W3C home > Mailing lists > Public > public-bpwg@w3.org > March 2009

The Web has flexible presentation Re: (Nielsen article)

From: Charles McCathieNevile <chaals@opera.com>
Date: Fri, 06 Mar 2009 21:50:42 +0100
To: "Luca Passani" <passani@eunet.no>, public-bpwg@w3.org
Message-ID: <op.uqd1iskswxe0ny@widsith.local>
Inter alia, Luca Passani <passani@eunet.no> wrote:

> The success of the web was based on the basic assumption that whoever  
> could publish web content

Agreed, up to here...

> and they would know what end-users would see.

Nonsense.

This is just not correct.

Further explanation for those who wonder how I arrive at this conclusion:

It was unintentionally the case when there was one client and one server,  
but the *success* of the web was that anyone could write a client for it,  
and different clients provided different renderings - wildly different  
renderings. Developers have always had *some idea* of most of the range of  
options, but it has been clear for more or less the entire history of the  
Web that developers don't know what end-users see.

> Of course, some advanced users could fiddle around with X11 settings or  
> define a custom CSS, but virtually nobody did. The basic point stands.

So the basic point is "designers know what end-users will see unless they  
decide to see it differently"?

I can accept that this is true. It seems rather different to your original  
statement, since the implication is "you no longer know what they will  
see, although you often have a fair idea about what it might be".

Looking a little further, it turns out that browsers have had options to  
change rendering in other ways that are pretty simple (compared to copying  
a User CSS from a friend, or tweaking X11 settings). Opera always had zoom  
and the option to change various other rendering features in the User  
Interface of the browser. Other browsers changed text size (which has a  
significant effect on layout with images, since the realtive size of  
images to text is itself changed), and offered other basic features from  
the interface. Likewise, since the dominant paradigm for computing became  
the "WIMP" interface, such systems have generally provided simple ways to  
change basic presentation settings - a trend that has increased over time.

CSS has been moderately successful on the web, and the premise (and very  
clear reality as the web exploded in the late 90s) is that while an author  
might provide a "preferred" rendering, clearly that wouldn't always suit  
the user or even be implemented by the browser, so it was necessary to  
ensure that the effort of providing the content to the user didn't fail  
when the author's preferred presentation wasn't what the user chose (by  
various different means).

> Clear rules are those that convinced people and companies to invest  
> resources (be it time or money) to create web content.
> This must be maintained also for mobile. If you want web content to be  
> available to mobile users, the way to go is to enable content owners to  
> "mobilize" their web content, both technologically and by creating  
> enough incentive for them to do so.  Reformatting content behind their  
> backs (or even against their will) is not the way to go. It is simply  
> cheating and it will lead to chaos.

20+ million Opera Mini users don't seem to agree with you. Banks like  
Barclays, who think about the security of their customers before they  
recommend what to do (use Opera Mini) and what not to (click on links in  
email that has been designed to be nicely formatted), appear not to agree  
with you either.

So who exactly is being "cheated" here? Enabling people to use the Web,  
and making it look "as far as possible" like other websites (given the  
limitations of some devices that is a chellenge, but many browsers  
including Mini have been marketed on their ability to do this) doesn't  
seem *to me* to be cheating those first-time web users, nor those who  
already use a site and can now access it in new ways.

Although it is possible to block specific user agents, we don't see people  
who we know block various browsers often choose to do that to OPera Mini -  
it appears that in general they are happy to have further ways for their  
users to reach them.

> Some applications are naturally mobile only (think ringtone/wallpapers  
> downloads). Some are web only. And some are a mixture of both. Content  
> owners will know. Not Novarra and not Opera.

Well, installing Opera Mini is something that might leap to mind as an  
example, no? And it turns out that we do know a few things about how  
people do that, since millions of people have done it from our website.

> Convergence will happen if it makes business sense (and I am not
> saying that necessarily it will and that it will for everyone).

I agree.

> The way Opera is trying to "enforce" convergence now is messy and tries  
> to replace what content owners have created with a bastardized version  
> of it.

This seems odd. You may choose to disbelieve me since I work for Opera,  
but let me offer an explanation of what Opera thinks it is doing, from  
soneone who works for the company and has access to such information.

We're not interested in "enforcing" ways to use content. We are trying to  
offer users access to content and services in a way that suits them, on  
wahtever platform suits them. We have worked hard to expand the ways users  
can intereact with content on a wide range of devices, providing  
full-screen overview with intelligent zooming to Opera Mini and mobile,  
offering an *optional* mobile view ("Small Screen Rendering" which splits  
it to one column) to users of almost all Opera products (mobile, mini,  
desktop, ...). We provide User Agent strings and device identifiers in  
headers so developers know what they are getting the request from, and  
*can* adapt it should they feel the need.

> As I wrote in the past, transcoding (be it proxy-based, server-side or  
> client side) is very similar to those on-line translators. While  
> automated translation may be useful to users in some cases, this does  
> not imply that you can:
> - place a translator in the middle of HTTP as a proxy

Sure you can, in a technical sense. In practice, services such as this  
exist, and people choose to use them sometimes in order to be able to get  
some access to content that they otherwise can't use. Whether they use  
this always, sometimes, or never, is a choice that users make.

Technically I presume it is possible to block translating services from a  
website, but I would be surprised if many content providers thought it was  
even an interesting possibility. There are multiple reasons why Search  
Engine Optimisation is an industry, but one of them is because many  
Content providers *want users*.

> - detect a user's region with heuristics (browser language settings,  
> accept headers, IP number...vendor dependent of course)

Huh? That depends entirely on the proxy passing on information. It may not  
get that information itself, but passing on what it gets is something that  
can be done by HTTP proxies in general, and is done by servers like the  
Opera Mini service.

> - translate everything that is not already in (what the proxy assumes to  
> be) the user's desired language

Automated transformation of any kind is an heuristic process. Depending on  
what you are doing, it may work well or badly. Transparent caching is  
something that is extremely common around the world (except in a few  
countries with outrageously good infrastructure), and is realtively simple  
for the most part. Translating natural language is known to be a difficult  
problem (it turns out that humans haven't reliably solved it yet either,  
so automating something we don't know exactly how to do is unsurprisingly  
sub-optimal).

Rendering web pages (which is what Opera Mini does) turns out to be  
something that browser developers generally know how to do (because that  
is what browsers do). It would seem that people are reasonably happy with  
the rendering options that Opera Mini provides - both developers and  
users. Are you telling me those people are *wrong* because they have  
chosen a service that suits them? (We don't force anyone to use Mini. We  
offer it, in a free market, and millions choose to take up the offer by  
actively installing and actively opening and using it).

These might mean that for many producers it is less cost-effective to do  
specific server-side adaptation than to provide something that can be  
rendered by many different web clients in a reasonable way (what browser  
does the fellow on the Clapham Omnibus use, anyway?). Different people,  
including content producers, have wildly differing expectations of how  
important their preferred presentation is, with some choosing technologies  
that pretty much force it into line (like flash) at the expense of limited  
distribution, and others choosing technology that implies wider  
distribution at the expense of total control over presentation (like the  
Web). Yet others choose a mix of some server-side adaptation, and assuming  
some adaptation on the client (the Web, again).

The Web is built on the two premises that anyone can produce a web page,  
and anyone can produce a browser to look at it in a way that suits  
someone. That's why W3C works so hard to ensure the development of open  
and free standards. It is known that this results in different approaches  
to rendering that produce different results. People have worked very hard  
to *ensure* that web browsers provide flexibility in the way things are  
viewed, and the technology is clearly designed *not* to enforce a  
particular presentation.

Whether the adaptations are done on the server side (which is one of the  
things the technology is designed to allow), or somewhere between teh  
server and the end-user (which is another thing the technology is designed  
to allow) depends on both the content producer and the content consumer.  
Each one has the ability to choose what they do - and so thoughtful web  
developers take this into account rather than simply imagining they can  
somehow control exactly what the user sees.

So, if your point is really "Web developers have total control over  
presentation", I think you are flying in the face of reality.

There is a rather more subtle and significantly more interesting point you  
have alluded to from time to time in this particular bout of  
windmill-tilting, which is to do with the advisability of making it  
invisible to the website what the client actually is and does. Perhaps  
that case is actually worth discussing.

cheers

Chaals

-- 
Charles McCathieNevile  Opera Software, Standards Group
     je parle français -- hablo español -- jeg lærer norsk
http://my.opera.com/chaals       Try Opera: http://www.opera.com
Received on Friday, 6 March 2009 20:51:35 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:43:00 UTC