Tim Berners-Lee: Just Say "NO" to .mobi Domains

[Pat Logsdon]

<div class='os_post_top_link'><a href='http://www.brighthand.com/article/Web_Founder_Against_Mobile-Only_Domain' target='_blank'>http://www.brighthand.com/article/W...ile-Only_Domain</a><br /><br /></div>"Last March, several leading companies from the mobile industry -- including Microsoft, HP, and Nokia -- requested that a Top Level Domain (TLD) be created for web sites that are intended solely for handhelds and smartphones. Recently, Tim Berners-Lee, the inventor of the technology behind the World Wide Web, came out in opposition to this plan. At the time it made the request to the Internet Corporation for Assigned Names and Numbers (ICANN), the group of companies said they hoped a '.mobi' TLD will create business opportunities for companies servicing mobile customers. They also hoped it will spur the development of compelling new mobile services and applications. <br /><br />However, the W3C Technical Advisory Group, of which Mr. Berners-Lee is a member, has released a statement that says, in part, 'There are major problems with these proposals. There are costs in general to creating any new top level domain. There are specific ways in which the '.mobi' breaks the Web architecture of links, and attacks the universality of the Web.' Rather than creating special versions of web site appropriate for different devices, the W3C Technical Advisory Group would prefer that all pages be able to handle any type of device. According to it, this can be done through the use of Style Sheets and a variety of new methods just coming into use for the client to be able to tell the server what its capabilities are. These will allow the server to only send content appropriate for the client." <br /><br />What do you think? Are there merits to creating a .mobi TLD that outweigh the damage the W3C describes, or should everyone design pages that work on all devices?

[Zack Mahdavi]

I couldn't agree more with Tim Berners Lee. The internet is a "world wide web" that should be easily accessible by any type of device from anywhere in the world. Trying to separate the "mobile internet" from the "regular internet" would be a bad idea in my opinion.

I get very annoyed when I read stories about how Microsoft doesn't want to follow the W3's web standards. It bothers me that a company with such a high browser market share is trying to prevent other browsers from competing with them. The world wide web is a global resource and shouldn't be tailored to any specific browser or platform.

[Sven Johannsen]

Why should I have to go to www.pocketpcthoughts.mobi when on my PPC and www.pocketpcthoughts.com on my desktop. Jason and company have to create two formats in either case. Let him figure out what I've got and deliver the right stuff ;)

[James Fee]

Were does it end? Should anyone typing james.fee into the address bar take you to my homepage? 0X

[Shadowcat]

I think I disagree with the .mobi TLD as well. It will force people to create two versions of their websites; I believe many may just decide it's not worth the trouble. In addition, I believe the average user would just get confused if a website looks different and perhaps offers less features if it is viewed on a mobile device instead of a desktop. Most people don't want to bother learning a new interface to do the same thing.

I think the true solution would be for website designers to create mobile-friendly designs and for developers to add more features to their browsers. Perhaps VGA screens (I'm really not sure about this one) would improve mobile browsing experiences?

[paris]

yap i tolly agree with them, a website should check who is acessing it and act acconrdingly, if its a mobile phone load the wml version, if its a pocket pc a light html version and if its a normal pc the full page. Now even with XML (RSS feeds) there are more ways to effectively make availalbe a site content to a wide variety of devices. A .mobi domain would only cause problems, i would never register an aditional domain just for mobile devices.

http://www.pocketpcthoughts.com/xml/ &lt;- use this effectively and you can utilize any device out there

[iPaqDude]

The technology exists for a website to check on the browser type and either redirect the user or generate the appropriate presentatiion for the given device/browser. My take would be for this versus the .mobi extension - why should the user have to do this when the site is perfectly capable to do so?

If you rely on the user typing in the correct extension, you loose consistancy - some websites will provide the content based on browser type, some will end up requiring the user to select the type. My vote is on the website looking at the browser or device type and delivering the content appropriately.

Just my humble .02....

[Hyperluminal]

I agree with Tim Berners-Lee also.
In addition to what was said already, what about cybersquatting? Or simply not caring if you register a .mobi domain where someone has a corresponding .com? That would ruin the system entirely. Imagine pocketpcthoughtsmobile.mobi, or something like that. :roll:

I guess they could have a system where you can only register a .mobi if you have a corresponding .com, but that brings out another problem: what do you do with two groups who the same original domain name but different TLDs, like domain.com and domain.net? Who gets the .mobi? One would have the break the naming system and change their domain name for the .mobi version.
It's a mess.. they really should just make the servers adapt to the browser...

[Kacey Green]

I agree as well, for my TLD, I share with a land trust, I have the .com and they have the .net, not to mention I really don't want to do the extra work, when the server can detect the clients browser, and load the page I've already (have yet to do this) made for mobile devices.

[hang5lngbd]

With more and more devices coming out that have a high resolution, I feel it would be a better use of resources to create a maximum width of 640 for webpages, that way there is less code and you know that when you buy a device that it can either access a webpage or access a webpage that won't look right on my device.
With a maximum width, it would also be easier for QVGA devices to be able to shrink the webpage into the appropriate amount of pixels.
Webpages made for specialty devices would work the same way that they do now, if a company has a specialty site (as we do here) or simply the desire to make it availible on QVGA or less capable devices, then go right ahead. But we (web programmers) shouldn't be forced to make pages that work on any device.
All in all, I think that web designers should have the freedom to create a webpage based on 1 standard and not have to re-write the webpage based on who is looking at it. It should be the browsers job to interpret it correctly.

[Kati Compton]

With more and more devices coming out that have a high resolution, I feel it would be a better use of resources to create a maximum width of 640 for webpages
As someone that works on a 1280x1040 desktop, I'd be against limiting my browser width to 640 pixels...
But we (web programmers) shouldn't be forced to make pages that work on any device.
All in all, I think that web designers should have the freedom to create a webpage based on 1 standard and not have to re-write the webpage based on who is looking at it. It should be the browsers job to interpret it correctly.
I understand that, but on the other hand there's a few other issues. Making a normal domain also work properly for mobile devices probably also makes sure it works well for people with vision problems, etc. So to me, it isn't a bad thing to move towards design that's more "accessible" to everyone, whether it's on a mobile device, a voice reader, a 640x480 desktop, or a 2048x1536 desktop.

At some point, there will be such a wide variety of device screen sizes that counting the smaller screens as "mobile" and the larger ones as "not mobile" will not be sufficient. Through the intelligent use of CSS's and well-defined browser strings, I think that it would work much better.

Plus, web designers will probably be happier about having to spend a little more time to make the page a lot better - they can charge more for the service. :)

[Chucky]

I think if all the browsers supported system specific css (ie one for mobile, another for pocket/palm pc and finally one for desktop pcs) then that would cut down a lot of work. If a webpage is crafted out of simple semantic HTML elements and does not use tables for design, it is only a matter of minutes to create seperate stylesheets.

At least that way a webpage well be usable on any device, of course most people browsing a site with there mobile phone dont want to be forced to browse through as much content as someone on a desktop would, so it would still pay to make seperate pages. But I think as the resolution of mobile devices increase there will be less and less need for seperate versions. I personally think any device with a res of at least 640 would only need a seperate stylesheet, a seperate page would be a waste of time.

[bjornkeizers]

yap i tolly agree with them, a website should check who is acessing it and act acconrdingly, if its a mobile phone load the wml version, if its a pocket pc a light html version and if its a normal pc the full page.

Agreed. I don't want to have to bookmark two seperate pages - I would like the page to load according to my browser or ask me which version I'd like to load [PPC Thoughts Mobile is nice, but sometimes I want the big page on my Clie]

[heliod]

I totally agree as well. The problem with mobile devices will not be solved by having a new TLD, but by having Website owners respecting mobile devices and opening pages that support them.

Having to memorize or bookmark 2 addresses for each site damages any kind of universality that the WEB is supposed to have. Rather than that, webmasters ought to create sites that support both kinds of presentations, or that automatically redirect mobile devices to the appropriate pages in their sites.

[Bill Gunn]

Not every website is meant to be "universally" available. You may choose to only target mobile devices. The people here who advocate for Web "standards" would go berserk if MS insisted that all PPC's have identical features because they need to maintain a standard, or how about if they just insist that everyone "standardize" on IE. (Oh wait, they do)

Innovations always deviate from the "standard".

[JvanEkris]

i agree with the W3C-guys as well. I want ONE link to evey website. I want to copy/sync my pages to my PDA and use them without restrictions.

IMHO, the TLD's represent the goal of the organization (global, european, local etc.). This philosophy is completely ruined by this: global and local companies will have to fight for the same domain.

Let the browser-detect and good stylesheets do their work.....

Jaap

[Kati Compton]

Not every website is meant to be "universally" available. You may choose to only target mobile devices. The people here who advocate for Web "standards" would go berserk if MS insisted that all PPC's have identical features because they need to maintain a standard, or how about if they just insist that everyone "standardize" on IE. (Oh wait, they do)

Innovations always deviate from the "standard".
I think it's perfectly acceptable to have a mobile-targeted site - that would likely make it close to "accessible", and certainly still readable on a desktop. I also understand that a site dedicated to high-end video probably doesn't really need to cater to those with vision problems, unless a journalist needs to access information there for a story. And individuals with home pages are likely not going to have the knowledge or time to make several versions.

The main point here, though, is that the big companies (MS, HP, Sony, Dell, Coca-Cola, Pepsi, Nike, Nokia, Motorola, etc etc etc) should really be making mobile versions of their sites, and the people on this thread at least do not want to have to visit a different domain to get that.

[mhuie]

I think its a good idea.

Companies are not going to agree to a standard way of accessing a mobile version of their site... some go mobile.domain.com , others domain.com/mobile or pocket.domain.com, etc. etc...

I think its a good way of separating what is designed for what and make companies really put an effort into creating a mobile site.

[Kati Compton]

I think its a good idea.

Companies are not going to agree to a standard way of accessing a mobile version of their site... some go mobile.domain.com , others domain.com/mobile or pocket.domain.com, etc. etc...

I think its a good way of separating what is designed for what and make companies really put an effort into creating a mobile site.
I think the people on the "against" side of this are wanting, for example, www.apple.com to automatically detect mobile vs. non-mobile and format accordingly. The whole point is to *not* have to remember more than one URL (mobile.apple.com vs. apple.com vs. apple.mobi, etc).

[lapchinj]

The main point here, though, is that the big companies (MS, HP, Sony, Dell, Coca-Cola, Pepsi, Nike, Nokia, Motorola, etc etc etc) should really be making mobile versions of their sites, and the people on this thread at least do not want to have to visit a different domain to get that.
I couldn’t agree with this more but once MS wants to push forward some protocol or something money seems to be the big motivator (probably a lot of it). Besides building extensions into any protocol is bad practice and produces bad results. The protocol should be able to handle whatever challenges or issues happen to comes along.

A case in point is the Windows ‘C’ compilers of 10-15 years ago. Standard ‘C’ code using any vendor’s compilers on any OS platform is able to be compiled as long as there were no proprietary code involved (AKA extensions). The platform compiler libraries usually took care of platform specific issues. Enter the proprietary code (extensions) and this soon became none movable code only to be compiled by one vendor’s compiler. This soon became a mess and this same scenario will be played over again if companies start creating extensions to the TLD. We’ll end up having to get separate browsers and hardware to view different TLDs.

Jeff-

[Pony99CA]

The main point here, though, is that the big companies (MS, HP, Sony, Dell, Coca-Cola, Pepsi, Nike, Nokia, Motorola, etc etc etc) should really be making mobile versions of their sites, and the people on this thread at least do not want to have to visit a different domain to get that.
Exactly. If the company is going to create two versions of their site, why not just auto-detect the browser and go to the correct one?

The one problem with that is that they may feel that detecting browsers is too clumsy (and getting worse as more browsers come out), and they don't want to do that. Having a .mobi domain would mean they didn't have to do that work. Of course, there are two problems with that approach.

First, you can get almost the same effect today with subdomains -- mobile.domain.com. It's a little more typing, which makes visiting a site on a device without a good text input mechanism more difficult, but do they think the extra characters will deter people from visiting the site if they have a compelling reason to?

Second, and worse, what exactly is a "mobile" device. Mobile devices have many different screen sizes, color depths and input methods, which would still necessitate detecting the browser or device and serving the correct content. The .mobi TLD has just partitioned the problem a little, not solved it.

So how do we avoid the kludge of detecting browser strings and make detection of device information simpler? The solution is the Composite Capability/Preference Profiles (http://www.w3.org/2001/di) (CC/PP) model the W3C is working on. It provides a standard method for detecting device capabilities and user preferences. This is probably what was alluded to when they said "a variety of new methods just coming into use for the client to be able to tell the server what its capabilities are."

Of course, browsers will need to support CC/PP, which may take longer than implementing a .mobi TLD, but browser strings and CSS Media attributes can tide us over, I hope.

Steve

P.S. My Web uses the CSS Media attribute to provide modified content for handheld devices. Here's a fragment of the code:

&lt;LINK REL="STYLESHEET" TYPE="text/css" HREF="svppc-handheld.css" MEDIA="handheld">
I'm not exactly sure how the browser detects that the device is a handheld, but it seems to work. Visit my site from your PC and from a Windows Mobile 2003 Pocket PC (or a Pocket PC with a browser that supports CSS) and see the difference.

[Weyoun6]

If a web page is done properly to W3C spec - then there is no need for a .mobi domain. Its only that microsoft is lazy in its IE implementation and wants to add special "features" to its web pages. Innovation is one thing, but you can innovate within the standard, such as using plugins like flash, which dont "break" the page if they are not installed, unlike microsofts "innovations". Having a whole domain just for mobile web pages is expensive and totally unnessesary.

[Pony99CA]

If a web page is done properly to W3C spec - then there is no need for a .mobi domain. Its only that microsoft is lazy in its IE implementation and wants to add special "features" to its web pages. Innovation is one thing, but you can innovate within the standard, such as using plugins like flash, which dont "break" the page if they are not installed, unlike microsofts "innovations".
First, I'm not sure why you're blaming Microsoft here; Netscape did the same things back in the day. I believe plug-ins and JavaScript were non-standard at one time, and both were created by Netscape.

Second, what specific features are you referring to that break pages?

While conforming to standards is generally a good thing, sometimes innovation requires going beyond the standards. In fact, standards often evolve because of extensions created by other companies. (I speak from experience here -- I helped make a programming language conform to a standard and was involved in the standards program for extending that language back in the '80s.)

Steve

[Weyoun6]

Actually I'm just picking on MS because Netscape doesnt exist anymore - and Mozilla does follow standards

[Kati Compton]

The one problem with that is that they may feel that detecting browsers is too clumsy (and getting worse as more browsers come out), and they don't want to do that. Having a .mobi domain would mean they didn't have to do that work. Of course, there are two problems with that approach.
What *I'd* like to see is an additional string besides the browser string. A user-settable "browser type" string. Where that "type" could be "small", "medium", or "large" for display. Or "simple", "medium", or "complex", more accurately. That way, on the desktop, if you have a (gasp) 28.8 modem, you could choose the simple version that has low graphics even though you're on a desktop. Likewise, the standard could specify that the "simple" version should be constrained to be "accessible" for those with disabilities, etc. Then you have medium, which the company might redirect to "simple" if they only want 2 versions - this could be for the VGA handhelds or PDAs, or desktop users with only minor vision issues. Then "complex" could be for "normal" desktop browsing.

So how do we avoid the kludge of detecting browser strings and make detection of device information simpler? The solution is the Composite Capability/Preference Profiles (http://www.w3.org/2001/di) (CC/PP) model the W3C is working on. It provides a standard method for detecting device capabilities and user preferences. This is probably what was alluded to when they said "a variety of new methods just coming into use for the client to be able to tell the server what its capabilities are."
Yeah - with a quick glance, this looks like it could do it, but I didn't see (again, I looked only *very* briefly) the actual propsed solution - mainly a discussion of the core problem, and some general techniques.

One thing that will be important is to limit the # of different display "types" so that it really isn't too much of a pain, and they don't have to make 30 different versions.

[Pony99CA]

So how do we avoid the kludge of detecting browser strings and make detection of device information simpler? The solution is the Composite Capability/Preference Profiles (http://www.w3.org/2001/di) (CC/PP) model the W3C is working on. It provides a standard method for detecting device capabilities and user preferences. This is probably what was alluded to when they said "a variety of new methods just coming into use for the client to be able to tell the server what its capabilities are."
Yeah - with a quick glance, this looks like it could do it, but I didn't see (again, I looked only *very* briefly) the actual propsed solution - mainly a discussion of the core problem, and some general techniques.

One thing that will be important is to limit the # of different display "types" so that it really isn't too much of a pain, and they don't have to make 30 different versions.
I haven't really looked at the CC/PP stuff -- reading RFCs and standards proposals isn't my idea of fun. :-D However, I would hope they wouldn't discuss display "types" at all, but rather allow specifying the actual details of the device being used -- horizontal and vertical resolution, color depth, audio capabilities, input mechanisms (keyboard, keypad, mouse, stylus, voice, etc.), perhaps even bandwidth (imagine not having to specify whether you want the dial-up or broadband video at various sites).

The preferences part would be used to pass browser settings to the server and could also be used to override the device capabilities -- besides being useful for users with disabilities, it would be very useful for testing Web sites without actually having to use all of the devices.

Steve