Hello all. I've decided that it's time to put a little bit of effort into deciding where to go next with ActiveWarehouse so here goes. First, let's look at where we are today:
* The ActiveWarehouse Rails plugin does a decent job of defining classes that can be used to model a dimensional data warehouse. This includes classes such as Fact, Dimension and Bridge. These classes, in conjunction with their migration generators can be used to construct a basic dimensional model. Additionally we have mix-ins for SCDs, Hierarchical dimensions and classes for dimensional views (role playing) and date dimensions. * The ActiveWarehouse ETL library is fairly full-featured. It has support for file source data including delimited, fixed-width and XML as well as inline data produced by a Ruby Enumerable, data from relational databases and data from Rails model objects. It supports a variety of built-in transforms as well as custom transforms and inline block transforms. Finally it supports output to flat files as well as databases and supports bulk upload through the adapter extensions library for MySQL, PostgreSQL and SQL Server. These are the bright spots. The areas which need improvement are: * ActiveWarehouse Rails plugin reporting tools. Right now AW doesn't give you much in terms of a reporting and query front end. The report_helper.rb code is horrendous and is basically all there is for reporting. * Aggregation in AW is all over the place right now with none of the aggregation implementations delivering much in the way of high-performance aggregation schemes. You can still execute queries through the NoAggregate implementation however for anything substantial you're talking about potentially slow queries. So what's next then: * Release ActiveWarehouse ETL 1.0. I think it's time to freeze this thing. It has documentation, a reasonable number of tests and is essentially usable right now so I think it's time for a release. * Focus on the AW front end. AW needs to come ready to go with at least some basic reporting tools on the front. * Clean out all of the old aggregation techniques and settle on something specific, perhaps based on aggregation techniques in other open source data warehouse libraries? What other items do you think should part of upcoming AW development? If you have time and are interested in contributing please let me know as AW can always benefit from more developers with a passion for data warehousing. V/r Anthony Eden -- Cell: 321 473-4966 Current Location: Melbourne, FL Home Location: Melbourne, FL _______________________________________________ Activewarehouse-discuss mailing list Activewarehouse-discuss@rubyforge.org http://rubyforge.org/mailman/listinfo/activewarehouse-discuss
