The Web, Mobile Web, The Internet, The Mobile Internet, Web Standards and Open standards ..

This blog is an extension of a previous blog ( There is only One Web i.e. Only one thematically consistent Web powered by Web Technologies ..) and incorporates extensive feedback from Luca Passani, Dean Bubley, Jag Minhas, James Pearce and Prof Jeff Sonstein. Many thanks for all these great comments.

The original question was:

What is the Web and What is the Mobile Web?

Is there one Web or a separate Mobile Web?

To this, I add

What is the Internet and

What is the Mobile Internet?

This blog reflects a set of comments which constitute a pragmatic view rather than an idealistic view. This was the intention behind the original blog as well – which talked of Microformats as a standardization process.

While the questions themselves sound like silly questions, they are not. Many people view the Web as ‘Online’ – however, with Offline browsing, technologies like Gears etc, that is not the case i.e. the Web can be offline

So, Is the Web primarily about the browser?

The browser itself has evolved and with technologies like browser plugins and widgets – the browser becomes a different software object than the one it started off as.

Things get MUCH more interesting if we go to the Mobile Web.


What is the ‘Mobile Web’?

Again, we have some technological grey areas.

Widgets can be based on web technologies but are not necessarily invoked from a browser. Similarly, we are likely to have offline browsing on mobile devices

Hence, to come back to the basic question:

How do we define the Web and the Mobile Web?


Is there a separate Mobile Web or is there only One Web?

But first we ask ourselves

What is the Internet and

What is the Mobile Internet?

Acknowledgements to Dean Bubley for this section:

Firstly, there is no mobile Internet, just The Internet on mobile” It’s specifically the interconnected network of networks that’s accessible via standard IP addresses.

Anything walled-garden or behind a NAT or needing a separate DNS (eg the GSMA IPX) is an IP network, maybe even an internetwork (eg military), but it’s not The Internet.

The second question is to distinguish the The Internet accessed from a mobile phone vs. The Internet accessed from another device (PC, MID etc), connected via the mobile network

A “mobile phone” is something that is (a) voice-optimised, and (b) small enough that you can also hold it to your ear.

So, what is the Web? (acknowledgements to Prof Jeff Sonstein for this section)

The Web is built on top of the Internet and uses HTTP (the HyperText Transport Protocol)

to transmit requests and responses. The Web itself does not mandate that a resource be

in any particular format (like the HyperText Markup Language or HTML) just that requests and responses flow using HTTP

When the Web was first invented HTTP (the hypertext transport protocol) was used to access HTML (hypertext mark-up language) documents only. This is not the case any more

and has not been so since very early on. Many kinds of resources are available on the Web

image files (GIFs and PNGs and so on) archive files (zip and gzip, for example) sound files

etc etc

Indeed the original term “URL” (Uniform Resource Locator) has been deprecated for a long time

in favour of “URI” (Universal Resource Indicator) simply because the Web rapidly became populated with resources which could be obtained over the Web (obtained using HTTP)

but might or might not actually be *located* at that address (see for example the discussion of REST at Wikipedia)

Anything which can be asked for via HTTP is a part of the Web and these things need not be in HTML form

The Web is largely based on Open standards.

Open standards are standards driven by consensus and created by a consortium with some form of community/committee process. The term “Open” is usually restricted to royalty-free technologies while the term “standard” is sometimes restricted to technologies approved by formalized committees that are open to participation by all interested parties and operate on a consensus basis. Protocols like BIND are examples of Open standards

Thematic consistency

The Web as seen by W3C(world wide web consortium) is more than ‘browsing’ but rather it is based on the usage of Web technologies which are open and not controlled by anyone(hence are interoperable)

The W3C see the Web to be beyond a ‘web of documents’ and believes in the idea of the One Web which means going from a web of documents to a One web of Data and Services on Everything for Everyone … where Web technologies provide the means of accessing and interacting with content via and between all devices (computing, communications, PIM, entertainment, embedded, transportation, industrial, health care, etc.) … worldwide.

Many people struggle with the idea of One Web because they view One Web in context of browsing whereas it is to be viewed more like Powered by Web technologies – at which point it makes a lot of sense of course.

From a user perspective, the vision of One Web is to make access to the Web on mobile devices as – seamless, reliable, cost-effective and useful as desktop / laptop Web access unfragmented by devices, browsers, operators, content providers …

When we talk of One Web from a user interface perspective, again there is considerable debate since people assume that it means you must serve exactly the same page format for mobile as for desktop use.

Of course, that would be silly ..

What One Web really means is Thematic consistency – a topic explained by Jo Rabin in this very insightful blog One Web – Why does this stir up such emotion?

The key is Thematic consistency (as per Jo’s blog above)

A key point here is that One Web doesn’t say that you must serve exactly the same page format for mobile as for desktop use. That would not be sensible. What it does say is that when you serve content it should the thematically similar – i.e. that a page served from a particular URI should be about the same thing, even if the format or the exact expression is not the same on different devices.

In practise, this means following the Best practices guidelines and Device description technologies

A few caveats:

a) I am a fan of Open technologies and agree to the idea of One Web as described above. However, other ‘LITE’ open technologies can also complement Web standards

b) Specifically, I am a fan of microformats,

c) And also of technologies like Wurfl from Luca Passani

d) Is there a space for proprietary technologies? In my view – yes. Technologies like Flash are clearly useful and they drive YouTube and much of the Web content. However, I see the role of proprietary technologies in consumption oriented situations rather than communication oriented scenarios.

e) Open source can also have a place in the standardization process – Open source vs. Open standards – cooperating vs. competing

Now some more additional caveats and practical comments – all from Luca Passani as per below

Should One Web be mandated by the W3C

So, what is One Web? it’s an elusive concept, so let’s start with W3C’s explanation:


“[THEMATIC_CONSISTENCY] Ensure that content provided by accessing a URI yields a thematically coherent experience when accessed from different devices.” What it means

This is a realization of the One Web (see 3.1 One Web) principle, whereby content should be accessible on a range of devices irrespective of differences in presentation capabilities and access mechanism. Web sites may paginate their content in various ways corresponding to differences in device characteristics; therefore the navigation structure of the site, and possibly its technical realization, may vary according to the device class that is being served. (See also [WebArch] Section 3.5.1).

A bookmark captured on one device should be usable on another, different type of device even if it does not yield exactly the same experience. If the page that was bookmarked is not appropriate for the device that is now using it, an alternative that is suitable should be provided.

URIs may be decorated to provide session or other information. If a URI is decorated with session information that is no longer current, then the user should be directed to a point in the navigation hierarchy that is appropriate to their device, in order to establish appropriate session and other parameters.


Now, to me this definition is a pendulum, in the sense that it can be interpreted in very different ways. One extreme is that, because of the lack of “consistency metrics”, there are no practical implications for developers and content owners. Let’s say that my mobile site lets you download ringtones, and the web site, apart from having the same logo, tells you how many wonderful ringtones you can download if you access my mobile site. Do I fulfil the requirements for Thematic Consistency? some will say yes. Some will say no. Nobody can tell for sure. There is no metrics.

The other extreme is the case where you need to make exactly the same content available. And even make sure that the content is available from the same URL. In that case, very few real mobile sites would be thematically consistent with their web counterparts. What all practitioners in mobile development know is that only a fraction of the functions available on a web site make sense to a mobile users. Because of this, content owners should make sure that the mobile UI is not cluttered with links and buttons which introduce features users are unlikely to use anyway.

But there is more: the part of the W3C definitions which ventures into giving recommendations for bookmarks always annoyed me. It shows that the W3C crowd who created BP never really built real mobile apps. Even assuming I have a unique entry point for web and mobile (by telling web clients and mobile clients apart), the most natural implementation is to redirect to different URL groups (wap.*, m*, *.mobi, /mobile/ VS. www.*, typically) and handle each class of users in the best way. Once the user is on one of those URLs and they bookmark it, the bookmark will typically be either web or mobile. Of course, there often are technical ways to redirect web users to web also from a mobile URL, and the other way around. But this is typically a lot of extra work (both wrt development and system performance) with very little added value (automatically redirecting a user to a different resource which may not even be what the user wants!).

In short, no matter how you look at this, One Web does not make any practical sense at all. It is simply a distraction from other more serious development tasks.

Why do we talk about One Web then?

I have my own explanation. Back in 2004, W3C figured that money was in mobile. Until that day, mobile had been the son of a lesser god for W3C. Until that day, W3C used to make web standards, while the mobile web was the domain of the OMA/WAP Forum crowd, which would go out of their way to stay complaint with whatever W3C produced without much attention to mobile.

FFW to 2004, How could W3C put a foot in mobile? One Web was obviously the answer: just postulate that the “web is one” and off you go: W3C is entitled to create specs for mobile. How sad.

Paradoxically, W3C has been the main responsible for the separation of web on one side and mobile on the other. This has happened as a side effect of the introduction of XHTML 1.0.

XHTML was supposed to be the evolution of HTML. The problem is that the totality of web content today is HTML. Even those who believe they are doing XHTML, are still doing HTML, because web browsers would revert to quirk mode (the tag-soup parser) whenever the HTML MIME type or transitional DTD is found. To add to that, real XHTML (MIME Type + DTD) would mean the “save as…” dialog on MSIE and endless catastrophic errors on Opera, Firefox and WebKit. Nobody uses XHTML on the web. A total failure of W3C here.

With mobile though, things are different. One can theoretically use HTML for mobile content too. But few do that for a simple reason: you have even less control on the actual rendering of your pages than with XHTML MP.

Now, let’s take one step back. XHTML MP (an OMA standard) is XHTML basic (W3C standard) + a few tags and custom CSS properties.

XHTML Basic is an abridged version of XHTML 1.0 (W3C, standard, but not practice).

In order to create mobile content, XHTML MP is the most solid mark-up there is today (which does not mean it is solid enough, but still the most solid). The overwhelming majority of mobile sites today use XHTML MP. It’s the best bet for developers.

In short, One Web is not happening. Separation is mark-up deep and W3C, the main one-web promoter, is the main responsible for this.

So, is XHTML MP is the way to go (and that’s an OMA standard)?

believe it or not, this is a complicated question.

Short Answer: Yes, it is.

Long Answer: for the time being XHTML MP is certainly the best option if you aim at building one simple “mobile web” application accessible to as many users as possible (someone might argue that WML is better than XHTML-MP fopr that, but I think I can push back on that notion in 2008. It would have been different just 2 or 3 years back).

Will XHTML MP still be the best solution 3 years from now? hard to say. Probably yes, but there are a few things to consider:

- OMA and W3C have agreed to make XHTML Basic 1.1 the common “convergent” mark-up of the mobile web.

- XHTML MP will be supported for backward compatibility for a long time to come

- XHTML Basic 1.1 adds the script tag (needed for mobile Ajax)

- XHTML Basic 1.1 deprecates the style attribute (I tried to fight that, to no avail. Anyway, deprecation means that @style is still there).

In practice, this means that developers will need to multiserve their mobile apps in the future. Even more than they are doing today already. By multiserve, I mean to serve different applications to different classes of devices in order to better exploit device capabilities and better work around device deficiencies.

So, thematic consistency is the only way to go from a customer standpoint

It is obvious that, for many sites, it makes sense to provide (in whole or in part) the same web functions to mobile users.

I would argue that this cannot be mandated, while this is exactly what W3C has been trying to do (purely for political reasons, IMO).

A company may have a web only presence, a mobile only presence, both with great overlap of functions, both with little or no overlap of functions. These are all legitimate configurations and decisions about these should be up to each company, based on their respective business model or whatever goal they happen to have.

Also (from Ajit)

Of course the issue of One Web, Open source and Open systems stretches beyond the technology. It is also a business model issue(free vs. ARPU), Uniqueness of Mobile issue(what can mobile do uniquely which the Web cannot) , at what level of the stack can the functionality be implemented The search for the IMS/NGN application: A multimedia version of Kindle (Amazon book reader) the usage of web services on mobile devices – driven by the Web or by Mobile? (Am I the only one who uses gmail on my blackberry and what does it say about Mobile applications ), Closed vs. open, Application distribution(appstores), Enablers(ex payment), Access to device APIs from the Mobile Web, Access to network APIs from the mobile Web(GSMA third party network access initiative ) and finally the meaning of the ‘network’ itself i.e. extending the cellular network to WiFi, Wimax, Bluetooth etc. Add to this issues of Context, Security and Identity and we have a complex ecosystem