Hi,
I'll assume the HTML+JSON version is better, for now, and work further from
there. Also, I've written some blog posts today, you can check them out at
dbpv.wordpress.com.
I have been thinking about a two-step loading process. The idea is to
parallellize page loading as follows:
first load the properties of the entity we are viewing (outgoing links) and
display themthen load the triples where this entity is the value of a property
of some other entity (incoming links)even for large entities (e.g.
dbpedia:Sweden), there are relatively few outgoing links (properties this
entity "owns"). These properties should be shown first (like now).the idea is
to fire two queries at the SPARQL endpoint simultaneously, the first one to get
the outgoing links and the second one to get the incoming linksassuming there
are much less outgoing than incoming links for large entities, the first query
will return first and we can already display that. Also, the outgoing links are
probably the most relevant oneswhen the answer to the second query arrives, we
add that after the outgoing links
However, I'm not sure if this is a good idea. Here, we have two queries for one
entity display and I'm not sure about the overhead this generates. What do you
think of this?
Greetings,
Denis
From: lukovni...@outlook.com
To: hellm...@informatik.uni-leipzig.de; jimk...@gmail.com;
cstad...@informatik.uni-leipzig.de; kontokos...@informatik.uni-leipzig.de
Date: Thu, 11 Jul 2013 02:11:51 +0200
CC: dbpedia-developers@lists.sourceforge.net
Subject: [Dbpedia-developers] [GSoC2013] DBpedia Viewer: tradeoff server
load/traffic VS. client load
Hi,
So we'll continue this discussion on the developers list.
A little recap. So far, I've been making mockups, setting up my DBpedia mirror,
researching and most recently playing around with AngularJS and how to get it
to work with the DBpedia SPARQL interface. What is accomplished so far (on a
toy Angular setup, not the DBpedia mirror):
fetching entities from the SPARQL endpoint using a descriptive SELECT query and
using HTTP POST (to prevent any errors due to the 1900 char limit of URLs in
Virtuoso) is donejust to test, fetching DBpedia Lookup results from its API
also workshashless routing enabledno-JavaScript support tests:the first test
was to test overlap of URL addresses (between static files on the webserver and
Angular routes)
the second test on my toy setup was to get the Angular implementation to look
for a JSON hidden in the loaded HTML page and use that to show triples. If
there is no such JSON on the page, DBpedia SPARQL endpoint is queried. Also,
when we want to view another entity from this initially loaded HTML page (with
preloaded JSON description), the DBpedia SPARQL endpoint is queried (at least,
if you have JavaScript enabled) to get us to the next entity (instead of a full
page load from the server). Without JS, all queries just go to the DBpedia web
server.
Conclusion:
AngularJS seems to work properly for our purposes
now I have a concrete idea how to support no-JS with AngularJS
Now, the question with respect to the noscript version is: is it better to put
the JSON in HTML at the server or is it better to use RDFa scraping code at
client side?
RDFa scraping at client will increase the load on their browsers, and you're
reverse-engineering the source in a sense.
adding JSON at the server will add load on the server, but probably less (per
request) than RDFa scraping at the client (here you're
"forward"-engineering).it will also make the pages sent to the clients much
bigger (but I guess that's not really an issue)
My main worry is that it may take quite a while at the client to scrape a 10MB
page (a big entity, e.g. <dbpedia:Sweden>) in JavaScript before showing the
pretty version from the scraped data. Whereas including some 3 MB of JSON at
the server would have just little computational overhead when building the page
(although will increase traffic). But since I don't have as much insight in the
VSP code as the one who wrote it (Dimitris?), I'm not completely sure about
that last fact ("little computational overhead"). Is it correct?
Which one do you think is best (HTML+JSON or HTML->RDFa-scraper->JSON)?
Greetings,
Denis
------------------------------------------------------------------------------
See everything from the browser to the database with AppDynamics
Get end-to-end visibility with application monitoring from AppDynamics
Isolate bottlenecks and diagnose root cause in seconds.
Start your free trial of AppDynamics Pro today!
http://pubads.g.doubleclick.net/gampad/clk?id=48808831&iu=/4140/ostg.clktrk
_______________________________________________
Dbpedia-developers mailing list
Dbpedia-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/dbpedia-developers
------------------------------------------------------------------------------
See everything from the browser to the database with AppDynamics
Get end-to-end visibility with application monitoring from AppDynamics
Isolate bottlenecks and diagnose root cause in seconds.
Start your free trial of AppDynamics Pro today!
http://pubads.g.doubleclick.net/gampad/clk?id=48808831&iu=/4140/ostg.clktrk
_______________________________________________
Dbpedia-developers mailing list
Dbpedia-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/dbpedia-developers
