Having recently undertaken some large-scale internal development work on the Conscious Platform we've been taking a look at how our testing and QA processes need to adapt to ensure that we're still putting our best foot/feet forward. A huge part of this has been re-evaluating our policy on web browser support, to bring it in-line with the new technology we're now using. We've also taken the opportunity to re-organise the policy into a more digestible format, hopefully to counteract any confusion caused by the sea of names and version numbers.
Our new system is constructed around three 'tiers' of support: Full, Partial and Unsupported. We organise the various browsers into these categories and define a level of support for the category as a whole. The current browser support tier list can be found below, as well as an explanation as to how we decide what goes where.
Tier 1 - Full Support
| Desktop Browsers | Mobile Browsers |
|---|---|
Google Chrome 33 to current (38); | iOS Safari 7 to current; |
Browsers classified under Tier 1 are tested to meet the following criteria:
- All content is fully consumable and legible.
- All functionality is useable.
- Styling and layout are rendered to specification.
- User experience is consistent across browsers and devices.
Tier 2 - Partial Support
| Desktop Browsers | Mobile Browsers |
|---|---|
Google Chrome 32; | iOS Safari 6; |
Browsers classified under Tier 2 are tested to meet the following criteria:
- All content is legible.
- Where functionality is diminished, this is graceful enough to not be obtrusive.
- Where styling and layout do not render to specification, we try to minimise any negative impact on the user experience.
Tier 3 - Unsupported
| Desktop Browsers | Mobile Browsers |
|---|---|
Opera; | iOS Safari 5 and earlier; |
Browsers classified under Tier 3 are not tested, and we're not able to make guarantees regarding user experience on these browsers.
How we Classify Web Browsers
When it comes to deciding on a level of support for the variety of browsers in use today, a degree of the decision-making process is essentially a foregone conclusion; our new responsive platform relies on some more modern technology to adapt to a user's screen size, and this technology is not generally available in older browsers. The prime example of this is Internet Explorer 8, which, at 5 years old, doesn't feature support for the underpinnings of a responsive site. Our best option in this case is to adopt a 'mobile first' methodology, building our sites on a foundation of a functional mobile layout which is progressively enhanced for larger devices. Older browsers, Internet Explorer 8 again being a notable culprit, are not able to support this progressive enhancement, and are therefore unable to adapt to display the site beyond the initial mobile base upon which everything else is constructed*. This we consider to be 'Partial Support', with the content being available, but perhaps not optimal. For more recent versions, where support for responsive design is greater, we consider these to be fully supported by our platform.
Beyond these clear-cut decisions, we also have to make some less obvious calls based on the age and usage of a browser. With the amount of granularity and variety in the browser landscape (multiple browsers, seemingly infinite versions), it is necessary to establish cut-off points for older or lesser-used versions. The more subtle background self-updating paradigm presently being used by Chrome and Firefox, as opposed to the oft-minimised "Would you like to...?" prompt, makes this more straightforward where these two browsers are concerned. Users of both browsers are now typically only spread over one or two of the more recent releases, meaning that dropping official support for older versions is acceptable, since their user base has long since moved on, even if they're unaware of having done so. Similarly, making decisions based on our active user base means that niche browsers like Opera, which only typically accounts for 0.5% of users, are categorised into Tier 3 as it is a poor use of resources to support a browser with such a relatively minor user base.
Conversely, this approach also ensures that we don't drop support for slightly older versions of browsers that are still in widespread use, such as Internet Explorer 9.
We're working on collating recent data about our users' browser usage in the second half of 2014 into a blog post, so check back for that in the next few weeks.
*While we do not support it as standard, we are able provide a quote to retrofit custom styling for Internet Explorer 8.