Re: The Test of Independent Invention (was: What Does Point Number 3 of TimBL's Linked Data Mean?)

I hope you realize that the point of that thought experiment is to 
ensure that the technology in question is sufficiently powerful and 
flexible, so that *if* a parallel technology were discovered, the two 
could be extended to encompass each other with minimal added cost -- 
*not* that it is in any way desirable to have such parallel technologies.

David

On 06/22/2013 08:55 AM, Melvin Carvalho wrote:
> It took me quite a while to understand this fully.  IMHO, it is really
> worth digesting.  I think it also sheds light on how to approach some of
> the topics raised in the last week.


>
> [[
>
> *The Test of Independent Invention*
>
> There's a test I use for technology which the Consortium is thinking of
> adopting, and I'll call it the Independent Invention test. Just suppose
> that someone had invented exactly the same system somewhere else, but
> made all the arbitrary decisions differently. Suppose after many years
> of development and adoption, the two systems came together. Would they
> work together?
>
> Take the Web. I tried to make it pass the test. Suppose someone had (and
> it was quite likely) invented a World Wide Web system somewhere else
> with the same principles. Suppose they called it the Multi Media Mesh
> (tm) and based it on Media Resource Identifiers(tm), the MultiMedia
> Transport Protocol(tm), and a Multi Media Markup Language(tm). After a
> few years, the Web and the Mesh meet. What is the damage?
>
> A huge battle, involving the abandonment of projects, conversion or loss
> of data?
> Division of the world by a border commission into two separate communities?
> Smooth integration with only incremental effort?
>
> Obviously we are looking for the latter option. Fortunately, we could
> immediately extend URIs to include "mmtp://" and extend MRIs to include
> "http;\\". We could make gateways, and on the better browsers
> immediately configure them to go through a gateway when finding a URI of
> the new type. *
>
> The URI space is universal: it covers all addresses of all accessible
> objects. But it does not have to be the only universal space. Universal,
> but not unique.*
>
> -- Tim Berners-Lee
>
> ]]

Received on Monday, 24 June 2013 03:05:34 UTC