Web-enabled Information Delivery and Analysis System
Search:
Our Product Customer Solutions About Us Contact Us
Features Benchmarks Applications FAQ

WIDAS Frequently Asked Questions

Why does WIDAS rely on its own proprietary database management system (DBMS) and not on commercially available DBMS's? How large can a WIDAS database be? What platforms does WIDAS run on and what efforts are required to migrate to a new platform? How efficient are Web-based WIDAS applications that involve extensive interaction with a remote database? What data analysis components are integrated in WIDAS? How is the quality of the software and the validity of the statistical components ensured? MSI is still a small company. What assurances can you give for the perennity of WIDAS software? How long does it take to become skilled in WIDAS application development? Are API modules (dynamic link libraries) supported by WIDAS?

Why does WIDAS rely on its own proprietary database management system (DBMS) and not on commercially available DBMS's?

Large-scale database and data mining applications require aDBMS that is well adapted to handling complex on-line queries and that is highly efficient at performing mass selections of data structured in multidimensional and relational models.

The development and maintenance of a DBMS that meets these requirements does indeed require substantial efforts and resources. However we have found out from our own experience that integrating software tools from different sources (DBMS's, data analysis systems, application modules, application generators etc.) requires even greater efforts than developing them ourselves.

Our approach to building WIDAS as a highly integrated system, made up of components that we have developed ourselves, has proven to be economically advantageous and technically sound. WIDAS does not have any of the interfacing and performance problems that assail typical application systems built out of heterogeneous parts.

How large can a WIDAS database be?

The size of databases is limited in practice by the capacity of your storage media. Databases with data volumes in the terabyte range can be readily handled by WIDAS. An individual data table in WIDAS can have up to 8000 columns and 200 million rows. Several productive WIDAS applications, such as STATINF (the official statistics provided by the Swiss Federal Statistical Office) operate with tables approaching these dimensions.

What platforms does WIDAS run on and what efforts are required to migrate to a new platform?

Currently the following platforms are supported:

WIDAS is perfectly consistent on all supported platforms. In fact, the WIDAS developers are free to work on their platform of choice.

In order to migrate WIDAS to a new platform, the WitCC source code is transferred as ASCII files to the new platform for compilation. The WitCC program generator then takes care of the platform specifics, such as I/O, character sets, screen handling, graphical output etc.

How efficient are Web-based WIDAS applications that involve extensive interaction with a remote database?

WIDAS Web-based applications make extensive use of current browser technologies. Through the intelligent use of Ajax and DOM, data transfers between the browser and the WIDAS application are kept at a minimum.
The user's browser window is refreshed locally in order to obtain a stable, flicker-free display.

What data analysis components are integrated in WIDAS?

WIDAS contains a full-fledged repertoire of statistical procedures, ranging from descriptive to inferential statistics, including parametric and non-parametric hypothesis testing. Also included are sophisticated multivariate methods, such as multiple regression, factor analysis, analysis of covariance, discriminant analysis, cluster analysis and analysis of time series. The focus in developing the data analysis components of WIDAS has been on ease-of-use and ease-of interpretation. We have also emphasized data presentation through powerful graphics and reporting facilities. Some of the visuals produced with WIDAS can be seen in WIDAS "the Product".

WIDAS has been driven throughout by the practical requirements of its users. This has led, for example, to the implementation of strong data management features that deal with missing values and to the development of a uniquely efficient solution for performing cluster analyses on very large data sets. (It is well known that cluster analyses normally require computational times proportional to the 3rd power of the number of observations, which becomes prohibitive in cases with 10'000 or more observations).

How is the quality of the software and the validity of the statistical components ensured?

WIDAS software is subjected to rigorous quality assurance procedures encompassing a large number of published test cases. As far as the statistical components are concerned, they are based on verifiable methods published in the literature. In addition, the continuous feedback received from experts in the field, particularly from scientists and statisticians working at the six Swiss Federal Agricultural Research Stations and the Swiss Federal Statistical Office, contributes significantly to the assessment of the quality and stability of WIDAS.

MSI is still a small company. What assurances can you give for the perennity of WIDAS software?

How long does it take to become skilled in WIDAS application development?

An experienced programmer, knowledgeable in one of the current programming languages will learn WitCC within one to two weeks. The fact that the sources of WIDAS applications are contained entirely as WitCC code makes development and maintenance of applications very efficient.

Are API modules (dynamic link libraries) supported by WIDAS?

The WitCC generator of WIDAS/3 can automatically generate API modules. These modules can be used as dynamic link libraries by other software such as ORACLE Forms or Microsoft ACCESS. Through API modules applications outside of WIDAS can access WIDAS databases.