Re: [ANN] SView: a Web-based system for browsing Linked Data

Thanks for your comment.

It touches some interesting topics such as Linked Open Data and the standardization of Linked Data browsers. 

>From a practical point of view,  SView should become robust to cope with different situations. 

The SView team has just updated a related component. Now, it can deal with the situation that the input URI denotes an RDF file. 

When such a situation occurs, the user can view the RDF sentences (grouped by subjects) in the fetched RDF file.  

Please check it, and any comment is welcome.

Regarding another issue you mentioned,  integrating with sameAs.org or some other sameAs services, we will consider the possibility. 

Actually, we have had tried to do so. But, we met with two big challenges: 

1. Some services consume two much time, e.g. several seconds. 

2. Different principles between services and the SView system, e.g. coreference relationship should be an equivalence relation or not. 

Anyway, we will continue to study this particular topic.  


Best regards, 


Yuzhong


--------------------------------------------------
 

> Thanks, very interesting.
> 
> On 14 Nov 2013, at 14:06, Yuzhong Qu <yzqu@nju.edu.cn> wrote:
> 
>> Dear Hugh,
>> 
>> The developer team investigated the reason why "RDF description not found” occurs to the three URIs you pointed out.
>> 
>> 1. For the first URI, the robots.txt of the website [1] does not allow agents to fetch the file.
>> 
>> User-agent: *
>> Sitemap: http://southampton.rkbexplorer.com/sitemap.xml
>> Disallow: /id/
> I think you must be the first Linked Data browser that respects robots.txt :-)
> Certainly all the other browsers access this page.
> I would expect many Linked Data sites to Disallow, so this may be a more general problem for you.
> I’m not sure that robots.txt is intended for something like this - a hit on the site prompted by a request yo get doesn’t really count as a “spider” or “wanderer”, which was the intention for robots.txt
> However, I can see a problem could happen.
> Since you don’t have a robots.txt Disallow, then a spider might start on you, and then you would be guilty of spidering on a site that didn’t want it.
> I note for example that http://graphite.ecs.soton.ac.uk/browser does access http://southampton.rkbexplorer.com/id/person-07113, but also does have its own robots.txt Disallow, which would prevent this - maybe you want to consider going down the same route?
>> 
>> 2. For the other two URIs, SView can dereferences the HTTP URIs. However, the captured RDF data do not provide any triples about the particular two URIs.
>> 
>> The SView system primarily aims to view entities instead of data files. So the input URI and the publication of related data should follow the rules of Linked Data [2], in particular the third rule:
>> When someone looks up a URI, provide useful information (about the URI), using the standards.
>> 
>> In this case, I think these two URIs denote RDF files instead of entities. Of course, there is a subtle issue, say HttpRange-14 [3].
>> 
>> In summary, the SView system keeps the principle of Linked Data in its design, and your feedback inspires us to cope with the situation that the input URI denotes an RDF file.
> Ah, interesting observation.
> So your "SView has not found RDF description about …” is telling me that it resolved the NIR and found a document (IR), but there was no statement about the NIR in the document, so it didn’t render it.
> That sounds pretty reasonable.
> By the way, you are right that the second (differentFrom) example is more cavalier with the principles, but I think not the first.
> http://www.sameas.org/?uri=http%3A%2F%2Fdata.ordnancesurvey.co.uk%2Fid%2F7000000000003822
> is an NIR that does Conneg to
> http://www.sameas.org/html?uri=http%3A%2F%2Fdata.ordnancesurvey.co.uk%2Fid%2F7000000000003822
> or
> http://www.sameas.org/n3?uri=http%3A%2F%2Fdata.ordnancesurvey.co.uk%2Fid%2F7000000000003822
> or whatever.
> And this doesn’t have statements about the NIR.
> (This is because we want to discourage people from considering the original NIR as a NIR for the things on the page, in case you were wondering.)
> 
> So you are in fact being quite right here.
> It is a good question whether you should allow for “bad Linked Data”, as you suggest you might.
> I suspect that you are right, but others may have a different view.
> Best regards
> Hugh
>> 
>> 
>> Regards,
>> 
>> 
>> Yuzhong Qu
>> 
>> 
>> [1] http://southampton.rkbexplorer.com/robots.txt
>> 
>> [2] Linked Data - Design Issues
>> http://www.w3.org/DesignIssues/LinkedData.html
>> 
>> [3] Dereferencing HTTP URIs
>> http://www.w3.org/2001/tag/doc/httpRange-14/2007-05-31/HttpRange-14
>> 
>> 
>> 
>> --------------------------------------------------
>> From: "Hugh Glaser" <hugh@glasers.org>
>>> Hi there,
>>> This looks good.
>>> But it doesn’t seem to be fetching quite a lot of Linked Data.
>>> I can get the ones on the home page, and I found
>>> http://data.ordnancesurvey.co.uk/id/7000000000003822
>>> worked.
>>> But
>>> http://southampton.rkbexplorer.com/id/person-07113
>>> http://www.sameas.org/?uri=http://data.ordnancesurvey.co.uk/id/7000000000003822
>>> http://differentfrom.org/symbols/http%3A%2F%2Fdata.ordnancesurvey.co.uk%2Fid%2F7000000000003822
>>> all give "RDF description not found”.
>>> As far as I can verify, these three all do Conneg for
>>> Accept:application/rdf+xml
>>> 
>>> Also, would you like to integrate with sameAs.org, or some other sameAs
>>> services?
>>> 
>>> Best
>>> Hugh
>>> 
>>> On 13 Nov 2013, at 05:50, Yuzhong Qu <yzqu@nju.edu.cn> wrote:
>>> 
>>>> Dear all,
>>>> 
>>>> A new system for browsing Linked Data, called SView, is now available
>>>> [1]. It enables end-users to conveniently browse entities described in
>>>> the Linked Data through views and navigators. Besides, it provides a
>>>> built-in service for users to resolve URI coreference collectively.
>>>> 
>>>> It has the following features:
>>>> 
>>>> 1. Home View
>>>> This particular view presents a personalized list of properties and their
>>>> values of the visited entity. Users can sort, remove and add properties
>>>> of the visited entity to form a better view of the concerned entity. Each
>>>> user's viewpoint on property sorting is collected and then turned into a
>>>> consensus sorting. The consensus sorting on involved properties is shared
>>>> among all users.
>>>> 
>>>> 2. Entity Identity Module
>>>> The "ID" component becomes available when browsing an entity (visible at
>>>> the left-upper corner). It enables users to interactively identify
>>>> coreference URIs of a visited entity. Each user's viewpoint on entity
>>>> coreference resolution is learned from the user's interaction, and all
>>>> users' viewpoints are aggregated and then shared among all users.
>>>> 
>>>> 3. Navigator Panel
>>>> This panel provides several navigators, such as navigation path, similar
>>>> entity and filtering facility, to help users locate or find semantically
>>>> related entities.
>>>> 
>>>> While a number of views and navigators have been built in the system,
>>>> more views and navigators are expected to be plugged in to make the
>>>> system better. It is under consideration that whether or not the SView
>>>> should become an open platform that receives “plug-ins” (e.g. views,
>>>> navigators) from the developer community.
>>>> 
>>>> The WebSoft group at Nanjing University (http://ws.nju.edu.cn) will
>>>> continue to improve the SView system. Any comments are welcome.
>>>> 
>>>> Have fun with SView :-)
>>>> 
>>>> Yuzhong Qu
>>>> 
>>>> [1] http://ws.nju.edu.cn/sview
>>>> 
>>>> 
>>>> 
>>> 
>>> -- 
>>> Hugh Glaser
>>>  20 Portchester Rise
>>>  Eastleigh
>>>  SO50 4QS
>>> Mobile: +44 75 9533 4155, Home: +44 23 8061 5652
>>> 
> 
> -- 
> Hugh Glaser
>   20 Portchester Rise
>   Eastleigh
>   SO50 4QS
> Mobile: +44 75 9533 4155, Home: +44 23 8061 5652
> 
>