About Count From Zero (Draft)

For communications email cfz@countfromzero.biz.

The Concept

Count From Zero (CFZ) indexes software engineering libraries and tools. CFZ supports hierarchical categorization and multiple categories per software package. For example, if a user wants software for

      part-of-speech tagging

they can find it under

       top: artificial intelligence (AI): natural language processing (NLP): part-of-speech tagging (POS tagging)

and a package listed there might also be found under

       top: artificial intelligence (AI): natural language processing (NLP): shallow parsers

if applicable. This scheme enables site users to easily find software fulfilling specifically what they are looking for rather than sifting through dozens of irrelevant software packages.

CFZ's business model seeks advertising revenue from hosting a community for software publishers and users. CFZ favors long-term growth over making a quick buck. Fulfilling the community's needs takes precedence over squeezing out every dime of ad revenue.

The People

At this early stage CFZ is a sole proprietorship owned by myself, Olinga K. Abbott, and has no employees or other contributors. I am a Java software engineer living in downtown Chicago. I started CFZ from frustration experienced when scouring the Internet to find software to assist with my work. I work on this site full-time as my only employment, financially supported during the startup phase by my wonderful wife Sarah.

My resume is available. Email me if you are looking for software engineering talent and entrepreneurial spirit.

The Technology

CFZ started development in November 2005, went live at alpha-stage in August 2006, and entered beta-stage in February 2007 along with the first public announcement for the site.

Go Daddy hosts CFZ on a virtual dedicated server. Software development uses agile techniques when applicable for a single-programmer environment. CFZ is programmed in Java with some Perl sprinkled in.

The software CFZ uses includes

• Sun JDK,
• PostgreSQL,
• Apache HTTP server,
• Apache Tomcat,
• JUnit,
• Apache Ant,
• Eclipse,
• Jakarta Cactus,
• several packages from Jakarta Commons,
• HTTPUnit,
• Fedora Core Linux (production),
• SUSE Linux (development),
• Perl
• AspectJ,
• JMock,
• Mockrunner, and
• Subversion.

Most software used by CFZ beyond the basic operating system is compiled from scratch using a SSH command line interface (the script to do this is almost 300 lines long!) rather than the default software provided by the hosting account. CFZ does not use a control panel like Plesk or cPanel.

The Future

CFZ currently exists in beta and only supports about 20% of the full functionality intended and contains perhaps 2% of the desired data. The current feature set appears free of significant bugs. Among the planned features:

• LOTS more packages. CFZ hopes to index tens of thousands of software packages eventually and branch out into other languages (CFZ currently caters mostly to Java).
• A wiki page for each category and package. The category wikis will describe the categories in detail, explain the essential differences among the member packages, and make recommendations for member packages based on usage scenarios. The package wikis will primarily be "Quick Start" articles providing orientation and explaining how to get up and running with the packages as quickly as possible. In some cases links to external sites, especially Wikipedia, will be provided instead when suitable.
• Forums for packages that do not have an external forum of their own. CFZ will link to external forums when available rather than create a competing forum.
• Unless spam becomes a problem, users of the wiki and forums can choose to remain anonymous.
• Lists of external articles for categories and packages. This feature is already supported but little data has been entered.
• Software publisher accounts. Publishers can directly update the data about their software without waiting on CFZ to key it in.
• User accounts.
• Customizable RSS and Atom webfeeds. Users can subscribe to the master feed, feeds for individual categories, or create their own. The data can include updates, new entries, or forum posts.
• Preferences. Users can choose to only view software written in Python.
• Search capability, sorely needed since the site currently can only be navigated through the categories.
• Obtain Web designing services. The site looks hideous!
• News for the front page that lists updates and additions.
• A blog.
• More ways for users to slice and dice the data, perhaps even allowing SQL if it can be secured.

Some other long-term ideas for the site:

• Job boards for each package and category. Looking for a qualified applicant with 7 years of experience in programming multi-dimensional indexes for real time systems? Now you can easily find them!
• Software package ratings, dependent on publishers' reactions.
• Additional data for packages and categories, such as multi-licensing and listing all the languages and protocols each package supports.