Syndicated content-push in the Internet is a huge success, and web feeds are being used to convey various types of content: from news headlines, to podcasts and video podcasts, to being a feature in Web 2.0 websites. This diversity lead to the appearance of several frameworks, each tailored to a specific content type. At the same time, interest in social networking exploded, as more and more websites of this purpose were launched. Syndicated content and social networking websites are now intimately connected.
In this work, a generic, modular, p2p content-push architecture is proposed. It provides an evolutionary upgrade path based on the technologies already in use in the Internet for interoperability, thus without disrupting current infrastructure or changing participants’ habits. It also leverages social networks for content discovery and aggregation, using peer-to-peer protocols for distribution. A partial implementation of this architecture, dubbed SEEDS, is also presented.

Since the dawn of times, curiosity and necessity to improve the quality of their life, led humans to find means to understand everything surrounding them, aiming at improving it. Whereas the creating abilities of some was growing, the capacity to comprehend of others follow their steps. Disassembling physical objects to comprehend the connections between the pieces in order to understand how they work together is a common human behavior. With the computers arrival, humans felt the necessity of applying the same techniques (disassemble to comprehend) to their programs.
Traditionally, these programs are written resorting to general-purpose programming languages. Hence, techniques and artifacts, used to aid on program comprehension, were built to facilitate the work of software programmers on maintaining and improving programs that were developed by others. Generally, these generic languages deal with concepts at a level that the human brain can hardly understand.
So understanding programs written in this languages is an hard task, because the distance between the concepts at the program level and the concepts at the problem level is too big.
Thus, as in politics, justice, medicine, etc. groups of words are regularly used facilitating the comprehension between people, also in programming, languages that address a specific domain were created. These programming languages raise the abstraction of the program domain, shortening the gap to the concepts of the problem domain.
Tools and techniques for program comprehension commonly address the program domain and they took little advantage of the problem domain. In this master’s thesis, the hypothesis that it is easier to comprehend a program when the underlying problem and program domains are known and a bridge between them is established, is assumed. Then, a program comprehension technique for domain specific languages,
is conceived, proposed and discussed. The main objective is to take advantage from the large knowledge about the problem domain inherent to the domain specific language, and to improve traditional program comprehension tools that only dealt, until then, with the program domain. This will create connections between both program and problem domains. The final result will show, visually, what happens internally
at the program domain level, synchronized with what happens externally, at problem level.

Due to referential transparency, functional programming is particularly appropriate for equational reasoning. In this thesis we reason about functional programs by calculation, using a program calculus built upon two basic ingredients. The first is a set of recursion patterns that allow us to define recursive functions implicitly. These are encoded as higher-order operators that encapsulate typical forms of recursion, such as the well-known foldr operator on lists. The second is a point-free style of programming in which programs are expressed as combinations of simpler functions, without ever mentioning their arguments. The basic combinators are derived from standard categorical constructions, and are characterized by a rich set of equational laws. In order to be able to apply this calculational methodology to real lazy functional programming languages, a concrete category of partial functions and elements is used.
While recursion patterns are already well accepted and a lot of research has been carried out on this topic, the same cannot be said about point-free programming. This thesis addresses precisely this component of the calculus. One of the contributions is a mechanism to translate classic pointwise code into the point-free style. This mechanism can be applied to a λ-calculus rich enough to represent the core functionality of a real functional programming language. A library that enables programming in a pure pointfree style within Haskell is also presented. This library is useful for understanding the expressions resulting from the above translation, since it allows their direct execution and, where applicable, the graphical visualization of recursion trees. Another contribution of the thesis is a framework for performing point-free calculations with higher-order functions. This framework is based on the internalization of some basic combinators, and considerably shortens calculations in this setting. In order to assess the viability of mechanizing calculations, several issues are discussed concerning the decidability of equality and the implementation of fusion laws.

Current usage of computers and data communication networks for a variety of daily tasks, calls for widespread deployment of fault tolerance techniques with inexpensive off-the-shelf hardware and software. Group communication is in this context a particularly appealing technology, as it provides to the application programmer reliability guarantees that highly simplify many fault tolerance techniques. It has however been reported that the performance of group communication toolkits in large and heterogeneous systems is frequently disappointing. Although this can be overcome by relaxing reliability guarantees, the resulting protocol is often much less useful than group communication, in particular, for strong consistent replication. The challenge is thus to relax reliability and still provide a convenient set of guarantees for fault tolerant programming. This thesis addresses models and mechanisms that by selectively relaxing reliability guarantees, offer both the convenience of group communication for fault tolerant programming and high performance. The key to our proposal is to use knowledge about the semantics of messages exchanged to determine which messages need to be reliably delivered, hence semantic reliability. In many applications, some messages implicitly convey or overwrite other messages sent recently before, making them obsolete while still in transit. By omitting only the delivery of obsolete messages, performance can be improved without impact on the correctness of the application. Specifications and algorithms for a complete semantically reliable group communication protocol suite are introduced, encompassing ordered and view synchronous multicast. The protocols are then evaluated with analytical and simulation models and with a prototype implementation. The discussion of a concrete application illustrates the resulting programming interface and performance.

Protocolos de difusão fiável em grupos de processo são uma ferramenta de programação utilizada em sistemas distribuídos e confiáveis para isolar cada aplicação da complexidade decorrente da comunicação e das diversas faltas que podem ocorrer no sistema. Isto é conseguido oferecendo às aplicações uma abstração de um canal de difusão que oferece garantias muito fortes quanto à qualidade de serviço, o que é suficiente para garantir a correção de um largo espectro de aplicações com um esforço mínimo por parte do programador.
A concretização de protocolos de difusão fiável sobre redes em grande escala introduz um nível adicional de complexidade, pois as características destas redes dificultam a concretização correta e eficiente de serviços que oferecem garantias muito fortes.
Diversas propostas procuram ultrapassar este problema disponibilizando serviços que oferecem menos garantias de qualidade de serviço, mas que apresentam desempenho superior nas situações em que são utilizáveis. No entanto, quanto menos garantias se oferecerem, menos aplicações podem ser correctamente suportadas, uma vez que diferentes aplicações toleram diferentes relaxamentos da qualidade de serviço.
Nesta tese propõe-se um protocolo de difusão fiável que possa ser configurado para cada aplicação, de modo a poder ser simultaneamente eficiente e correcto porque adequado à aplicação em causa.
Para o efeito, começa-se por decompor a especificação de protocolos de difusão segundo um conjunto de parâmetros aplicáveis em separado a cada uma das mensagens de uma sessão, que podem ser combinados em especificações de protocolos adequados a cada aplicação.
Propõe-se então uma estratégia de concretização aberta orientada por objetos, que permite compor protocolos a partir de módulos independentes para cada mensagem, correspondentes aos parâmetros de especificação.

In this work we introduce a formalism for specifying the behaviour of computational systems, based on logic.
Our basic unit of software specification is called agent and is nothing more than a particular logical system. The agent logical language allow us to express the relationship between the occurrence of events and the properties an agent exhibits. The events are seen as logical connectives and the logical assertions reflect the causality between events and properties - the essential cause of a property in face of an event.
Because we are interested in knowing which sequence of events “justify” a property, we have traces coming from the past to the present. Our perspective is that of knowing which behaviour could justify the characteristics of au agent in some instant. Obviously, these characteristics will condition the future behaviour of the agent. This dependency is expressed in a deduction system that reflects the notion of causality.
For the purpose of characterizing the agents semantically a connection is made between the logical system and its topological models. The Stone duality provides a junction between the logical system and its semantics and it provides two alternative perspectives:
the algebraic or logical perspective and the topological perspective.
We present two types of models for the agents: the trace model, a denotational model intimately connected to the notion of event and trace, and the state model, au operational model based on the notions of state and state transitions.
We establish a relation between these two types of models: we see how a state model may be obtained from a trace model. The states arise as equivalence classes of past behaviours. We also study the properties of the state models which result from the conversion of trace models. We see how topological properties — such as separation conditions, soberty and coherence — characterize the state space of these modeis, and how from behaviour we can generate a minimal state space.