W3C home > Mailing lists > Public > semantic-web@w3.org > April 2008

Re: discussion about Semantic Web realization

From: Dan Brickley <danbri@danbri.org>
Date: Wed, 30 Apr 2008 18:29:53 +0100
Message-ID: <4818AC91.5090108@danbri.org>
To: Ioachim Drugus <sw@semanticsoft.net>
Cc: Maciej Gawinecki <mgawinecki@gmail.com>, semantic-web@w3.org

Ioachim Drugus wrote:
> 
> We are a group of developers in Moldova, who think that SW is the 
> project to *build the planet Earth  intelligent infosphere* - we believe 
> this is put it right terms.
> 
> Before I write my thoughts on this topic, I would share on the  status 
> of SW in Moldova, second to the smallest ex-Soviet republic, now an 
> *independent* country. The smallest (2mln) republic was Estonia - in 
> post-Soviet times they built the Skype. In Moldova (2.4mln) we hope to 
> build some good solutions for SW.  The capital Chisinau city was an 
> electronic city during USSR and we had a strong mathematical logics 
> school due to systematic research on superintuitionistic logics, which 
> started here. Currently Moldova is the "poorest county in Europe",  it 
> is not member of EU, it has great economic problems and lowest salaries. 
> But also in Moldova the expenses are the lowest in Europe, which invited 
> massive outsourcing it became a "small India in Europe". Unfortunately - 
> no projects on SW.
> 
> I am sure, Moldova can serve as a good development base for the 
> outsourced SW projects! This is also an invitation - I would be happy to 
> help to any company or organisation willing to spend less on development 
> of SW by outsourcing it to my country.

Thanks for the intro to Moldova. I'll admit that I didn't know much 
about the country, but am happy to learn.

And I'm inclined to agree re the prospects for SW-related outsourcing, 
too...

> At SemanticSoft, Inc. located in Moldova, we finance our SW research and 
> development from money we earn on custom software and our salaries are 
> lower than salaries of foreign companies in Moldova. Despite this, many 
> professionals prefer to work with us and develop SW because they believe 
> in SW great future. We cooperate with Moldova State University, 
> Technological University and Economics Academy. They already started 
> work on adding semantic technologies to their post-graduates curriculum. 

Good to know! Are any of these materials online? Even if not in English, 
it is good to be able to take a look around such things.

> Even though slowly, the interest to SW is steadily growing in Moldova. I 
> don't know where and how to start looking for research on SW to be 
> outsourced to our universities, and will be grateful if somebody 
> suggests or helps.

Another suggestion here (and a general one, to all 
listmembers/countries) is that arranging for inter-site visits may be a 
good way to build up skills and connections that bind the Semantic Web 
community together better. A related approach is through events like 
workshops and conferences. I think in general outsourcing is easier when 
there has been some prior face-to-face contact, or 
collaboration/discussion through standards work (such as this list), 
opensource toolkits etc.

> Now, I will share two thoughts on Maciej's message, which precisely 
> articulates the problems of growth of SW.
> 
> 1. Search engines are not alternative, but complementary to SW. Really, 
> SW is about *integration* on the global scale and *building* the logical 
> layer of the web, and SE is about *search* in the *presentation layer*.  
> SE and SW live in different spaces - they cannot compete.
> 
> 2. I will try to articulate the problem which stay in SW progress 
> towards industry and on  how to overcome it.
> 
> SW works only with *descriptions* - SW tools process only descriptions. 
> On the other side, we have the classic web, which works with *resources* 

(I'll read that as "document-like resources"; "resource" in the strict 
SW sense simply means "thing", while in other groups it meant something 
a bit like "page")

> - all classic tools work with *resources* with poor meta-data 
> management. Their is a gap between the two. In order to fill this gap, 
> we need to build tools of the *resource description* class.  A *resource 
> descriptor* (our term) is a tool which links the two sides and can make 
> classic web and semantic web work together. When such a tool is built, 
> the SW will become a great business. Our SemanticServer is built along 
> these lines
>
> S-Server does two main things:
> - Management of resources in repositories built on the Java Content 
> Repository standard, JSR-170  (classic web)
> - Description of content by ontologies (semantic web)
> 
> S-Server has incepient  *resource description* features. We started 
> another project to build a separate *resource descriptor* product which 
> could work with any other tools.

Sounds like nice work. Do you see any scope for a common 
architecture/design with the recently announced Drupal RDF effort? They 
again are building RDF/SW descriptions into a system based around a 
resource/document repository, eg see http://groups.drupal.org/node/9010 
and http://boston2008.drupalcon.org/session/rdf-and-semantic-web-drupal

There are also some commonalities with OpenLink and Talis's work. Which 
is all good: not meant as a criticism. I think we're all figuring out 
what shape RDF/SW application software will take; at the moment the 
closests we have to interchangable software pieces for RDF are parsers, 
databases and SPARQL engines. I'm very interested to see whether higher 
level functionality also gets created to shared interfaces, so 
implementations could compete on their merits, and users have improved 
app portability between them.

BTW looking at http://www.semanticsoft.net/semanticwebtools.html ... I 
read "Currently, Semantic Server comes with its own SPARQL processor 
which is also visual. The user can draw a diagram of the query - the 
tool builds the query and displays it result." ... is this your own 
SPARQL implementation, or do you build on top of an existing toolkit?

cheers,

Dan
Received on Wednesday, 30 April 2008 17:30:34 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 07:42:04 UTC