MAHMOUD SAMIR wrote: > hi > > >> So what I would like to see from you (in order to convince ME to try >>> your framework) is one thing: please describe in plain text how you >>> would implement the following task in your framework; with as much >>> technical detail as necessary. >>> >>> The task is: >>> "There's a table called EMPLOYEES in a DB2 database under AIX. There's >>> a table named DEPARTMENTS in an Oracle database under Solaris. And >>> finally there's a table named GROSS_REVENUE stored in a MS SQLServer >>> table. >>> Combine the data from all three tables such that the output of the >>> program contains: >>> - name of every department, >>> - total revenue generated by all programmers in >>> this dept. >>> Keep in mind that departments are stored in a hierarchical manner. >>> " >>> [end of task description] > > ok, solving this problem (task) in programming without coding means that we > should > 1 - Design High Level RPWI (Real/Rapid Programming without Instructions) > Package > 2 - Implementation of the RPWI Package > 3 - Testing it, to make sure that everything done well > 4 - resuing it again and again as much as we need > > High Level RPWI Package means developing GUI that can be used without > understanding the programming language syntax and rules, and interaction with > this GUI gives results that do something (Goal) we want. > > Designing High Level RPWI Package (Design start from what we want , not > what we can do) is an art, and solutions differ from one to another > > So i will give you a solution, but you can present another one , and so on. > > [1] - the first step in designing High Level RPWI Package , is to determine > the *logical sequence* that you uses to solve the problem > in our task , for example, i will put the following logical squence > 0 - Create Connections to differnet data sources > 1 - Set state = firstdepartment (first record) = our first root > 2 - use our Connection to Oracle Database > 3 - Create view from Department table to get the department record by > (state) > 3.1 - Start Scan (Record by Record) in our view > 3.1.1 - Use our Connection to DB2 Database > 3.1.2 - Create view from Employees table to get all employees from > our department > 3.1.3 - Start Scan (Record by Record) in our view > 3.1.3.1 - Connect to MS SQL Server Database > 3.1.3.2 - create view from GROSS_REVENUE table for selected > dep & emp > 3.1.3.3 - print/store program output on screen/file/printer > 3.1.4 - update the cursor (for scan) > 3.1.5 - Continue/End scan as required > 3.2 - update the state to point to the next department in the > hierarchical strucutre > 3.3 - Continue/End scan as required > 4 - Clear Views > 5 - Close Connections > 6 - Return to caller > [2] the second step is to write a real code for each step written about in > the logical sequence
Aha! Snake oil. You DO have to write code at some point. And probably for each language you want to export to as well. Exactly what I was hoping you would admit. The above sentence is completely contradictory to the subject line of this whole thread. Another major problem I see with your tool is it only will work with sequential programming. Event-driven programming paradigms are difficult to impossible to implement with your tool. The main problem for people like me is we use Safe C++ methodologies when developing software. I've already done what your tool does within C++. Segmenting my code into base library and application layers is one of the best things I've ever done. This approach makes your tool useless. -- Thomas Hruska CubicleSoft President Ph: 517-803-4197 *NEW* MyTaskFocus 1.1 Get on task. Stay on task. http://www.CubicleSoft.com/MyTaskFocus/
