On Gnu/Linux and Unix platforms, we have a set of established packaging systems (apt, rpm, bsd ports etc). However they all address the issue of automated application provision (install->configure->activate) specificaly for operating systems. What's missing is a tool that performs simillar functionality, but aimed at web applications. Which is precisely the intention behind the development of StAc. It is an application deployment framework, built to automate provision of multiple instances of applications, specificaly web applications. It also keeps the record of end users of the applications that it deploys.

First alpha relase with limited functionality is planed to be released in September 2003. Functionality will be limited to command line interface, installation and configuration of at least two tools (twiki and web serving space). Ability to create mailing lists using StAc will depend on what interfaces or hooks Sympa (mail manager that we will be using) offers. StAc source code will be available in public CVS repository.

Ability to provide(deploy) web application in fraction of time required by manual provision. Make updates and backups of web applications automated across the web servers. As a result of all that, (we) web servers maintainers will have much more time to focus on the end users and software tools, and boring repetitive technical tasks will be cut down. It will also provide us with coherent web servers, which is a significant advantage over ad-hoc (usual) way. For example, we will know at at any moment for all web applications who are they provided for and where all their files live on the server.

StAc is very modular and intend to keep it that way. Here are its main components right now:

( TODO: description of the role of the Worker, most important component, needs to be written ).

Task	Implementation	Comment
UI	Curses::Application, Getopt::Long
Local Storage (Stac's data)	XML::LibXML, XML::PatAction
Remote Storage (Tools)	Cvs
Caching	Cache::Cache
Logging	Log::Log4perl	Retire your debugger, log smartly...
Internationalization	Locale::Maketext, I18N::LangTags
Exception handling	Exception::Class
Templating	Templating Toolkit
Fuzzy matching	String::Similarity	will be used for project name matching (code written, but disabled for now)
Config file format	Config::ApacheFormat	XML is another possible implementation, but apache format is more human readable (less clutter) and known to almost all sys-admins

for detailed dependencies on other perl modules(classes/frameworks) and versions needed, see StAcMakefile.

concept diagram

see StAcDiagrams#concept_diagram

UML class diagram

see StAcDiagrams#UML_class_diagram (needs updating after first alpha release)

data model

see see StAcDataModel

storage data model

see StAcDataStorageModel

File Layout

see InfraStructureFileLayout^?