[ih] DNS History

[Richard Bennett]

Revised OSI Model: Layer 0 = Authentication; Layer 8 = Money.

On 3/8/2010 7:18 PM, John Day wrote:
> Yes but much this pays no attention to issues of security, access 
> control or scope.
>
> At 17:51 -0800 2010/03/08, Richard Bennett wrote:
>> And now there's this Semantic Web thing and the Bob Kahn Digital 
>> Object Identifier systems that aim to expose structure in web sites 
>> so that the content can be more easily indexed, searched, and 
>> grabbed. In the end, it's all about granularity and aggregating local 
>> indexes.
>>
>> On 3/8/2010 5:26 PM, Dave Crocker wrote:
>>>
>>>
>>>
>>>> Could you say the same thing about X.500?
>>>>
>>>>> Nope -- early attempt to do the web.
>>>>>
>>>>>> Wasn't all that Archie and Veronica stuff an attempt to provide the
>>>>>> Internet with a directory service?
>>>
>>>
>>> This exchange is confusing things a bit.
>>>
>>> The Web publishes documents and has evolved into something that is 
>>> probably best viewed as allowing interaction with documents.  (That 
>>> might be a Procrustean view, given the lofty views of web 2.0, etc., 
>>> but I'm trying to stay with basics.)
>>>
>>> Google, et all, scan the web and index it.  A search engine is not 
>>> 'the web', although it is a tool of the web.  The web is either the 
>>> documents or the full set of things that touch the documents.  But a 
>>> search engine is not 'the' web.
>>>
>>> Anonymous FTP published documents.  Lousy usability characteristics. 
>>> Gopher published documents. Reasonable usability, but limited 
>>> document style. They were the early sequence that led to the actual 
>>> Web.
>>>
>>> Archie indexed ftp.  Veronica indexed gopher. Early search engines. 
>>> These are services that are layered on top of the publication 
>>> service and the publication service is passive, in that there was no 
>>> organized registration of the documents, particularly, with respect 
>>> to the indexing (more recent active web page support of search 
>>> engines not withstanding.)
>>>
>>> X.500 was a user name registration scheme, originally designed to 
>>> lookup users, especially for email. It started with the premise 
>>> that, done in scale, a human name is not unique so that other 
>>> attributes would be needed to distinguish the target user.  Since if 
>>> flowed from X.400, the concept of a simple, global, unique email 
>>> address was already a lost cause.  (Your global address was relative 
>>> to your provider, which led to some interesting business cards, for 
>>> folks who had multiple providers.)
>>>
>>> In its earliest discussions, the function description was strikingly 
>>> similar to what we built for MCI Mail, so that
>>>
>>>    crocker, brandenburg, california
>>>
>>> might produce my address.  (My first participation in the X.500 
>>> discussions was shortly after we had MCI Mail running, so I was able 
>>> to confirm the utility of this basic model, though not the later 
>>> technical design for achieving it in scale.  MCI Mail was a closed 
>>> system.)
>>>
>>> But note that the data base that X.500 used was for actively 
>>> registered email users, not passively available (rather than listed) 
>>> documents.  This was meant to be more like a White Pages than a more 
>>> general searching service, even as constrained as a Yellow Pages.  
>>> (But yes, goals expanded.)
>>>
>>> Besides having a search function, X.500 differed from the goals of 
>>> the DNS by being finer-grained, targeting personal addresses, rather 
>>> than host addresses.
>>>
>>> The differences between document publishing, personnel registration, 
>>> name lookup and name (or, more generally, attribute) searching each 
>>> warrant distinction from the other.
>>>
>>> d/
>>
>> -- 
>> Richard Bennett
>> Research Fellow
>> Information Technology and Innovation Foundation
>> Washington, DC
>

-- 
Richard Bennett
Research Fellow
Information Technology and Innovation Foundation
Washington, DC