New Publications ER

[Till Nagel]

Dear all,

here is a new version of the ER (and a small class diagram) for publication data, according to the requirements we discussed in our last VC.

The diagram only is for discussion. An extensible and flexible RDF schema may be more appropriate. Also, I did not fully design all the sub-classes for publications and conferences, as this can get rather extensive, and standards already exist.

Important issues from my side:

a) Affiliation with address and geo-location
- How to extend the SWRC affiliation property of a Person to also include address, and geo-location?
- Proposal: vCard in RDF: http://www.w3.org/2006/vcard/ns
- How to handle legacy data? According to the latest SWRC OWL an affiliation is an organization, which according to [Sure 2005] (http://www.aifb.kit.edu/web/Inproceedings1003) only has a single name property.

b) Ternary association between paper, author, affiliation
- Only with a ternary association all occurring relations can be mapped. Some analysis only is possible if these exists.
- Not all data sources provide these data.
- To be able to do both, the database schema should be seen as suggestion, only, and the BuRST as exchange format. Then every client can adopt the data according to his needs.

c) Unification / Duplicate handling
- Xavier and his team as well as I came up with unification heuristics. It would be great to offer these cleaned-up data to the community.
- Allow same-as relations (?)
- Question: Is it needed to provide both, the original and the unified data? Or can we just go with the cleaned data (with the potential false positives)?

I would appreciate any feedback, as well as your opinion on how we proceed using and extending BuRST. I would love to use it, also for further discussion, but at the moment some things are not very clear to me.

Best,
Till

[Xavier Ochoa]

Dear scitels,

We have been converting the data from ED-Media, E-Learn and ECTEL to the format of the database and the most difficult issue is the identification of the authors.

When you fuse together the papers and the citations hell breaks loose with the name of the authors. For the conference papers, thanks to the information present in the proceedings,  we know that Xian Lee from X Taiwanese University is the author of some papers in ED-Media and maybe some in ECTEL.  However, when citations are included, we have a ton of X. Lee named authors.  Close inspection of the papers indicated that X. Lee is in reality 14 different authors.  

For now we are managing the papers authors and citations authors in different tables to avoid problems.  We should come up with some magical algorithm to relate authors complete name and authors abbreviated name.  But appart from an intelligent agent scouting the web searching for the compete references of the citations, it is not a problem that will be solved soon.

By the way, we expect to release the data this week in the database format and linked data version.

Kind regards,

Xavier

--
Professor
Electric and Computing Engineering Faculty
Director of the Research Program on Teaching and Learning Technologies
---------------------------------------------------------
CTI - Information Technology Center
Escuela Superior Politécnica del Litoral
Campus "Gustavo Galindo"
Km. 30.5 Via Perimetral
Guayaquil - Ecuador
Tel.: (593)-(4)-2269773
Fax: (593)-(4)-2269776
web: http://ariadne.cti.espol.edu.ec/xavier
----------------------------------------------------------

--
Dipl.-Inform.(FH) Till Nagel
Interaction Design Lab
Fachhochschule Potsdam
http://interface.fh-potsdam.de


--
You received this message because you are subscribed to the Google Groups "SciTEL2.0" group.
To post to this group, send email to scit...@googlegroups.com.
To unsubscribe from this group, send email to scitel20+u...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/scitel20?hl=en.

[Bram Vandeputte]

Hi All,

Till proposed to organise a new flashmeeting next week to discuss further about this topic.
Please indicate your availability in this dooel :
http://doodle.com/cyte2fksq9q64fd7

greetings,
Bram

> Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm

>
> --
> Dipl.-Inform.(FH) Till Nagel
> Interaction Design Lab
> Fachhochschule Potsdam
> http://interface.fh-potsdam.de
>

> <scitel-er-4.png><scitel-uml-1.png>--

> You received this message because you are subscribed to the Google Groups "SciTEL2.0" group.
> To post to this group, send email to scit...@googlegroups.com.
> To unsubscribe from this group, send email to scitel20+u...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/scitel20?hl=en.
>

--
Bram Vandeputte

Katholieke Universiteit Leuven
Dept. Computer Science
Celestijnenlaan 200A, A03.010
B-3001 Leuven, Belgium

Phone: +32 16 327659
Mob: +32 474 667796
Fax: -

[Bram Vandeputte]

Hi,


The doodle has found us a nice spot where everybody can join :
Wed, 21 Apr 2010; 14:00:00 +0100
Please make sure you use the correct timezone !

link to meeting : http://fm.ea-tel.eu/fm/609c12-21210

See you then !


greetings,
Bram

[Till Nagel]

Dear all,

sorry for last week's late cancellation notice due to my sickness. Now I am fit and healthy again...

So, please again provide dates when you're up for a Flashmeeting re the SciTEL 2.0 infrastructure.

http://doodle.com/en6iztzhvsrvfrx4

(I just chose afternoon times, so Xavier could join)

Best,
Till

--
Dipl.-Inform.(FH) Till Nagel
Interaction Design Lab
Fachhochschule Potsdam
http://interface.fh-potsdam.de

[Till Nagel]

This Doodle replaces the previous one from Bram, as those times did not really work out.

> http://doodle.com/en6iztzhvsrvfrx4


Sorry for any confusion.

-Till

[Till Nagel]

Dear all,

so, let's meet next Tuesday at 4pm CET.

http://fm.ea-tel.eu/fm/2cf2fc-21357
Tue, 04 May 2010 15:00:00 +0100

Best,

Till


On 27.04.2010, at 22:15, Till Nagel wrote: