Re: Linked Data Book in Early Access Release

On 12/7/12 1:30 PM, Melvin Carvalho wrote:
>
>
> On 7 December 2012 19:24, Kingsley Idehen <kidehen@openlinksw.com 
> <mailto:kidehen@openlinksw.com>> wrote:
>
>     On 12/7/12 12:30 PM, David Wood wrote:
>
>         On Dec 6, 2012, at 17:33, Kingsley Idehen
>         <kidehen@openlinksw.com <mailto:kidehen@openlinksw.com>> wrote:
>
>             On 12/6/12 5:22 PM, Kingsley Idehen wrote:
>
>                 On 12/6/12 5:12 PM, David Wood wrote:
>
>                     This seems like good guidance for anyone wishing
>                     to provide LD the easiest possible way: Add a
>                     Turtle file, link to it and provide a link header.
>                      Of course, we need to live with a bogus HTTP
>                     Content-Type, but that's unfortunately quite
>                     common even for people who control their own server.
>
>                     Thanks again to Manning for the quick response!
>
>                 Exactly!
>
>                 Linked Data should be discoverable to a variety of
>                 user agent profiles:
>
>                 1. HTML oriented -- user <link/> based Web Linking
>                 pattern in <head/>
>                 2. HTTP oriented -- repeat via "Link:"
>                 3. RDF aware -- in the content include
>                 wdrs:desceribedby, foaf:topic etc. relations.
>
>
>             David,
>
>             In addition to the above, is it possible to have the mime
>             type fixed? At the current time, I still get:
>
>         Hi Kingsley,
>
>         Maybe.  Manning js using the Yahoo! Small Business Web Hosting
>         service.  I have asked some people at Yahoo! to repair the
>         Content-Types and I will revisit this if they don't.
>
>
>     How about this:
>
>     1. They add another <link/> relationship to <head/>
>     2. The link points to a Turtle doc in a data space you control --
>     absolute worst case (as I believe you already have your own
>     space), *you* just can get free data space from the likes of
>     Dropbox, SkyDrive, Amazon S3 etc
>     3. You set the correct content-type for the Turtle doc in your
>     data space
>     4. Ditto fixing any other issues e.g., entity name ambiguity etc..
>     5. Done :-)
>
>
> I've often thought there should be an open world, distributed, fault 
> correction service for sites that give the wrong content type.  Could 
> solve a lot of problems.

Yep!

Web-scale Linked Data virtuosity strikes again :-)

-- 

Regards,

Kingsley Idehen	
Founder & CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen