Only for fun: Making a text to speech converter using VB .NET 2008

Powerful speech-based interfaces, available in your language, will enhance the user experience and let people access information more naturally and effectively wherever they are. Microsoft’s speech technology already powers great ready-to-use applications like Outlook Voice Access in Microsoft Exchange 2007 Unified Messaging, Windows Speech Recognition on Windows Vista, Speech Server on Office Communications Server 2007, and Voice Command on Windows Mobile. In addition, developers can leverage powerful APIs and tools to build their own speech-enabled applications using the Speech Server Developer tools for Visual Studio 2005 and the System.Speech API in .NET 3.0.

Transcripts from ISSRE 2009 presentation - Business Rules Separation and Reuse Using MDA, OWL and AspectJ

Motivation

Usually the developers inserts code in the specific platform model (PSM) directly. So that, it doesn't have guarantee of dependability by the programming errors, beside that tedious. The lack of mechanisms by automatic transformation of business rules becomes one important trouble for model, reuse and maintenance trustworthy of a software base on MDA. For that, so necessary increase trustworthiness and security in the functioning (dependability), that are characteristic essentials for existing critical applications in the industry.

Objectives

In our approach, an ontology specifies domain properties and translates it to aspect-oriented code. The implementation code could be automatically merged. As a result, we will be able to perform approval tests at run-time and to create business rules reused in other MDA specific models. This novel approach shall be evaluates in a toolkit for re-engineering an industrial furnace management application.

Development Process Outline

An ontology about application domain is conceived with the domain properties formed by the functional requirements found after domain analisys of software. In its original conception, the MDA defines business rules in the platform specific models, causing some problems already reported. Own proposes defines the business rules on ontology level and confirm its semantics validation in a conceptual domain before PIM generation model.The domain ontology is generated into a platform independent model (PIM) using an ontology edition tool and then is converted to UML language. The PIM business rules will be transformed in a PSM model across of OWLtoAspectJ. Several PSMs models can be generated, being: one representing the business rules and the others for specific platforms (e.g. Java IF, Java EE, RMI, Corba). Finally, the final application combines the business rules and application PSMs using a composition process of models.

Main contributions

In making an approach to emerging technologies (MDA, OWL and AOP), was built a tool specifically for it (OWLtoAspectJ). In this boarding, was add a set of activities to orient the development process using MDA, OWL and AOP. Our approach uses OWL-DL and a reasoner tool (Racer) instead of OCL. It helps to verify the domain properties in a formal way for critical systems. The UML language does not support it actually.

Case Study – Furnace management system overview

Reengineering of industrial furnace management application of Ferroalloys (e.g. Ferrochrome). For each alloys iron exist a recipe that has inputs (e.g. carbon, chromium, silicon). Theses inputs are put on in the industrial ovens in according to recipe made by production analyst. Sensors sends the plan data for a programmable logic controller (PLC) system managed.

Case Study – Domain Engineering and Development Process

•Conceptual modeling of application domain
oFunctional requirements and business rules analysis
oUML for knowledge representation
oDomain modeling using ontology
oVerification of domain properties
1. Domain modeling - Definition of concepts, axioms, properties, relationship between concepts.

2. Verification of domain properties - Business rules definition using PAL (Protégé Axiomatic Language)

•Instance creations and business rules

•Generation of application domain from ontology to OWL
oClasses, properties, relationship between classes, axioms

•Building PSM model to Java platform
oUML backend (Protégé plugin for XMI)
oGenerate from PIM UML model to application Java code
oImporting domain model XMI using UML tool

•Building PSM model to business rules
oSeparation and generation automatically of business rules from conceptual domain in OWL-DL format

Verication of domain and business rules

The protégé ontology tool allows to create objects during ontology domain modeling. It favors the model creation and very close to the implementation of the final application code. It too, is an ease manner, by not having the need to instantiates objects using a programming language to check the modeling. It usually is made when entering in the implementation code phase. In this step, were created several objects to test the comprehension of furnace management field, such as inheritance, business rules validation, association and aggregation classes. Other uses that can be conceived too, for example: design patterns test, in a high abstraction level.

Business rules in OWL format

The OWL language, express the structure, relationships and characteristics of the concepts of a field, satisfied by logical axioms of an ontology. The OWL is based on description logics and has a similar structure to XML. The description logics uses algorithms designed for the reasoning services, having the main goal, inferred about the correctness on concepts modeling in an ontology. The business rules are known by the tagt (owl: Restriction) in XML format. The domain restriction consists of an associated property with a class, it is referenced by the tag (rdfs: domain rdf: resource). The tag rfds: range rdf: resource in turn, determines which value should be accepted, such as instance objects. In this case, should be at least one oven and operator kind of instances to occur the charge of inputs in the furnace management application.

OWLtoAspectJ tool

The OWLtoAspectJ tool translates business rules from OWL format to aspects on Java language. Each business rule was transformed into an aspect, meaning, each aspect unit, represents a business rules base on conceptual PIM model, conceived and asserted about its satisfatibility using both: ontology and reasoner tools. The OWLtoAspectJ tool, was built using the JENA framework and a process translation, developed by our team, for the business rules processing.

Business rules translated to aspectJ

The PSM model to aspectJ should be generated automatically. in our case, the method to be intercepted by the aspects, has classe name related to concept, in your signature (eg Operator.charge_Oven() method). Therefore, we may use designators (eg execution, call, Get, Set), such as pointcut, with the same name of associated class. Thus, is not necessary to write the full method signature. The property "owl: ObjectProperty rdf: ID" in OWL format, represent the business rule name, therefore, was used to name the aspects, facilitating the business rules identification. The business rules, that has references to restrictions on objects types, (eg owl: allValuesFrom, owl: hasValuesFrom and owl: someValuesFrom) will use the instaceof operator in a conditional structure. While the cardinality operators rules will use the Java original operators (eg ">=", "<=" and "= ="). Business rules validation in run-time

The figure shows an attempted to execution of charge of oven by an administrator. This causes an exception occur, during application test, and pay your attention for the business rule method, where that was violated (in red lines). Despite the current limitations of the static business rules, this process is extremely useful during the validation and testing of the system. Therefore, it allows the inconsistencies identification, in according to checks activities made, for each stage of system development, which in a manner, provide dependability of furnace charge. such as can be saw the figure in the presentation.

Related works

AOP and MDA, crosscut concerns specifically about non-functional requirements and business rules. CALA group, at INRIA lab in France. PSM models generation, in Aspect-oriented Language, too in the CALA group at INRIA lab, and in our distributed systems laboratory. MDA and OWL, domain modeling and models reuse using ontology at Object Management Group. Bridge, between OWL and UML, using Ontology Definition Metamodel in LORESFU group at Surrey University. Searches about UML formalization, used on both representation, Ontology Web language and Description Logic at Informatics center in federal university of Pernambuco, and in our group in Brazil.

Conclusion

In this approach, we thought that across of, needs to check the domain properties, defined for dependability of critical systems, was created a domain modeling process based on ontology. The definition of business rules, by an ontology editor, and validations them using reasoner, the Racer tool, has warranty of functional requirements comprehension. The business rules extraction, and validation, before generating the PSM, on PIM level, improved reuse of business rules easier, and can avoid implementation code fail. Because, it is generates automatically, in final application code merged.

Future works

Treat the business rules formalization, before, translates them to aspectJ, for increasing dependability of systems. To study aspects concorrency, and an evaluation process for that, from view point of business rules overlap. improve OWLtoAspectJ functionalities, to standardize the violation messages, and building a rules repository, to merge rules in software produtc lines. Other works in progress in LASID, involving formal verification, MDA model transformation with SPEM metamodel 2, and, ontology web language semantic rules for web services composition.

It see the presentation in ppt

Off topic - RIP Dr. Chandra Kintala

Dr. Kintala spent over 20 years doing research and managing research at Bell Labs in AT&T, Lucent and Avaya in New Jersey. Among other things, he invented with Dr. David Belanger a language and a software tool now called "Backtalk" in 80s; it is still used in AT&T for data analytics on very large databases. With Dr. Yennun Huang, he invented the concepts and components for Software-implemented Fault Tolerance and Software Rejuvenation in 90s which are now widely used in industry and academia. He was responsible for managing research on Distributed Systems and Network Software Research in Bell Labs and was also Vice President of Research Realization Center in Avaya Labs. While working in Bell Labs, he was also an Adjunct Professor and later Distinguished Industry Professor at Stevens Institute of Technology in New Jersey. In Sept 2006, he moved to India as the Director of Motorola Labs in Bangalore where he was responsible for research and academic partnerships in India. He moved to Yahoo Labs in Bangalore in Aug 2008 where he is the Director of System Sciences and Academic Relations in India.

Dr. Kintala received his Ph.D. in Computer Science from Penn State University and published 48 refereed research papers and received 6 US patents and a Smithsonian medal sponsored by Computer World in 1998. He was the General Chair of IEEE’s conference on Dependable Systems and Networks in Philadelphia in June’06, Acting Chair of IFIP WG1.2, Member of IFIP WG10.4, Senior Member of IEEE, keynote or guest speaker at several academic and industry events and member of several technical program committees. He was a member of FICCI and Pacific Council’s Joint Task Force on Global Innovation Economy – Enhancing India-US Relations.

Dr. Kintala passed away 11/05/2009 from a heart attack in New Jersey.

Paper accepted on Brazilian Information Systems Periodic (annual periodic)

Other paper was accepted, thus more boarding piece's it's now on Brazilian Information Systems Periodic of these year. More than full papers accepts that's the full growth of my boarding. I'm an assiduous research about business process, derivation rules in MDA (Model-driven Architecture) and software architecture advanced, mainly within the professional environment since 2006. A sum of fifteen is not so bad for a beginner. I'm so happy by gain this recognition and the respects by publication's.

Reference Cards

The reference cards was written by bestselling authors and leading experts with reliable information on major developer topics filled with useful tips and source code examples. PDF looks great on-screen or printed from your printer.

Airline Crew Scheduling Problem

I tried to read and understand the airline crew scheduling problem on past month. Is there really an NP-complete, I read some thesis and papers quickly (below). I think challenge extremely relevant, because this area does not well and certainly knowledge, not enough to define an algorithm to perform searches for each subgroup of data and find the best routes, cost, etc... You need to create an artificial intelligence so that it can adapt to the data volume that will be growing, so a big problem!

There is a need to distribute the processing and it involves logical clocks, timing and distributed systems fundamentals. I know well distributed algorithms: from ground to specification and development. As part of the growth difficult is returned from flights range, because it require advanced concepts about analysis and complexity of algorithms.

I guess also necessary to search for a boarding and to do comparisons between implementation technologies such as MPI, CORBA, ORB, OD, etc... I remember a colleague that has implemented a weather events simulator with MPI, because of synchronization problems and only a few tools provide security and efficiency of message exchange in this board.

I just bought 3 books in this year: numeric algorithms, algorithm analysis (Cormen best seller) and AI (Takashi Yoneyama). I thought to enroll on USP sun course: algorithms optimization and analysis, In April, I needed to learn data mining algorithms for SQL Server 2008 growth smart searching and to found value patterns. Of course that's really easier!.

I read the book written by Cormen since January when I have time, it's an area which, coupled with modern software engineering, can bring good results to improve performance in systems. I saw some boardings about distributed algorithms, whereas its implements backpack problems, I'll try a first approach for using it thought the crew scheduling problem.


Quick reading about:

•Joyce W.Yen, John R. Birge. A Stochastic Programming Approach to the Airline Crew Scheduling Problem.

•Jenny Díaz Ramírez, An Approach to the Integrated Single-Fleet Aircraft and Crew Scheduling Problem. ESCUELA DE INGENIERÍA Y ARQUITECTURA INSTITUTO TECNOLOGICO Y DE ESTUDIOS SUPERIORES DE MONTERREY.

•Ugur Özdemir. METHODOLOGY FOR CREW-PAIRING PROBLEM IN AIRLINE CREW SCHEDULING.

•António J. M. Castro, Eugénio Oliveira. Quantifying Quality Operational Costs in a Multi-Agent System for Airline Operations Recovery.

Off Topic - Composer day!

Today is the composer day. With this in mind, Whereas I writing a lot about my research activities here, in spite of it, I cannot forget my love by the writing songs. I am a composer, although I have done harmony for only 30 songs on the guitar. Without a doubt that I learned to read and play songs alone, In its turn, I have used just to make music arrangements and harmonic analysis that I so like.

Congratulations for all good musicians, specially: Steve Vai, Chick Korea, Yngwie Malmsteen, Jorge Lynch, Kiko Loureiro, and other that I heard very times when was youngest.

Business Rules Separation and Reuse Using MDA, OWL and AspectJ on ISSRE IEEE Proceedings 2009

This work presents a proposes to separate the business rules within the platform specific model (PSM) and allows its sinking in the implementation code. In this boarding, the business rules writes in Ontology Web Language (OWL) are translated for a LOA (Aspect-oriented Language). These are, in turn, combined and compiled to create the final application code. To certify own propose, will be made reengineering of industrial furnace management application that currently controls pyrometry of ovens, it demands critical needs for its functioning.

These paper was accept on 20th International Symposium on Software Reliability Engineering (ISSRE 2009).

Using Data Mining Algorithms in MS Office 2007 - Within the 22º Microsoft Architecture Journal

On the business environment currently, the data mining is gaining attention because such resource treats about exploration and analysis in an automatic or semi-automatic form and many data should help to found value patterns and rules. These patterns and rules improve understanding of companies’ areas such as marketing, sales, client support operations. After years, the corporations increase large databases and applications, as enterprise resource planning (ERP), customer relationship management (CRM) or other operation systems. The people believe in the existence of hiding values in these large databases and the data mining algorithms can be help in find this values.

In this paper will be made an explanation about data miner algorithms provided by the Microsoft Analysis Services 2008 currently (e.g. Decision Trees, Naive Bayes, Clustering, Neural Network, Logistic Regression, Linear Regression, Time Series, Sequence Clustering, Association Rules) and how use it for achieve the expected results. There a demonstration about data mining from Microsoft Office 2007 with use Add-in Data Mining proposing an overview about how to work the tables data using Excel 2007.

Short list of reader's takeaways from business and technical perspectives:

1. Use it to know the strategies for data mining on MS Excel 2007.
2. To know about data miner algorithms and its applied.
3. To use the correct algorithm for solver real troubles in the enterprise.

Efficient Management of Business Rules in Software Product Lines

Despite advances in the scientific community about MDA models building and reuse, many companies needs to use software products lines to implement the same business rules in several products for the same organization. Any change within the model business rules will imply its maintenance and updating of both PSM as PIM. The purpose of this work is to create a new way to keep ahead of software products to the constantly business rules changing in this scenario, through an effective approach for Efficient Management of Business Rules.

Full paper