraded to Internet Explorer 11 (i.e. IE8, IE9, and IE10 users).
- IE9 on Windows Vista SP2 will move into "extended support" that lasts through April 2017. From that date onwards, for all practical purposes IE9 is unsupported.
- IE10 drops support for Windows Vista and will only run on Windows 7 Service Pack 1 and later. From Jan 2016, it's effectively unsupported (Windows Server 2012 only)

And the obligatory browser stats (with a sample size of 3 million sites and 16 billion page views per month):

Global - StatsCounter (Desktop only)
Jan 2016 (Jan 2015)
IE8 1.67% (1.92%)
IE9 1.32% (2.7%)
IE10 1.33% (2.09%)
IE11 10.96% (11.65%)


Global - StatsCounter (Desktop and Tablet)
Jan 2016 (Jan 2015)
IE8 1.53% (3.74%)
IE9 1.21% (2.46%)
IE10 1.22% (1.92%)
IE11 10.07% (10.63%)

So we'd make the CMS inaccessible between 2.9% (excl. IE10) and 4.3% (incl. IE10) of current users. Slightly more in in North America and Asia, slightly less in Europe and South America (and New Zealand!).

In terms of feature support, a minimum requirement of IE10 would allow us to remove pretty much all polyfills (while still transpiling ES6 of course). Dropping IE9 would allow us to natively support history.pushState without polyfills. IE11 makes CSS flexbox more viable for a fast and concise application layout. IE10 is the first version with webworker support.

The case for not supporting IE10 isn't very strong in terms of feature support, but we would like to simplify our quality assurance by limiting tested browsers. We'll already have our hands full with different IE Edge versions.

As a comparison, Wordpress Calypso (and by extension, wordpress.com) already only supports IE11+ today - and they'll have a very large global user base to consider.

SilverStripe 4.x will come out "sometime in 2016", so we'd expect legacy browser use to drop further. Since SilverStripe 4.x follows semantic versioning, we can't change browser requirements until the 5.x release line - so are "stuck" with this choice at least until mid 2017.

Ingo

Thomas B. Nielsen (Nobrainer Web)

unread,
Feb 24, 2016, 10:00:38 AM2/24/16
to SilverStripe Core Development
Drop

Roman

unread,
Feb 24, 2016, 10:10:18 AM2/24/16
to silverst...@googlegroups.com
I'm in favor of dropping support for IE10 and lower. Basically the same as WP does.
But this is easy to say for me, since I don't have companies as clients where old versions of IE stick around due to IT policies ;)
--
You received this message because you are subscribed to the Google Groups "SilverStripe Core Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email to silverstripe-d...@googlegroups.com.
To post to this group, send email to silverst...@googlegroups.com.
Visit this group at https://groups.google.com/group/silverstripe-dev.
For more options, visit https://groups.google.com/d/optout.

Jonathon Menz

unread,
Feb 24, 2016, 2:12:44 PM2/24/16
to SilverStripe Core Development
I'm in favour of dropping IE10 too.

Corey Sewell

unread,
Feb 24, 2016, 2:19:46 PM2/24/16
to SilverStripe Core Development

+1 For dropping IE10 and lower

Stevie Mayhew

unread,
Feb 24, 2016, 2:28:39 PM2/24/16
to SilverStripe Core Development
I am for dropping all support for browsers which do not have continuous update functionality.

Paul Clarke

unread,
Feb 24, 2016, 4:08:34 PM2/24/16
to SilverStripe Core Development
Usage for IE10 is going down as fast as IE9, I would think IE10 would be below or pretty much the 1% mark in half a year (along with IE9). We would be doing these users a favour by giving them a reason to upgrade :). I'm for it
browser stats.png

Ralph Slooten

unread,
Feb 24, 2016, 4:19:37 PM2/24/16
to silverstripe-dev
In theory I agree with everyone about dropping IE10, however I wonder if this decision is "just because", or whether continuing to support IE10 actually poses a real significant compatibility/maintainability issue in the CMS?

For example I have two "largish" clients who (due to their "IT department & strict policies" ~ don't even get me started...) are bound to using older IE versions. I realise that in these situations one could simply just stick with SS3 - problem solved - however if dropping IE10 is really just a matter of principal, then I don't see the advantage.

Just curious?

simone...@silverstripe.com

unread,
Feb 24, 2016, 4:39:21 PM2/24/16
to SilverStripe Core Development
I agree. As of IE9, IE isn't that bad anymore. I think it'll just work fine, even if it's not officially supported.

Nicolaas Thiemen Francken - Sunny Side Up

unread,
Feb 24, 2016, 4:42:09 PM2/24/16
to silverstripe-dev
I wonder what would happen if we were to drop support for IE10 and you'd use the CMS with IE10.  What scenario are we going for?

a. would the user get a message saying that we are now in the 21st century, basically stopping people from using the CMS with IE10.
b. would the user not be able to use the CMS at all because major functionality is broken.
c. would the user be able to use the CMS reasonably well with some annoyance?

If we are dropping the support, as suggested, then I think it would still be worthwhile assessing the work involved in getting to (c). If getting to (c) would require just a few hacks then why not add them (danger danger!) while continuing the not supported status for IE10. 

@Ralph, by the time you will be using 4.0 would it be likely that your client has fired their IT manager? 

Jonathon Menz

unread,
Feb 24, 2016, 5:00:12 PM2/24/16
to SilverStripe Core Development, n...@sunnysideup.co.nz
I think if you decide not to support a browser the answer is d) all or none of the above. We would basically just ignore it exists I think. We wouldn't code things in a way that deliberately excludes unsupported browsers, so the likelihood of IE10 continuing to work for a while is probably pretty good, but further down the track there may be major problems.

Mark Guinn

unread,
Feb 24, 2016, 5:09:52 PM2/24/16
to silverst...@googlegroups.com
I’m all for dropping 9 and 10. I think it’s possible that someone concerned with supporting those two could write a module with some extra CSS and js polyfills too.


Paul Clarke

unread,
Feb 24, 2016, 5:20:15 PM2/24/16
to SilverStripe Core Development, n...@sunnysideup.co.nz
Microsoft itself has dropped support for IE10 along with IE9 seven years prior to what it was going to support it, also anyone who has IE10 should be receiving upgrades to IE11 without too many issues. I guessing the clients Ralph was talking about are stuck on IE8 or IE9, and not stuck on IE10. Saying that I don't see that the CMS couldn't be used in IE10 (but I also don't have a crystal ball unfortunately), it just wouldn't be encouraged to use if it was to drop.

Ralph Slooten

unread,
Feb 24, 2016, 5:58:55 PM2/24/16
to silverstripe-dev
Not necessarily always upgradable unfortunately - these clients are stuck on whatever is dictated up the food chain, and this kind of scenario is not unique. One of these clients is still using IE8 dictated by the international corporate head office in Australia (though I am in no way suggesting to support 8!). The other client was "upgraded" to 10 only about 6 months ago, again dictated by their national IT department. Neither client has the ability to upgrade manually, and/or install other software (ironically citing security concerns). @Nic I don't know, but highly unlikely ;-)

I'm not sure exactly why exactly these particular versions of IE - I just assume it has to do with software integration with other software as 12 years ago I was in the same support role for a software package that required a particular older version of IE.

Anyway, from what I gather IE10 should still continue to work, hopefully. I also think that it would be good to keep IE10 compatibility provided it doesn't result in lots of CMS maintainability issues.

Patrick Nelson

unread,
Feb 24, 2016, 9:04:22 PM2/24/16
to silverst...@googlegroups.com
Let's put it this way. If you're stuck with IE10 due to IT department restrictions, then you're also going to be stuck with SS 3.x or maybe a somewhat hindered SS 4.x. If the difference between IE10 and IE11 isn't that great and the issues that do pop up are a big deal to whoever needs to support it then I think they should have that responsibility, since the rest of the community will be able to move forward with more advanced (and simpler) implementations which do not account for quirks in older browsers.

My 2c anyway. And I work for a large company which has (thankfully) updated everyone to IE11. 

Sent from my iPad

Anselm Christophersen

unread,
Feb 25, 2016, 1:37:32 AM2/25/16
to silverst...@googlegroups.com
I like Mark’s idea of that whoever needs IE10 support could write a polyfill.
I doubt there’ll be big issues with IE10 anyway, but like to spearhead and be able to go ahead and user newer technologies in the future without having to have to consider IE10.
I doubt it’ll happen for the release of 4, but it’s nice if it’ll be possible before the release of 5.

Naomi Guyer

unread,
Feb 25, 2016, 1:57:34 AM2/25/16
to silverst...@googlegroups.com
+1 to drop IE10. 

Nedmas

unread,
Feb 25, 2016, 4:54:21 AM2/25/16
to SilverStripe Core Development
As much as I would dearly love to support dropping IE10, I am in a similar situation to @Ralph with clients that are stuck with old technology. I work for an agency in the UK where large parts of the public sector still run Windows XP! And although @Paul is correct that Microsoft has dropped official support, they do still offer private support contracts for outdate operating systems and browsers.

We as an agency will have to continue to support our current and future clients on IE8 & IE9 and understand that SilverStripe 4 won't be the right fit for those clients. This does raise the question of how long SilverStripe 3 will continue to be supported? Which I think is also relevant to the issue at hand.

If SilverStripe 3 will continue to be supported then that makes it easier from our point of view to support dropping IE10 in SilverStripe 4. If however SilverStripe 3 won't be receiving continued support then it would make it harder to support dropping IE10.

swaiba

unread,
Feb 25, 2016, 10:09:06 AM2/25/16
to SilverStripe Core Development
I'd have to echo this - much of our work is with local authorities and they are very stuck to IE9 and IE10 - I'm very happy to say I don't hear about older versions anymore.
It isn't about the smallness of the numbers when the bulk of those folk are in one group (i.e. local authorities) and are very unlikely to move forward quickly.

Patrick Nelson

unread,
Feb 25, 2016, 11:22:14 AM2/25/16
to silverst...@googlegroups.com
Nedmas said:
 
We as an agency will have to continue to support our current and future clients on IE8 & IE9 and understand that SilverStripe 4 won't be the right fit for those clients. This does raise the question of how long SilverStripe 3 will continue to be supported? Which I think is also relevant to the issue at hand.

My assumption is that no major changes would be implemented that wouldn't be supported in the already defined list of supported browsers for that version anyway. At least that should continue to be the case. A

Swaiba said:

I'd have to echo this - much of our work is with local authorities and they are very stuck to IE9 and IE10 - I'm very happy to say I don't hear about older versions anymore.
It isn't about the smallness of the numbers when the bulk of those folk are in one group (i.e. local authorities) and are very unlikely to move forward quickly.

If you think about it though, since this is for a totally new (and as of yet unreleased) version of SilverStripe, they will be able to think about if they wish to use SS 4.x. One of the considerations the developer need to take into account is the fact that the primary user base for the CMS is still constrained to IE9 (for example). If that's the case, they should simply install the latest version of SS 3.x and ensure that the powers that be (primarily IT) are informed of the additional technical debt that they are incurring upon themselves by not upgrading to IE11 (if indeed it's worthwhile to really stick with SS 3.x), due to the backward incompatible API changes implicit in a major version change, should they ever decide to upgrade to SS 4.x ex post facto.



--

swaiba

unread,
Feb 25, 2016, 11:38:19 AM2/25/16
to SilverStripe Core Development

Swaiba said:

I'd have to echo this - much of our work is with local authorities and they are very stuck to IE9 and IE10 - I'm very happy to say I don't hear about older versions anymore.
It isn't about the smallness of the numbers when the bulk of those folk are in one group (i.e. local authorities) and are very unlikely to move forward quickly.

If you think about it though, since this is for a totally new (and as of yet unreleased) version of SilverStripe, they will be able to think about if they wish to use SS 4.x. One of the considerations the developer need to take into account is the fact that the primary user base for the CMS is still constrained to IE9 (for example). If that's the case, they should simply install the latest version of SS 3.x and ensure that the powers that be (primarily IT) are informed of the additional technical debt that they are incurring upon themselves by not upgrading to IE11 (if indeed it's worthwhile to really stick with SS 3.x), due to the backward incompatible API changes implicit in a major version change, should they ever decide to upgrade to SS 4.x ex post facto.


That is one way to think about it if every site was a simple CMS, another way is within our business where we have a product based on SS which relies alot on these folk that are slow to move forward with this - if there is a potential that SS4 isn't going to handle this well then we would be slow to move to SS4 - which is all I'm highlighting.  I'm not happy about it but it is then a difficult decision internally as to adoption of SS4.

Nedmas

unread,
Feb 25, 2016, 11:41:16 AM2/25/16
to SilverStripe Core Development
Patrick said: 
My assumption is that no major changes would be implemented that wouldn't be supported in the already defined list of supported browsers for that version anyway. At least that should continue to be the case. A
 
What Patrick stated would be my assumption as well, however I was trying to get at how long SilverStripe 3 would be officially supported with security and bug fixes? Maybe my question wasn't clear enough.

My concern is that SilverStripe 4 will be released and before SilverStripe 5 is released support for SilverStripe 3 will be dropped. This would leave our customers with no officially supported version of SilverStripe that they could access. Which would be a big problem for us.

Patrick Nelson

unread,
Feb 25, 2016, 12:53:50 PM2/25/16
to silverst...@googlegroups.com
Per Nedmas :

My concern is that SilverStripe 4 will be released and before SilverStripe 5 is released support for SilverStripe 3 will be dropped. This would leave our customers with no officially supported version of SilverStripe that they could access. Which would be a big problem for us.

Welcome to the world of enterprise software ;) i.e. supporting legacy code!

--

David Alexander

unread,
Feb 25, 2016, 2:35:00 PM2/25/16
to SilverStripe Core Development
What about the idea of SS adopting an Ubuntu style release cycle with LTS (Long term support) versions ?


Such a schedule would satisfy the needs of both Legacy Laggards and Bleeding Edgers. If SS3 were deemed LTS, then development on SS4 can ignore legacy issues altogether and push the envelope by using the latest and greatest, even experimental features.

Ralph Slooten

unread,
Feb 25, 2016, 3:23:45 PM2/25/16
to silverstripe-dev
On 26 February 2016 at 08:34, David Alexander <da...@spiritlevel.nz> wrote:
What about the idea of SS adopting an Ubuntu style release cycle with LTS (Long term support) versions ?

I think this is a great idea if SilverStripe is seriously intending to drop all support except the latest browsers and IE11, specifically for the reasons mentioned before. We're not talking about any new features/development in 3, just security patches where required to keep it going for a few more years (?).
 


Such a schedule would satisfy the needs of both Legacy Laggards and Bleeding Edgers. If SS3 were deemed LTS, then development on SS4 can ignore legacy issues altogether and push the envelope by using the latest and greatest, even experimental features.


I totally agree with David. The downside is that it creates an extra branch to maintain, but at the same time frees up SS4 to move forward, effectively satisfying existing as well as new "customers". If/when we (developers) hit the issue of legacy browsers then at least there is a choice that doesn't mean the client has to run unsupported (upstream) software, at least for the next couple of years.

I'd be very keen to hear what SilverStripe themselves have to say about this idea.

Sam Minnée

unread,
Feb 25, 2016, 4:25:48 PM2/25/16
to silverst...@googlegroups.com
We as an agency will have to continue to support our current and future clients on IE8 & IE9 and understand that SilverStripe 4 won't be the right fit for those clients. This does raise the question of how long SilverStripe 3 will continue to be supported? Which I think is also relevant to the issue at hand.

We've yet to formalise that, but it looks likely that we'll be providing security fixes until at least March 2018. By that stage Windows Vista (which is what locks people into IE9) will be out of Microsoft's extended support and anyone still running it will be well into no-mans land.

--
Sam Minnée
CEO
SilverStripe Limited

Ingo Schommer

unread,
Mar 2, 2016, 5:27:35 PM3/2/16
to SilverStripe Core Development
Thanks for your input everyone! We've held a core committers vote at the bi-monthly core committers hangout, and voted to drop IE8 and IE9 support from SilverStripe 4.
This was following a discussion about our Long Term Support (LTS) plans in the same hangout, which we'll detail here once they're solidified.
In short, we'll support SilverStripe 3.x for longer than Microsoft has "extended support" for IE9.

We specifically haven't voted on dropping IE10. Judging from the feedback above,
most of the "legacy users" seem to be on IE8/IE9 rather than IE10.
Would dropping IE10 pose any more of an upgrade barrier for anyone here?

Keep in mind that IE10 is already unsupported by Microsoft since Jan 2016
for all OS capable of running IE11, which is Windows 7 SP1 or later.

Thanks
Ingo

Sam Minnée

unread,
Nov 17, 2016, 8:02:43 PM11/17/16
to SilverStripe Core Development
Hi everyone,

In order to help decide whether to drop IE10 support in SS4, we've created this card to let us know your usage circumstances. Please respond and share it widely! :-)


Thanks,
Sam
Reply all
Reply to author
Forward
0 new messages