Hi Dimitris,

I've submitted a proposal (called DBpedia Viewer) at the Melange site. Could 
you please take a look and give some feedback and further directions? The idea 
about DBpedia Search is also included in the proposal (in short).

Regards,

Denis

From: [email protected]
To: [email protected]
Date: Sun, 28 Apr 2013 17:57:59 +0200
CC: [email protected]
Subject: Re: [Dbpedia-gsoc] DBpedia Better Display




Hi,

Thanks for the quick response.

About Spotlight annotations: I was also worried about the load on Spotlight. 
I'd like to note the annotations of course will work asynchronously (first page 
is loaded and an annotation request is made, in the following seconds the reply 
will be displayed). However, going directly to the DBpedia Spotlight REST 
interface for every DBpedia page query would mean an enormous amount of 
duplicate work because the rdfs:comment and rdfs:abstract rarely change even in 
DBpedia LIVE. This feature would need a large cache of rdfs:comment and 
rdfs:abstract annotations for every page. After such a cache is constructed, 
there would be no more load on DBpedia Spotlight for the "dead" versions of 
DBpedia and minimal load to update pages that are changed.
However, I'm not yet sure where to put this cache database.

About Wikidata insertions: how do you want to insert the data to Wikidata? So 
that users can add/edit all triples of some entity and push it to Wikidata? The 
interface part of editing triples is not so difficult. I have had an idea for 
quite some time now to let users edit HTML live, in-page, without transition to 
a special "WYSIWYG" editor or something. This idea would be simple to implement 
for editing triples on DBpedia. But after the edit/add of the page in the 
interface, the changed data need to be sent to Wikidata and processed. And this 
requires authentication. And if we want to push to Wikidata, we probably need 
Wikipedia logins. Plus there is probably the whole anti-vandalism system that 
would complicate things. Or am I getting something wrong here?

Greetings,

Denis

From: [email protected]
Date: Sun, 28 Apr 2013 18:25:46 +0300
Subject: Re: [Dbpedia-gsoc] DBpedia Better Display
To: [email protected]
CC: [email protected]

Hi Denis,


On Sun, Apr 28, 2013 at 5:28 PM, Denis Lukovnikov <[email protected]> 
wrote:





Hi,

So I'll take a shot.
One of the ideas on the ideas page that caught my eye was is the idea to design 
a better display for DBpedia.
I'm not exactly sure what kind of solution you want. You mentioned 
dbpedia-vad-i18n somewhere in earlier threads.


The solution I have in mind right now is to make a fat client that does a lot 
of SPARQL. 
We could try to reduce client load by using some PHP code that does SPARQL via 
SQL to the triplestore at page query (as mentioned at the end of 
http://www.openlinksw.com/dataspace/doc/dav/wiki/Main/VirtuosoPHP). But I would 
keep this PHP very simple. This PHP page could also serve as a query point for 
AJAX requests from the main page.


Alternatively, if (AJAX->)PHP->SQL->SPARQL doesn't prove to be better, 
everything could just be done in JavaScript->SPARQL.
Also, I would minimize or even ban any VSP as it is Virtuoso-specific.



We 'd also like to minimize or eliminate VSP code. The problem is that this 
plugin serves also the Linked Data interface (content negotiation) thus, 
removing it completely will generate more work.


A simple / minimal PHP page within the plugin that loads the stored data and 
JavaScript->SPARQL for loading external data seems like a very reasonable 
option.
  


For the styling part, I would follow the flat design style that's quite "in" 
right now (http://bits.blogs.nytimes.com/2013/04/23/the-flattening-of-design/). 
Bootstrap would be my choice of front-end framework. It allows for easy 
cross-device web development and supports more IE versions that its 
competitors. Also, I'm quite familiar with Bootstrap as I've worked with it for 
a job last summer. I also have quite some experience with jQuery (from summer 
jobs and pet projects), which can be used together with Bootstrap. jQuery 
allows for easy AJAXing (and thus SPARQLing over AJAX) and easy manipulation of 
HTML DOM.



As for the layout structure, here is a lot of freedom and I have several 
idea's. Something concise but flashy, interactive could be more attractive to 
"normal" users and something more factoid looking would be better for academics 
maybe. I could do both. Would you like me to include some mock-ups in the 
proposal?



We are very open to design suggestions here and bootstrap & jQuery seem fine. 
The two modes (academic / normal) also look like a good idea. 
Regarding the mockup, there is no need for one at this point but, if you 
already figured out the layout and it's easy for you to create one then you can 
include it in your application



 Also keep in mind that we want this interface to work for multiple DBpedia 
editions so, some parts (like domain name, default graph, ...) must be 
parametrizable.
  


Also, I just got a wild idea to make DBpedia more useful and promote DBpedia 
Spotlight. The rdfs:comment and rdfs:abstract properties (everything with plain 
text) could be annotated with DBpedia Spotlight! (with a mention it's annotated 
with DBpedia Spotlight somewhere next to it).



Maybe the Spotlight guys can make a comment on this. My only concern here is on 
the load of the Spotlight API.
  


Could this be the solution you are looking for? What is your advise on it and 
on the steps I further need to take?

Yes, this idea is quite straightforward and this is more or less what we want 
to achieve.


You could also give some extra thought on the last part of the idea about 
inserting information back to Wikipedia / Wikidata from the interface.

You can also start preparing your application. If you submit early we can help 
you improve it.



Cheers,
Dimitris
 Maybe something more about myself. I consider myself fluent in Java and 
Python. I've done my share of web development so I'm good with HTML(5), SQL, 
JavaScript, PHP, XSLT, CSS. Last summer, I've written a Drupal 7 module for 
documentation import and collaborated on a redesign of the website of my 
employer (using Bootstrap and moving to Drupal 7).



Kind regards,

Denis




> From: [email protected]
> Date: Sun, 28 Apr 2013 12:55:47 +0200
> Subject: Re: [Dbpedia-gsoc] DBpedia GSoC Application Feasibility


> To: [email protected]
> CC: [email protected]
> 


> Hi Denis,
> 
> see comments in line:
> 
> On Sun, Apr 28, 2013 at 1:43 AM, Denis Lukovnikov
> <[email protected]> wrote:


> > First, I'll be graduating this summer. For the next year, I'm hoping to land
> > a PhD position or to study further (probably management plus some more CS
> > stuff), but I'm not sure which one yet. Am I eligible for GSoC 2013 in the


> > case I choose not to stay in the academic world next year?
> 
> Somebody on the mentors list pointed out:
> 
> > From the FAQ:
> >
> > As long as you are accepted into or enrolled in a college or university


> > program as of May 27, 2013, you are eligible to participate in the program.
> > and separately:
> >
> > In order to participate in the program, you must be a student. Google
> > defines a student as an individual enrolled in or accepted into an


> > accredited institution including (but not necessarily limited to) colleges,
> > universities, masters programs, PhD programs and undergraduate programs
> 
> So it looks like you would be eligible.


> 
> 
> > Second, I found out DBpedia is in GSoC just a few days ago. After reading in
> > the mailing list, I got the impression it takes a lot of effort to gain
> > notability. [...] And there already are candidates for


> > the most of the idea's I found interesting.
> > [...]
> 
> What you say it true. An early engagement votes well for students, but
> the project proposal and how you work with us in creating it is still


> the biggest factor for acceptance.
> 
> 
> > Also, is it possible to make more than one DBpedia proposal or do we have to
> > chose one idea? (I'm new to GSoC...)
> 
> Yes, but the proposals should be thought out thoroughly and be


> improved after feedback from the mentors. This might be difficult for
> multiple proposals.
> 
> 
> Cheers,
> Max
                                          

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

Try New Relic Now & We'll Send You this Cool Shirt

New Relic is the only SaaS-based application performance monitoring service

that delivers powerful full stack analytics. Optimize and monitor your

browser, app, & servers with just a few lines of code. Try New Relic

and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_apr
_______________________________________________

Dbpedia-gsoc mailing list

[email protected]

https://lists.sourceforge.net/lists/listinfo/dbpedia-gsoc




-- 
Kontokostas Dimitris
                                          

------------------------------------------------------------------------------
Try New Relic Now & We'll Send You this Cool Shirt
New Relic is the only SaaS-based application performance monitoring service 
that delivers powerful full stack analytics. Optimize and monitor your
browser, app, & servers with just a few lines of code. Try New Relic
and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_apr
_______________________________________________
Dbpedia-gsoc mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dbpedia-gsoc                       
                  
------------------------------------------------------------------------------
Try New Relic Now & We'll Send You this Cool Shirt
New Relic is the only SaaS-based application performance monitoring service 
that delivers powerful full stack analytics. Optimize and monitor your
browser, app, & servers with just a few lines of code. Try New Relic
and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_apr
_______________________________________________
Dbpedia-gsoc mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dbpedia-gsoc

Reply via email to