SD at ICMS 2016

[Hans-Gert Gräbe]

Dear SD followers,

let me report on several discussions we had at ICMS 2016 about the SD
project.

-------------------------------------------------
(1) Our SD People Database contains (today) 1119 entries and we decided
to make a new Author Disambiguation run with ZBMath (the last one dates
from Aug 2014). Author Disambiguation is a point in several contexts
and I moved and renamed the corresponding RDF graph to

http://symbolicdata.org/Data/AuthorIdentification/

Math Reviews also have a very clear Author Disambiguation scheme (using
numerical author IDs) and it would be nice to have disambiguation (i.e.,
alignment between our URIs and their AuthorIDs) also in that direction.
Unfortunately they have a much more restricted access policy for non AMS
members (as me). Nevertheless I already looked up manually and added
sd:hasMRAuthorID to five entries, e.g.
<http://symbolicdata.org/Data/Person/Robbiano_L>.

With sigsam.bib <ftp://tug.ctan.org/pub/tex/bib/sigsam.bib> there is a
third source that could be evaluated for author disambiguation. A first
try to translate that to RDF looks quite optimistic. Before I spend more
effort on that I would like to learn more about that source. In
particular in the intro there is a reference to the "ACM Portal
database" that probably supersedes this information, so it may be of
less interest to transform sigsam.bib, but to align Author IDs also with
http://dl.acm.org (although I have no idea how this could be done).

-------------------------------------------------
(2) With Albert Heinle I discussed the question about providing research
input (and output) data and how to reflect that within SD. Albert has a
web page with links to own such data and a first attempt to model that
is an RDF graph <http://symbolicdata.org/Data/BenchmarkReferences/> with
entries of RDF type sd:BenchmarkReference with sentences of the structure

sd:Person sd:providesDataAt <URL> .

similar to the Author Disambiguation idea. For the moment there is only
one such entry. As a first proof of concept look at
<http://symbolicdata.org/info/people.php?name=Albert>

-------------------------------------------------
(3) With Albert Heinle and Viktor Levadovskyy I discussed also the
situation of the G-Algebra collection within SD. We have only 8
instances within <http://symbolicdata.org/XMLResources/GAlgebras/>, no
more added for years and a quite cumbersome syntax. We decided to
reorganize that part - the G-Algebra community organizes its own store
of examples, either at a local site or within the repo at
data/OtherResources/ and deciding about the syntax of those files as the
Normaliz community did (they provide Data in Normaliz input format with
references to both local resources at
<http://www.home.uni-osnabrueck.de/wbruns/normaliz> and SD Resources at
<http://symbolicdata.org/OtherResources/TestSets>

>
> <http://symbolicdata.org/Data/TestSet/pluralityCondorcetEff> sd:hasNormalizPrimaryBase <http://www.home.uni-osnabrueck.de/wbruns/normaliz/InterChallExamples/CombinatorialVoting.zip> ;
> sd:hasNormalizSDBase <http://symbolicdata.org/OtherResources/TestSets/pluralityCondorcetEff.zip> ;
> sd:hasOrigin "Normaliz Collection, Tim Roemer 2015-03. From Example 10 of the Normaliz collection" ;

In SD we then store only fingerprints (also to be redefined?) of such
examples.

-------------------------------------------------
(4) With Matthew England I discussed several points of a close
cooperation with SIGSAM. We identified the following points of common
interest:

(4a) Upcoming and past conferences. SIGSAM maintains a
<http://sigsam.org/Resources/ConferenceSeries.html> a large list of past
conferences organized in conference series. I scraped that information
a little bit into the RDF graph
<http://symbolicdata.org/Data/ConferenceSeries/> to define URIs for each
of the series that can be referenced in conference records. For each of
the Conference Series entries I added a reference

> rdfs:seeAlso <http://sigsam.org/Resources/ConferenceSeries.html#nav...>

hence it is easy to get full information about a special conference from
the original SIGSAM web site.

For the single conferences it is harder to extract full information as
stored within our PastConferences.ttl (we store also the date of the
conference) and this can be processed from time to time for the
different conference series.

(4b) Matthew showed me that - similar to the CA Rundbrief of the German
Fachgruppe - CCA regularly publishes "Abstracts of Recent Doctoral
Dissertations in Computer Algebra" and it could be worth to join forces
to compile a common information stack. For the German Fachgruppe such an
RDF information source is provided at their CASN node at
<http://www.fachgruppe-computeralgebra.de/rdf/Dissertationen.rdf> and
used to serve the web page
<http://www.fachgruppe-computeralgebra.de/dissertationen/>. It is
extracted from the abstracts published in the CA Rundbrief.
Unfortunately that information was no more updated since early 2015.

I reactivated it as part of the CASN project and updated it with
additional information from CA Rundbrief back to nr. 45 in
<http://symbolicdata.org/Data/Dissertations/>. A typical entry looks as
follows:

> <http://symbolicdata.org/Data/Dissertation/Margolis_15> dcterms:abstract "Published in CA-Rundbrief 57 (2015)" ;
> dcterms:creator sdp:Margolis_L ;
> dcterms:date "2015"^^dcterms:W3CDTF ;
> dcterms:title "Torsionseinheiten in ganzzahligen Gruppenringen nicht auflösbarer Gruppen" ;
> bibo:degree bibo_degrees:phd ;
> bibo:institution "Uni Stuttgart, Germany" ;
> sd:hasReviewer sdp:Geck_M, sdp:delRioMateos_A ;
> sd:hasSupervisor sdp:Kimmerle_W ;
> sd:hasURL <http://dx.doi.org/10.18419/opus-5172> ;
> a bibo:Thesis .

I added 'dcterms:abstract' to refer to abstract publications and
'sd:hasURL' for an URL to the work itself. During my update I learned
that many of the old references published in the printed versions now
redirect to special information services that provide either a doi or an
urn as permanent address. Look at
<http://symbolicdata.org/info/dissertation.php> for a first proof of
concept. To integrate Information from the SIGSAM dissertation project
some structured information is required.

So much for the moment.

Best regards, Hans-Gert

--

apl. Prof. Dr. Hans-Gert Graebe, Inst. Informatik, Univ. Leipzig
postal address: Postfach 100920, D-04009 Leipzig
Hausanschrift: Augustusplatz 10, 04109 Leipzig, Raum P-633
tel. : +49-341-97-32248
email: gra...@informatik.uni-leipzig.de
Home Page: http://www.informatik.uni-leipzig.de/~graebe