Some of the common ways of providing information were by editing a text based form , or using a dedicated program which runs on a personal computer and interactively guides the submitter in the steps of submission and as an end result, creates a formatted text file. The text file is then sent by email. There are drawbacks to these methods. Editing a text file is not interactive and there is no machine based proccess to validate the completeness of data. The standalone program on the other hand, requires re-writing of the code for each platform. It is also very difficult to keep the program up-to-date with the dynamically developing database. These drawbacks led to a search for better ways of interactive data collection.
The recent development of WWW based data entry forms opened a door to an excellent interface between providers and maintainers of information. WWW forms have many advantages. They provide a uniform graphic user interface. They are machine independent and they provide means for a constant, online update and maintenance by the server administrator.
HTML documents, and especially forms are lend themselves gracefully to object oriented development. The fact that each element of an HTML document is created by a standard primitive fits well with the concept of objects. This is especially evident in form elements that have a "TYPE" primitive in their tags.
Processing of the data stream which is provided by a submitted WWW form is done by a CGI script. When the data submission process is complicated and involves many validation steps of field contents and many forms, a simple script is not sufficient and a need for a high-level program arises. The Python programming language is especially suitable for writing such programs. Python's unique combined ability to serve as a fast scripting language and at the same time as a high level object oriented development language turns it into a very powerful development tool for WWW development.
This article describes a data submission system that has been developed by the author at the EBI. The system is based on a collection of Python objects which define the basic properties of the system and the basic steps in the cycle of aquiring, validating and submitting information. The objects are defined in a generic way which enables the same basic system to cater different data collection schemes. The design of the system makes it easy to support various concepts such as data hiding, dynamic creation of forms, preservation of data items for later use and providing context sensitive help. A data submission system for the EMBL database has been released to the public in May 1995 and it has been operating successfuly since then, providing more than 850 data entries so far.