I have just finished reading the State of New York et al.,[Plaintiff] v. Microsoft Corporation [Defendant]. Much of the discussion centers on what the court calls "the application barrier to entry". This describes a situation where Microsoft, through its monopoly of over 90% market share of operating systems, prevents applications from third parties from gaining distribution.
The browser is now part of the operating system. This is no longer a controversial viewpoint - although whether it should be still remains controversial. As it is part of the OS, IE has the same power - to prevent distribution of third party applications - as the OS. However there is a difference. The browser speaks to a service - as middleware. If middleware is not available to the user through the browser then, due to the browser's dominance, it simply does not exist as far as the user is concerned. Over 500 million desktops simply cannot access it.
Microsoft therefore has the power of life and death over middleware. The decision not to renew the RealNames relationship literally kills the middleware from the point of view of the user. This is the impact of monopoly produced by a 90%+ browser market share.
The "application barrier to entry" is repeated, but now it is the browser and internet middleware that is being squeezed.
The following diagram shows the impact of Microsoft excluding all middleware but Microsoft's own MSN Search from the browser.
http://www.teare.com/afterrealnames.gif
In the post-RealNames world all access to content from the browser address bar will be mediated by an interstitial page - delivered by MSN Search. It will be impossible, in any language, to type in a name and go directly to a place. Microsoft will - of course - monetize the interstitial page. Is this a good user experience? Will brand managers appreciate all navigation to their brand being interrupted by a visit to MSN Search? Is this a good outcome? Is it a violation of anti-trust regulations concerning the "application barrier to entry"? I think we will have answers to these questions in time.
Chris you are missing the point. Any reading of my WebLog at teare.com must lead you to the conclusion that the inability of DNS to support multi-lingual characters requires fixing, and that right now ONLY RealNames fixes this natively in the browser that is on 90% + desktops.
Microsoft are now about to hard code the browser to Microsoft's OWN middleware - the MSN Search Engine. If you type "IBM Thinkpad" into the browser you will get an MSN Search result.
Even if you do not like RealNames (its a free world) you have to acknowledge that ending up on the ThinkPad page at ibm.com is the right outcome.
How you can support Microsoft tying the browser to exclusively Microsoft controlled middleware - and by so doing disable every language except English (7 bit ASCII actually) is baffling to me.
Incidentally the business model you describe was abandoned many years ago. Keywords were $50 per year flat fee or $500 if it was a top brand with high traffic.
Keith Teare
Former CEO
RealNames Corporation
The limits of the Domain Name System are connected to its roots as a naming architecture invented in 1984, with the primary purpose of giving names to networks and to people as endpoints in email services. The network has changed dramatically since then. DNS's main weaknesses include: a) DNS is only able to make use of 7 bit ASCII - 26 characters in the English alphabet and the 10 in the numerical system, 0-9, plus the hyphen (37 total characters), in forming a name. 7 bit ASCII cannot handle foreign characters, creating a significant problem for languages with non-Roman scripts. b) DNS cannot guarantee quality of service in delivering content. A DNS resolution points a user to a physical resource and is at the mercy of bandwidth constraints and traffic peaks. c) DNS is a poor global naming system. A company with multiple sites worldwide has to give each of them different names [ibm.com; ibm.co.uk etc]. d) DNS has no inbuilt reporting capabilities. In fact, reporting on DNS traffic is so complex and essential that an industry has arisen to provide the imperfect reports that are available today.
URIs and URLs have weaknesses as well:
a) DNS gave birth to the URI. These long strings - again restricted to ASCII - allow naming of a wider set of resources. The URI can address individual web pages (with URLs), but the URI can also address people's email address - as in mailto:person@company.com - and even their phone number - as in phoneto:16504865555. b) The URI is a major breakthrough as a means of addressing an unlimited number and type of resources on the Internet, but it is not a naming system. Rather it is a physical addressing system. Naming systems match a physical resource with an alias. A phone number, for example, is simply a memorable (one hopes!) alias to a physical switch address. A DNS name is an alias to an IP number. Physical addresses that are also forced to play the role of names are a bad idea because an identity is then tied to a physical resource identifier. If the resource moves or changes, the name will break. No persistent naming system for the Web was built, and the URL was adopted as the only available alternative. This is widely accepted to be a huge error. c) In addition, the URI is incapable of being human friendly. Home page URLs for well known things barely pass muster as human friendly, intuitive identifiers. http://www.coke.com is OK, but how could one expect to intuitively understand that the URL for the US Fish and Wildlife Service is http://www.fws.gov. d) URLs cannot be consistent pointers to all content across all network access devices Wireless URLs and Web URLs point to different versions of content. e) The URL, like DNS, cannot use non-ASCII characters, although it can use a wider set of ASCII characters than the DNS. Limitations in DNS and URIs spawned search engines - which compensate for the lack of a manageable, human friendly naming architecture for network resources.
While they solve a specific and relevant problem, search engines also have weaknesses: a) Search Engines can only index "static" web pages on the public network. These are pages with a physical existence on a web server. Today less than 25% of web pages are "static". Search engines cannot provide pointers to protected content. Similarly, search engines cannot provide access to dynamic content that is refreshed frequently, or content that resides in a content management server or searchable database. b) Search Engines employ a "full text index" approach to content. Even with algorithms which attempt to elevate one site above another based on relevance rankings, search engines inevitably find it hard to distinguish between a home page for an entity [a company, a product, a famous person] and a reference to that entity by a third party. Search is great for research but of limited value for navigation.
Domain names can only name hosts. Keywords can name resources - as granular as a single result from a database query, a single page, even a sentance in a page. Naming is for resources. When the dns was invented hosts were the most granular resource, later people [email addresses]. But now its way more granular than that. Search Engines even don't cut it because many resources cannot be indexed as they are not live/static.
We are still out there. Try "lookup ?" in your browser where ? is anything. Or "Lookup cc ?" where cc is an iso country code. This will go away on 6/30.
Netscape will not be affected. AOL runs its own Keyword thing at keywords.netscape.com. Its a small database of about 40k names but it works for many popular sites. Doesn't do Chinese and Japanese theough. RealNames main strength - appart from aliasing a physical address [URL]- which is always a good thing, was to do it in all languages. The URL is only ASCII.
Slashdot Mirror
I have just finished reading the State of New York et al.,[Plaintiff] v. Microsoft Corporation [Defendant]. Much of the discussion centers on what the court calls "the application barrier to entry". This describes a situation where Microsoft, through its monopoly of over 90% market share of operating systems, prevents applications from third parties from gaining distribution.
The browser is now part of the operating system. This is no longer a controversial viewpoint - although whether it should be still remains controversial. As it is part of the OS, IE has the same power - to prevent distribution of third party applications - as the OS. However there is a difference. The browser speaks to a service - as middleware. If middleware is not available to the user through the browser then, due to the browser's dominance, it simply does not exist as far as the user is concerned. Over 500 million desktops simply cannot access it.
Microsoft therefore has the power of life and death over middleware. The decision not to renew the RealNames relationship literally kills the middleware from the point of view of the user. This is the impact of monopoly produced by a 90%+ browser market share.
The "application barrier to entry" is repeated, but now it is the browser and internet middleware that is being squeezed.
The following diagram shows the impact of Microsoft excluding all middleware but Microsoft's own MSN Search from the browser.
http://www.teare.com/afterrealnames.gif
In the post-RealNames world all access to content from the browser address bar will be mediated by an interstitial page - delivered by MSN Search. It will be impossible, in any language, to type in a name and go directly to a place. Microsoft will - of course - monetize the interstitial page. Is this a good user experience? Will brand managers appreciate all navigation to their brand being interrupted by a visit to MSN Search? Is this a good outcome? Is it a violation of anti-trust regulations concerning the "application barrier to entry"? I think we will have answers to these questions in time.
RealNames Keywords were typed into the browser 187 million times in March 2002. So people do use it. Lots of people actually.
Keith Teare
Former CEO
RealNames
Chris you are missing the point. Any reading of my WebLog at teare.com must lead you to the conclusion that the inability of DNS to support multi-lingual characters requires fixing, and that right now ONLY RealNames fixes this natively in the browser that is on 90% + desktops. Microsoft are now about to hard code the browser to Microsoft's OWN middleware - the MSN Search Engine. If you type "IBM Thinkpad" into the browser you will get an MSN Search result. Even if you do not like RealNames (its a free world) you have to acknowledge that ending up on the ThinkPad page at ibm.com is the right outcome. How you can support Microsoft tying the browser to exclusively Microsoft controlled middleware - and by so doing disable every language except English (7 bit ASCII actually) is baffling to me. Incidentally the business model you describe was abandoned many years ago. Keywords were $50 per year flat fee or $500 if it was a top brand with high traffic. Keith Teare Former CEO RealNames Corporation
The limits of the Domain Name System are connected to its roots as a naming architecture invented in 1984, with the primary purpose of giving names to networks and to people as endpoints in email services. The network has changed dramatically since then. DNS's main weaknesses include:
a) DNS is only able to make use of 7 bit ASCII - 26 characters in the English alphabet and the 10 in the numerical system, 0-9, plus the hyphen (37 total characters), in forming a name. 7 bit ASCII cannot handle foreign characters, creating a significant problem for languages with non-Roman scripts.
b) DNS cannot guarantee quality of service in delivering content. A DNS resolution points a user to a physical resource and is at the mercy of bandwidth constraints and traffic peaks.
c) DNS is a poor global naming system. A company with multiple sites worldwide has to give each of them different names [ibm.com; ibm.co.uk etc].
d) DNS has no inbuilt reporting capabilities. In fact, reporting on DNS traffic is so complex and essential that an industry has arisen to provide the imperfect reports that are available today.
URIs and URLs have weaknesses as well:
a) DNS gave birth to the URI. These long strings - again restricted to ASCII - allow naming of a wider set of resources. The URI can address individual web pages (with URLs), but the URI can also address people's email address - as in mailto:person@company.com - and even their phone number - as in phoneto:16504865555.
b) The URI is a major breakthrough as a means of addressing an unlimited number and type of resources on the Internet, but it is not a naming system. Rather it is a physical addressing system. Naming systems match a physical resource with an alias. A phone number, for example, is simply a memorable (one hopes!) alias to a physical switch address. A DNS name is an alias to an IP number. Physical addresses that are also forced to play the role of names are a bad idea because an identity is then tied to a physical resource identifier. If the resource moves or changes, the name will break. No persistent naming system for the Web was built, and the URL was adopted as the only available alternative. This is widely accepted to be a huge error.
c) In addition, the URI is incapable of being human friendly. Home page URLs for well known things barely pass muster as human friendly, intuitive identifiers. http://www.coke.com is OK, but how could one expect to intuitively understand that the URL for the US Fish and Wildlife Service is http://www.fws.gov.
d) URLs cannot be consistent pointers to all content across all network access devices Wireless URLs and Web URLs point to different versions of content.
e) The URL, like DNS, cannot use non-ASCII characters, although it can use a wider set of ASCII characters than the DNS. Limitations in DNS and URIs spawned search engines - which compensate for the lack of a manageable, human friendly naming architecture for network resources.
While they solve a specific and relevant problem, search engines also have weaknesses:
a) Search Engines can only index "static" web pages on the public network. These are pages with a physical existence on a web server. Today less than 25% of web pages are "static". Search engines cannot provide pointers to protected content. Similarly, search engines cannot provide access to dynamic content that is refreshed frequently, or content that resides in a content management server or searchable database.
b) Search Engines employ a "full text index" approach to content. Even with algorithms which attempt to elevate one site above another based on relevance rankings, search engines inevitably find it hard to distinguish between a home page for an entity [a company, a product, a famous person] and a reference to that entity by a third party. Search is great for research but of limited value for navigation.
Here are the open source API's.
http://www.realnames.com/body/ar_techcenter.asp
or use Keyword "RealNames Technical Documents".
Domain names can only name hosts. Keywords can name resources - as granular as a single result from a database query, a single page, even a sentance in a page. Naming is for resources. When the dns was invented hosts were the most granular resource, later people [email addresses]. But now its way more granular than that. Search Engines even don't cut it because many resources cannot be indexed as they are not live/static.
We are still out there. Try "lookup ?" in your browser where ? is anything. Or "Lookup cc ?" where cc is an iso country code. This will go away on 6/30.
Netscape will not be affected. AOL runs its own Keyword thing at keywords.netscape.com. Its a small database of about 40k names but it works for many popular sites. Doesn't do Chinese and Japanese theough. RealNames main strength - appart from aliasing a physical address [URL]- which is always a good thing, was to do it in all languages. The URL is only ASCII.