Message-ID: <697924373.1007.1432322349024.JavaMail.confluence@ece-vmapps> Subject: Exported From Confluence MIME-Version: 1.0 Content-Type: multipart/related; boundary="----=_Part_1006_103921112.1432322349023" ------=_Part_1006_103921112.1432322349023 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Content-Location: file:///C:/exported.html
|4. Statistical Metadata Systems (Stat= istics Austria)||Statistics Austria=||6. Organizational and workplace= culture issues (Statistics Austria)|
One of the cornerstones of Statistics Austria's IT strategy is the use o= f the database DB2 on the mainframe computer. Web applications run on Linux= servers (which are logical partitions of the mainframe). Server programs a= re developed in Java and deployed with IBM Websphere as application server.= To a large extent, PC software is also written in Java.
However, Statistics Austria still has a relatively heterogeneous, histor=
ically grown IT landscape which is composed of a large number of legacy sys=
tems and components. As a consequence one can find a multitude of programmi=
ng languages, development environments and IT architectures. A few examples=
ISIS: The statistical output database is a mainfra=
me application which was developed completely by Statistics Austria. PL/I a=
nd assembler were used as programming languages.
ISISGui: At the beginning of the new millennium a =
graphical user interface for ISIS was implemented as Java applet.
Publication Database: The core of the Publication =
Database is a commercial document management software (Stellent). Extension=
s and adaptations developed especially for Statistics Austria are written i=
n Java and a proprietary Stellent script language.
IMS: If the decision is made to realize the integr= ated metadata system IMS, the implementation will be carried out in accorda= nce with the IT-strategy (storage of metadata in DB2 on the mainframe, busi= ness-logic-layer as Web application under IBM Websphere on Linux, client so= ftware written in Java).
Due to the existence of several isolated metadata systems, there also ex=
ist a number of different metadata management tools. Some examples:
Standard Documentations: The=
standard documentations are written in Word using a predefined template, s=
tored in the Publication Database, converted to PDF and finally replicated =
on the Web server.
ISIS: Management tools for ISIS are mainframe appl=
ications written in PL/I or assembler.
e-Quest: = Statisticians define electronic questionnaires using a graphical editor (&q= uot;e Quest Metadata Manager") which was developed in Visual Basic 6.0= .
Classification Database: Adm=
inistration software for interactive editing and processing of classificati=
ons does not yet exist.
IMS: The development of a management tool which ca=
n easily be extended by plug-ins, based on the Eclipse RCP (Rich Client Pla=
tform) architecture, is planned.
Classification Database: The=
Classification Database was based on the Neuchatel model; however, the sys=
tem is not a complete implementation of that model.
IMS: Two of the proposed IMS subsystems ("Var= iables" and "Value Domains") are based on ISO 11179.
Version control of metadata is a complex topic. Within the framework of = the IMS it is planned to specify version objects whenever version control i= s found necessary during analysis (for example, a survey would be composed = of one or more survey versions). This approach was also used in e-Quest and= proved successful.
Regarding the IMS subsystem "Concepts and Definitions", it was= decided that identifying versions of definitions is not necessary because = handling them would be too difficult. It is the opinion of subject matter e= xperts that most of the definitions can remain static. If there are changes= in some definitions, these will be described in the corresponding text of = the definition.
Concerning revisions, it should be a feature of the metadata repository =
software that prior versions of an item are not deleted, but remain availab=
In the Classification Database versions of classifications can be = accessed.
Due to limited human resources the outsourcing of software development o= ften is inevitable. However, project management is still carried out by emp= loyees of Statistics Austria.
The Classification Database was developed internally (although a prototy= pe was written by an external company).