UK government says ‘test by popularity’

Monday 15th September

According to wasp, the UK government’s browser guidelines are “daft” and “fundamentally flawed”.

OK, I’m a little late to respond to this one, but respond I must.

Guiding principles

Overall, the tone of the guidelines is set by the following statement (emphasis mine):

The objective of this guidance is to enable you to develop a robust browser standard that is [...] reasonable (i.e. may not support old or unpopular browsers)

The author of the WaSP reaction piece is an Opera employee, and understandably takes issue with this viewpoint. The guidelines begin in an attempt to define ‘popularity’, thankfully echoing my sentiments about exercising caution when it comes to analysing web traffic. However, they go on to discuss specific browser support, and here the confusion begins:

  1. Your browser standards must support the three main operating systems: Windows, Mac and Linux
  2. Operating systems used by 2% or more of your users must be supported

It’s unclear how this conflict should be resolved but, presumably, the latter point is in addition to the former. I’d be fascinated to know if any other operating systems are anywhere near commanding more than 2% of traffic for any website’s usage. Maybe there’s a huge mobile browser contingent accessing twitter, but theis point will be hugely irrelevant to almost everyone.

Next up:

This sounds OK until the guidelines get a little bit confused with themselves in the first Appendix A example, which recommends (in this case, of course) testing only in Mozilla on Linux. Firefox is ignored. The implication may well be that, since it’s the same rendering engine, the testing need not be repeated, yet the guidelines are not solely aimed at rendering, but take in functionality too. Although functionality should, I believe, be more-or-less identical in the two Mozilla browsers, this feels like a slippery slope. How about Chrome/Safari, for example?

The guidelines then go on to state:

This demonstrates a certain naivety which is quite telling. Have the UK government not noticed the huge typographical differences between Linux, Windows, and OSX? This might be a small point but, across environments, small point sizes matter. Text on one may certainly be unreadable in another.

Moreover, this advice is tantamount to: On second thoughts, toss out everything but Firefox, and IE on Windows, and Safari on the Mac. It’s not clear whether the guidelines think Safari Mac/Windows are ‘close enough’ to allow ignorance of one.

An alternative approach

I don’t want to go on further here; there are other points I could criticise, but let’s look at an alternative methodology: standards. If we’re going to try to legislate for this at all, how about the following:

Obviously, that last point is the controversial one; maybe it should start with public companies, then be rolled out further once the point was proven. But I really think it’s about time we started pointing the finger where it belongs: at software companies that produce web content tied to specific platforms, and the IT departments that feed them.

On the plus side

However, having said that, at least the guidelines are almost on track. They do recognise that testing on every browser / OS combo under the sun is unrealistic, and wasting huge amounts of money. If only they embraced web standards a little more, and cut down on the figures, we might start getting somewhere.

Comments

Leave a comment