Package Features As of version 0.5a4, PEAK features include: * A component binding framework that makes it easy to parameterize components and thus more easily combine and "wire" them together. Interfaces, adaptation, and "assembly events" (notification when components have been engaged as part of a "complete" application) are all available. * A comprehensive configuration framework that allows accessing "utilities" and "configuration properties" in context. Properties and utilities can be loaded or computed on demand, supplied by rules, defined in configuration files or code, in a supplied or custom format. Properties and utilities are contextual and can be safely acquired from parent/context components automatically. * A "document-driven testing" framework similar to "FIT":http://fit.c2.com/ that lets you automatically extract test data from design and requirements documents created with popular word processors, then create HTML reports on what passed or failed. * Naming system/framework that's midway between J2EE's JNDI and CORBA's cosNaming in features, but much easier to use and extend than either of those systems. * An "event-driven" programming framework that supports ultralight "microthreads" implemented via generators, and interacts with Twisted or stands alone. * A storage management and persistence system, including: - Atomic, multi-database transactions with two-phase commit. - "Data Manager" class framework for persistence management, that allows you to separate business logic from storage implementation. If you can write a few simple methods like "load" and "save" for a given object type and storage approach, you can create your own "DM" components. You can think of a DM as an advanced form of Python "shelve", that supports references to other objects, transactions, arbitrary back-end storages, and caching. - "Stackable" data managers: one DM might serialize a set of objects to XML, which could then be stored in a database record by another DM, and then the database record might be implemented via a DM that writes to disk files! Each DM only needs to know how to manipulate objects offered by the next-level DM, not the details of the next DM's implementation, so all the DM's are potentially replaceable with alternate storage mechanisms. - RDBMS and LDAP connection framework based on the Python DBAPI, that handles data type conversions (via the configuration framework) and seamlessly integrates with the transaction system and naming services framework. DB Connections can be accessed by name or URL, and bound as default collaborators or utilities for access by other application components. * CASE/modelling tools: PEAK includes APIs to read object models created in the XML-based XMI format. Many open-source and commercial modelling tools support XMI, inlcuding Argo/Poseidon and MagicDraw UML. PEAK includes pre-built support for UML versions 1.3 and 1.4, and MOF 1.3.1, using XMI versions 1.0 and 1.1. (UML 1.5, CWM 1.0, CWM 1.1, and XMI 1.2-2.0 are anticipated for version 0.6.) Also included is a MOF->Python code generator, which was used to generate the UML support, and which you can use to generate support for other modelling languages based on the MOF. For the specifications of XMI, MOF, CWM, and UML, visit: http://www.omg.org/technology/documents/modeling_spec_catalog.htm * A domain modelling framework for creating "business object models" with unidirectional and bidirectional associations, generated getters/setters and validators for fields, etc., and all necessary persistence support for use with the PEAK storage framework. Domain types can also define string parsing and formatting syntax, so you can create domain-specific data languages or just string formats for data types (such as specialized date/time or currency types). The business object framework supplies structural metadata about classes built with it, so you can query a class for its fields and links, and their names, types, etc. This can be useful for implementing model-driven storage or user interfaces. And the metadata is aligned with the MOF, so generating MOF, UML, or CWM from PEAK models (and vice versa) is possible (although not yet implemented for anything but MOF->PEAK). * Application Runtime tools, including: - a "command objects" framework for creating command-line applications, including the ability to create "executable configuration files" or "configuration interpreters" that can load a configuration file and run an application instance constructed using the configuration data. Supported formats include an .ini-like PEAK format, and arbitrary schemas defined using ZConfig. - a "periodic tasks" framework for executing tasks that perform "as needed", scheduling themselves in response to their available workloads - a CGI/FastCGI publishing framework that uses 'zope.publisher' to publish a PEAK component tree and its associated transaction service - an event-driven "reactor" framework that seamlessly integrates with Twisted, but can also be used without Twisted for applications that are mostly scheduling-oriented, or which use only third-party protocol implementations such as FAM, FastCGI, ReadyExec, etc. - a robust and flexible logging framework that can integrate with the PEP 282 logging module, or stand alone. It's simpler than the PEP 282 system for simple log configuration, and is configured on demand rather than "up front", and is thus more manageably configurable for large or complex applications consisting of components from diverse providers. - a "process supervisor" framework for multiprocess servers using long-running child processes (created via 'fork()') to take maximum advantage of multiprocessor machines for CPU-intensive services. * AOP and SOP: PEAK allows you to separate concerns as modules, then combine the modules via a "module inheritance" technique. This lets you define a generated business object model as a "structural" concern, and then combine it with a "behavioral" concern. This is as simple as writing classes that contain only what you want to add, and then telling PEAK that your new module "inherits" from the generated module. This is similar to (but designed independently from) the "MixJuice" tool for AOP in Java.