The Classify WorkBench

Highway To Productivity

Synchronization of source code and documentation has always been a major concern of MIS managers. The more a project progresses, the wider the gap between documentation and actual code. And with more advanced programming languages, programmers tend to skip technical designs right away and immediately reach for the programmer editor.

Yet, a good design and well structured end-user documentation is still much appreciated. Both by in-house users as well as by potential customers.

The Classify WorkBench is more than a code generator. It will actually help you create both technical and end-user style documentation and keep your code in line with the documentation.

Start With Documentation

Any software project should begin with a documented description of what should be built. This document should be presented to the future user (or customer) and after they have approved it, programming should begin.

Unfortunately, with the new, high-level programming languages, there exists a strong temptation to immediately start programming and building prototypes instead of first preparing a preliminary design. This is a temptation many developers cannot resist, especially when faced with the low rates, coming from working in the PC and network market.

When you decide to use the Classify WorkBench, it will allow you to start with the documentation. The WorkBench is based upon a large database 'tailored' to the Classify approach of application building. This means you will find the data structured in a way similar to class and object structures in Classify.

The Data Dictionary

In Classify, you will usually start by specifying the data dictionary. In the WorkBench, you will start by entering all information about the file into the database.

The entry screen has a number of familiar topics such as file name, flags to describe delete-behavior, rights and number of records. It also has ample space for a description and additional room for directions for the analyst or programmer.

Validation procedures and special 'hooks' can be selected from a library of functions. You can easily expand this library with new functions and new classes.

Also, you can specify all the fields in the file. For fields that occur in multiple files (like amount, invoice number, address) you can specify default characteristics.

Again, the information you specify per field is both documentation and a set of prescriptions that will easily translate into messages and properties.

Among the information specified are:

Field name and type.
Field description (both form and table oriented).
Data entry options.
Field-messages (validation, entry, exit, prompt).
Prompt and zoom object.
Short description (help line).
User documentation.
Directions for analyst and programmer.
Limited and normal tables.

Once you have entered all this information, the WorkBench will generate several documents for you. Depending on your selections it will include more technical information (like directions for the programmer) or just plain 'end-user' material. These documents contain formatting codes that you can specify yourself for easy import into any text-formatter or word processor. By default, the WorkBench is configured for use with Ventura Publisher and comes with a style-sheet matching the tags we use.

This document is more than 'just a report'. It "translates" property settings into readable information (again, the text is user-configurable). So, instead of "Set pDelete_Range false", the document will read something like "whenever the user tries to delete this record, the system will check if child-records exist; if so, the delete is aborted and an error is generated".

If you specify a validation table, the WorkBench will automatically list the choices from this table. And if you specify a field exit procedure, the WorkBench again will retrieve the documentation from this procedure and include it in the field description under the header "upon exiting the field".

Data Entry Modules

Once the data files are specified, you can specify zooms, subsystems and views. Again, these provide both descriptions and property settings (like auto activate, location).

Once the views have been specified, you can group the views into programs and the WorkBench again will generate documentation per program. Per program, it will list views, per view it will list the subsystem, per subsystem it will list the zooms and finally, per zoom it will list all fields in this file. This will allow you to find information very quickly without having to go back and forth through the manuals, and yet all information is entered (and maintained) only once.

Source Code Generation

Once the customer or potential user agrees to the specifications, you'll instruct the WorkBench to generate the code. It will generate images, data dictionary objects, zoom, subsystem, view classes and programs.

Through use of multiple inheritance and generation flags, the WorkBench allows you to add 'manual coding' to the generated program without losing the ability to regenerate.

The WorkBench will also automatically fill your system files (errors, tables and help files).

It's your ideal development partner!

Where can you get yours?

Ask us at emory@emory.com, or call

Calvin Consultancy USA, Inc.
Ft Lauderdale, Florida USA
(305) 742-2992 voice
(305) 742-2452 fax

Back to Emory Computer Consultants' home page