(Browser) Death and Taxes

[Brian Kardell]

I recently posted an article encouraging the community to build a campaign around nudging people as hard as possible to adopt evergreen browsers.  I thought that sites could consider how to participate/how hard to push by either a UK cookie bar style campaign or a SOPA blackout style (which didn't actually black out, it was just a splash screen that you click through).  I know I'm not the first, but I am specifically interested in moving to an evergreen web because without that we are doomed to just keep repeating some (unnecessarily) painful bits of history.  

In retrospect it was a probably mistake to present it the way that I did, but in that post I included something that I still think is related (how we declare a browser dead and what we do about it) and that seems to have taken over the conversation instead of the larger thrust about why evergreen is a Really Good Thing for both developers and users.  

When you get right down to it this is all about two things:  feature gaps  and the lag/lead times to close those gaps or add new ones (see https://rawgithub.com/bkardell/gaps/master/gaps-caniuse-2-2013.html or this https://rawgithub.com/bkardell/gaps/master/gaps-webbrowsercompatability-2-2013.html).  The problem is that the old model is very slow to catch up and the new model is very fast to pull ahead meaning that the gaps are becoming a giant and growing uncrossable canyon.  Lots of people jumped in with things about punishing/taxing users unfairly or how we should polyfill or progressively enhance - perhaps it is just me but I feel like they have missed my point that this becomes more and more impractical as the leading browsers set ever increasing UX expectations and provide more and more features allowing people to imagine with.

Historically we have this distinction where we declare a browser "dead" because it falls below some threshold of users reporting it, and no one really seems to have a problem with that, it is really the status quo.  A bunch of people said "progressive enhancement" or "polyfill", and like I said that's good - but I've yet to see a case where this makes sites infinitely backward compatible and the more you have to fill the more code you are giving the least capable browsers to deal with.  At some point sites really do stop working for a given browser - or the experience would be so bad that the company would rather not put you through it.  I began a post in response to my first one which I have since decided to delay and rethink, but you can see screenshots http://briankardell.wordpress.com/?p=180&shareadraft=511be2761ca08 illustrating what I am talking about...  There is no clean actual "death of a browser" - even 1% of users is still something like 240+ million users.   It seems to me that browsers really do "die" when sites stop working with them altogether and provide no explanation as to why and that all of these are, in fact, heavy taxes on users that most of them aren't even aware of and we can do much better.

[Brian Kardell]

I suppose actually I should clarify the last bit there... I mean to say:  Historically, we "stop supporting" generally based on % of users - it's never when that number reaches 0.   "Stops supporting" generally means "will probably break horribly... at some point... but we won't tell you when, and when it happens we won't tell you why."  That is, in my mind, sort of like failing to tell you that taxes are due until they are way overdue and then just taking you straight to jail :)

[Alex Russell]

Hey Brian,

Great to see you here! Comments inline....and apologies for not
responding sooner.

On Wed, Feb 13, 2013 at 8:29 PM, Brian Kardell <bkar...@gmail.com> wrote:
> I recently posted an article encouraging the community to build a campaign
> around nudging people as hard as possible to adopt evergreen browsers. I
> thought that sites could consider how to participate/how hard to push by
> either a UK cookie bar style campaign or a SOPA blackout style (which didn't
> actually black out, it was just a splash screen that you click through). I
> know I'm not the first, but I am specifically interested in moving to an
> evergreen web because without that we are doomed to just keep repeating some
> (unnecessarily) painful bits of history.
>
> In retrospect it was a probably mistake to present it the way that I did,
> but in that post I included something that I still think is related (how we
> declare a browser dead and what we do about it) and that seems to have taken
> over the conversation instead of the larger thrust about why evergreen is a
> Really Good Thing for both developers and users.
>
> When you get right down to it this is all about two things: feature gaps
> and the lag/lead times to close those gaps or add new ones (see
> https://rawgithub.com/bkardell/gaps/master/gaps-caniuse-2-2013.html or this
> https://rawgithub.com/bkardell/gaps/master/gaps-webbrowsercompatability-2-2013.html).
> The problem is that the old model is very slow to catch up and the new model
> is very fast to pull ahead meaning that the gaps are becoming a giant and
> growing uncrossable canyon.

That framing feels right to me. Webdevs need to put browsers in
"buckets", designing for the minimum number of discrete buckets while
still trying to reach the most users with the best experience. For
everyone else, there's a "fallback", although that phrase feels wrong.

Something the GDS folks blogged struck a chord, using the phrase
"inclusive design":

http://digital.cabinetoffice.gov.uk/2012/10/01/building-for-inclusion/

It seems to that's the right frame for all of this. The goal isn't to
get everyone into great browsers so we can stop doing the hard work of
making adaptable content, it's to ensure that when we *do* do
inclusive design, the most people get the best experience.

> Lots of people jumped in with things about
> punishing/taxing users unfairly or how we should polyfill or progressively
> enhance - perhaps it is just me but I feel like they have missed my point
> that this becomes more and more impractical as the leading browsers set ever
> increasing UX expectations and provide more and more features allowing
> people to imagine with.
>
>
> Historically we have this distinction where we declare a browser "dead"
> because it falls below some threshold of users reporting it, and no one
> really seems to have a problem with that, it is really the status quo. A
> bunch of people said "progressive enhancement" or "polyfill", and like I
> said that's good - but I've yet to see a case where this makes sites
> infinitely backward compatible and the more you have to fill the more code
> you are giving the least capable browsers to deal with.

The promise of HTML is that it will work everywhere; that your content
won't cease to be available just because your processing agent isn't
of some narrow make or model. Giving up on that ideal would be bad for
many of the web's users and developers (eventually).

> At some point sites
> really do stop working for a given browser - or the experience would be so
> bad that the company would rather not put you through it.

Hmm. I think it's fair to say that's Doing It Wrong (TM). Building for
inclusion means starting from markup that might not hit every client
with full-fidelity, but will allow the main bits to shine through
everywhere (assuming the main bit isn't some bit of art or WebGL-only
content).

> I began a post in
> response to my first one which I have since decided to delay and rethink,
> but you can see screenshots
> http://briankardell.wordpress.com/?p=180&shareadraft=511be2761ca08
> illustrating what I am talking about... There is no clean actual "death of
> a browser" - even 1% of users is still something like 240+ million users.
> It seems to me that browsers really do "die" when sites stop working with
> them altogether and provide no explanation as to why and that all of these
> are, in fact, heavy taxes on users that most of them aren't even aware of
> and we can do much better.
>

> --
> More at etherpad.mozilla.org/evergreen . Group at
> groups.google.com/d/forum/evergreen-the-web
> ---
> You received this message because you are subscribed to the Google Groups
> "Operation Evergreen" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to evergreen-the-...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>