Paul Houle wrote:
> Kingsley Idehen wrote:
>>
>> I would qualify the problem as being specific to the DBpedia Linked 
>> Data Space rather than Linked Data in general :-)
>>
>    I'd call the field "generic databases;"  Dbpedia and Freebase are 
> the clearest examples,  although there's a close relationship with 
> Cyc, YAGO and things of that sort.
>
>    Dbpedia has strengths and weaknesses.  It's got very wide 
> coverage,  so it's a reasonable "universe of discourse" for general 
> communication.  Of course,  breadth of coverage brings with it quality 
> problems.
>
>    Ultimately dbpedia is useful as a "list of things that people have 
> affective attachments to" -- these items,  for the most part,  are 
> linked to human readable descriptions,  so it's a pretty good start.  
> (All you need is "love.com")
>
>    The same problems are going to turn up in data that comes from Web 
> 2.0 sources:  for instance,  you could find videos on youtube that are 
> about many dbpedia topics,  but you're just as likely to find 
> something obscene.
>> Also, how does the geonames data space fair re. your analysis?
>>
>    Geonames has a good taxonomy for locations.  It has some 
> information about data quality.  It's also got about 10 times as many 
> points in it as does Dbpedia.  I still need something that represents 
> NYC as a polygon,  and I'm getting that by merging data from other 
> sources.
>> You are highlighting what could become anecdotal material re. why 
>> domain specific data spaces are important, in this case, one that's 
>> totally about data for reliable geo informatics etc..
>>
>    Well,  better ontologies help,  but it's hard to please everybody.
>
>    For instance,  I know a librarian who says that Dublin Core is a 
> big step back from what was in the MARC specification in 1969.  She's 
> right.  MARC was designed for the largest and most advanced libraries 
> in the world,  whereas Dublin Core is designed to be something that 
> anybody can understand.  Some people would be happy to have a 
> coordinate for the summit of Mount Everest and others would like to 
> draw a boundary between Everest and the mountains around it.  Other 
> people are concerned that the the concept of "Mountain" is not well 
> defined.
>
>    My overall vision in this area is to have interacting data spaces:  
> higher quality (or shall we say higher "resolution") spaces could 
> import data from lower quality spaces,  merge it,  clean it and 
> improve it,  maybe even push something back.
>
>    Behind all this is the concept of a persistent store,  something 
> which is often prohibited by "Web 2.0" API licenses (Flickr,  Amazon,  
> etc.)
>
Amen!!



-- 


Regards,

Kingsley Idehen       Weblog: http://www.openlinksw.com/blog/~kidehen
President & CEO 
OpenLink Software     Web: http://www.openlinksw.com





------------------------------------------------------------------------------
Enter the BlackBerry Developer Challenge  
This is your chance to win up to $100,000 in prizes! For a limited time, 
vendors submitting new applications to BlackBerry App World(TM) will have
the opportunity to enter the BlackBerry Developer Challenge. See full prize  
details at: http://p.sf.net/sfu/Challenge
_______________________________________________
Dbpedia-discussion mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dbpedia-discussion

Reply via email to