One of the main reasons for writing a new data processing software was to address some of the frustrating issues I faced with Quantum.
- I wanted to get rid of the 200 chars per line restriction - a source of many core dumps and unexpected behavious
- Edit restriction like a prefined limit on the number of edit statements
- Label number for loops and branching statements - that too an upper bound of 500 statements
- Subroutines could not be called from a loop
- Weak type checking
- Record length limit of 32768
- ... the list goes on
Layout independence from column numbers. Ideally I would like the data map to be embedded in the data file. The burden of dealing with column nos should rest on the compiler not the DP
There is a lot of scope for computer assisted table checking. For example while processing data - we have a raw column number - it is easy to program a function to check the base of the table vs non-blank holecounts. In fact I had this working in a interpreter prior to the compiler.
Print tables to a variety of formats including html and pdf
Designed for performance. Data file will be in binary format
Automatically detect incorrectly formed data files - and stop data processing,

Easy to use questionnaire scripting language - with support for C like data types.
Map should be modifiable even after data has been entered
Statements which can reserve columns for future use
Hidden questions which can be exposed after coding is done - during data entry on 2nd pass it will only prompt for hidden questions
Print a pdf questionnaire from the script with routing instructions

Data conversion S/W goals - The glue between the backend and front end

Ability to convert data files from one architecture to another - ex: i386 to SPARC
Convert data from qscript to xtcc backend format

dpassist Goals - This may seem a bit cryptic but will elaborate more when there is something to demonstrate

If just a binary data file is present - the program should setup axes based on the data types in the embedded map
If a qscript file and data file are both present - program should output a detailed axes with stub information, include files and filter conditions