Design bomb in W3C’s garden

The W3C's new logo for HTML5 with a whacking great question mark next to it. A post came up in my feed called the BOMB in the GARDEN by Matthew Butterick. It is from a talk he did in April 2013, but I just saw it after an update more recently. I found it… frustrating. Phrases went through my head like “flawed assumptions”. But I left it a couple of days and I suspect the frustration comes from a different perspective and different assumptions. I try to show those below.

I appreciate being made to think about this, and whilst I disagree with a lot of the content, it did make me think which I hope Butterick would consider a success. I should also point out my company is a member of the W3C, and I’m the (fairly recent) ‘rep’ to the W3C. My time has mostly been in the Authoring Tool Accesibility Guidelines committee.

Sources of design excellence

A lot of the world’s news­pa­pers are very well de­signed pub­li­ca­tions. We know they have de­sign­ers on staff. We know they have bud­get for de­sign. OK, so let’s go look at their websites.

I would question that organisations who’s history is primarily in print would be a good source of web design excellence. Why not look to the pure-digital providers? Depending on how much of the point is about advertising, I would look at publications like Alistapart, or Daring Fireball that makes a living from very subtle adverts, or even less typography based sites. (Pinterest?)

Design isn’t my strong point, it’s really the technical and standards stuff later that I want to pick up on.

Comparing to PDF

Because PDF is an ISO stan­dard. It just works.

This is a great point against what Butterick is arguing. Yes, simple PDFs ‘just work’, so do simple HTML pages. There is a lot in the PDF standard that doesn’t work that well across readers. Try PDFs which use forms, scripting, embedded movies, Flash or other interactive elements in non-Adobe products. I see these type of PDFs from organisations that only use MS & Adobe products. They don’t work in Preview, or other non-Adobe readers. I’m not criticising Adobe, it was a good thing to make PDF a (set of) standards. However, it is a good case of an expansive standard and only one full implementation.

Simple HTML just works to, PDF isn’t a case where this is different.

Directing standards

About Tim Berners-Lee:

he’s not the boss of the web, but he’s an influential voice in set­ting the agen­da for what the web stan­dards are.

He has a voice, of course, but the standards developed by the W3C are essentially driven by implementors. If someone wants to implement a payment mechanism, they can do so. They might choose to do so under the auspices of the W3C, but no-one is forced to.

If you want to complain a standard hasn’t been created, talk to companies who might create an implementation. If it takes co-ordination between different stakeholders and uses the web, the W3C might be the place to do it.

Compared to mobile platforms

Apparently on mobile platforms like iOS:

“the tech­ni­cal stan­dards are to­tal­ly stan­dard”

For iOS, Windows, and Android to some degree. But the web is not homogenous, you develop one site for multiple platforms. The argument isn’t about application standards vs web standards, it is about developing across multiple platforms, or develop once for the web.

Assuming you are not going iOS only, both approaches require cross-device testing, but the first requires completely separate and different code bases with different skills for each.

On programming

I hate HTML and CSS. That’s why I like pro­gram­ming…. HTML and CSS—those are not pro­gram­ming lan­guages. Those are data-en­try for­mats.

I prefer HTML/CSS, but that’s just preference. They are declarative languages, not programmatic, and that makes the web much more robust. They also lower the bar for creating sites. There is a place for programmatic websites/web-apps, but it would raise the bar to make the web require programming.


The web should re­al­ly en­cour­age au­thors to write books. But it doesn’t. It re­al­ly doesn’t. The best you can do on the web is make an ugly book and give it away for free.

ePub is essentially web-based (HTML, CSS etc), just packaged up in a wrapper (with optional DRM). Given that eBooks basically are web pages, I’m not sure what improvement Butterick is aiming for? Should web pages match ePub? They already do, just with a different set of browsers and a DRM wrapper.

Do we need a central web standards organisation?

rather than pun­ish­ing the bad browsers for their bad be­hav­ior, you’re in­dulging them

The problem there is that people use browsers, so taking action against ‘bad’ browsers means you punish users by proxy. Unless of course you are going to mandate what technology people use?

Would Microsoft (for example) stop producing IE if they were thrown out of the W3C? I don’t think so. So who would you punish?

The idea that stan­dards couldn’t hap­pen with­out the W3C to ‘bro­ker’ them is plain­ly false… if you wor­ry that the al­ter­na­tive is a web ruled by Google and Microsoft and Apple—I don’t think so.

I would worry, we saw that happened: the browser wars. Developers got so pissed off with all the divergent implementations they started the Web Standards project (WaSP). Then IE won, and we had a period of stagnation with one browser as the defacto standard.

PDF is an ISO standard, but how many good implementations of it are there? Have you tried filling in PDF forms in anything except Acrobat? Or even opening PDFs build for Acrobat X which use internal navigation and Flash objects.

As a solution, Butterick suggests that open source would be a better model it is has a:

different pow­er struc­ture: a ‘benev­o­lent dictator’ who is the ul­ti­mate ar­biter.

I’m not sure how different that is from the chair of a W3C committee? In fact, for the WHATWG Ian Hixie is that role, has been for a while.

A big reason that the open-source model wouldn’t help with a standard like HTML is that it has so many stakeholders. Benevolent dictator of an application (e.g. browser) is one thing, but also of users, developers, and tool makers at the same time?

Also, if people simply fork off a ‘standard’, how would an implementor know what to do? We end up with fragmented standards and implementators with no clear direction.

Web stan­dards can be—should be—defined by a ref­er­ence im­ple­men­ta­tion.

That sounds like the IE6 era. Perhaps Butterick is thinking of webkit on mobile? PPK found over 19 versions of webkit on different mobile devices, each with their own bugs. I don’t see that helping to keep a vibrant ecosystem.

As long as the W3C charges as much as it does for mem­ber­ship, mem­ber­ship will re­main re­strict­ed most­ly to wealthy cor­po­ra­tions

Check the fees, smaller organisations can join and don’t pay nearly that much. We (Nomensa) are a member, we’re a small/medium business, not a wealthly corporation.

More importantly, there is the ‘invited expert’ method used by many committees to include representatives who have expertise, no matter the organisation or lack thereof. There is also the open-source style HTML5 membership list where anyone can join.

af­ter many years of keep­ing DRM out of open web stan­dards, the W3C is prepar­ing to adopt a DRM proposal

There is a point to be made about whether DRM (or at least a window to it) should be included in HTML5. But I wonder how Butterick squares this with wanting payment for things like books? How would you enable charging for books on the web without that?

The only alternatives I’ve heard of lack DRM and don’t have any other means of preventing copying. That is a perfectly valid approach, but there isn’t an obvious technical fix that would need a standard apart from DRM.

the W3C has been drag­ging its feet on Do Not Track, which would en­hance in­di­vid­ual pri­va­cy but re­duce the ad rev­enue of W3C mem­bers

That’s a great example where you simply wouldn’t get a standard without a body overseeing it. We probably won’t see an agreed standard because members disagree so much and can’t find a compromise. But how is that better without a standards body? The fact we can see that committee failing to agree means we can see what’s going on!

Of course there are things the W3C could improve, but I don’t see any good coming from Butterick’s suggestions.

One contribution to “Design bomb in W3C’s garden

  1. IMHO H5 is designed for advanced interaction (in opposite to pdf). With CSS3 it introduces and standardizes many functions (animations, audio, 3D elements) which until now people tried to implement with big effort or using external technologies like java applets or flash.

Comments are closed.