Babase uses common and widespread Unix development tools and techniques[24] to minimize a new developer's learning curve. This is a vain hope. Babase is complex and contains a lot of moving parts.
The remainder of this section describes conventions and procedures that those working with the Babase source code are expected to follow. It is of interest primarily to those who work with, or are considering working with, the code. It is not a comprehensive list, guidance should be taken from the existing code.
Anything and everything that is part of Babase should be checked into the project's revision control system.
All data values used in the code should be abstracted,
either via m4 or PHP defines, using names that begin with
“bb_”.
Minimize hardcoding. The use of data values in the code should be minimized. By keeping the number of hardcoded values to a minimum, the values used within the system can be altered through procedural changes alone, expensive programming can be avoided, and the flexibility of the system is increased.[25]
All database extension, triggers, functions, etc. should be written in PL/PgSQL, supplemented by m4.
All stand-alone programs should be accessible via the web. They should be written in PHP and styled with CSS2. The web pages they produce should be XHTML 1.0 compliant and should pass W3C validation at http://validator.w3.org/. Style sheets should pass the CSS validator at http://jigsaw.w3.org/css-validator/. Programs that access the database should obtain their PostgreSQL login credentials from the user, preferably using the existing PHP library code.
Each database user must be assigned unique login credentials to the PostgreSQL database. Each user is responsible for the security of his own login credentials and should never use login credentials that are not her own. All code should support this paradigm.
Every file should begin with a statement of copyright.
Each program, function, or procedure should have documented: its input arguments; its return value; any side effects including changes to pass-by-reference arguments, changes to the screen, changes to the database cursors, etc.
Clarity in your code is more important than efficiency. If the code is not clear, it is less likely to work and more likely to have bugs introduced upon maintenance. There is no point in getting a wrong answer quickly.
See the README files in the source tree's directories for information on how the source code is organized.
[24] Usually.
[25] This is very important but the reasons behind it are not obvious, coding values into the programs means creating office procedures that cannot be altered without a programmer. For example, encoding the value of the unknown group into the system would make it impossible to create different unknown groups for animals disappearing from different groups, or different unknown groups for animals disappearing in varying states of health, or whatever.