OK, that was alot. I'll try to respond to each point starting from the
I agree. I think CDC is an incredible program, especially considering
that it was done all by one person up to this point. Patrick has done
an incredible job so far. I also see much potential in CDC to become
even more powerful and flexible.
I also worry a great deal about great programs like CDC being developed
by only one person. This program is relatively complex. If anything
were to happen to Patrick, we would all be at a loss. I don't think any
of us could pick this program up by ourself and continue work on it.
I agree with you that the priority for CDC should be documentation.
With a good set of documentation, anyone that want to can spin off their
own version. For that reason alone, I am willing to help out in
documenting CDC. Unfortunately, I don't know much of anything about
Pascal or Delphi.
That brings me to my next concern. Without knowing much about Pascal, I
wonder how closely tied this project is to Borland Delphi. I've
downloaded a trial of Delphi, but it is time limited. I went searching
for a free Pascal development tool. All I could find was Lazarus. I
could not figure out how to load up CDC and build it. But, like I said,
I'm pretty ignorant when it comes to Pascal. I wonder if the next step,
after documentation, would be to make CDC independant of Borland products.
The next step after that would be to try to fully OO-ize CDC. Having a
nearly 100% OO CDC would aide greatly in any future code conversions.
Which I think brings me to my final point. I think a Java/C# version of
CDC would be far more maintainable long-term project than a Pascal
version. There are far more people capable of coding in Java/C#. And,
a fully OO codebase would be easier to maintain as well.
So, with all that said, I'm here volunteering my time. Starting about
mid-January, I will probably have quite a bit of free time on my hands.
At least for a few months. Although I spend the majority of my time now
coding in C, I am capable of coding in Java and C#. Unfortunately, I
have no real experience coding for a windowing environment, so that may
be quite a steep learning curve for me.
Well, that enough for now. Let me know what you think. Tell me if I'm
crazy or not.
Mathieu Clerté wrote:
> Hi Dan,
> Well I indeed planned to port CDC in a java-like language (in fact,
> building the core in parallel for java and C#)...
> 1) The code in 80% undocumented at this stage, and I am working on
> with Patrick *when we both have time* ;
> I'll probably have a litlle more time from next month.
> 2) We really need to consider the *reasons* for porting it... (As the
> program runs now both on win32 and linux)
> and also would it be *porting* (=translate from pascal to java/c#) OR
> re-designing... since a part of CDC has came from the non-OO original
> 3) I am at this stage the only person who wants to do it (= any
> java/C# guys want to join?) and Patrick is not keen in moving to
> another platform (as he masters Delphi already), so it needs to be a
> project which HAS to evolve IN PARALLEL WITH the pascal version of CDC
> ; it means that we need to establish *interfaces*, not only CVS...
> Interfaces and mini-rules which would explain the changes made to the
> delphi version, and most of all which would require those changes to
> follow the interface so that they could be easily applyable to the
> java-like version.
> The first step is to design and document the original pascal code with
> the help of Patrick and build a solid doco to start from before
> implementing anything.
> The second step is *mlaybe* to re-think the software in OO terms.
> The more CDC will be split into modules, the easier it will be to
> upgrade all versions (and actually to understand what's happenning).
> I think it's a great software, but at this stage, it's home-made and
> Patrick would enjoy some help for other developers to clean everything
> and draw new roads for the future of CDC.
> All comments and remarks are welcome ;)
> Mathieu Clerte
> Queensland University of Technology, Brisbane, Australia
> This SF.Net email is sponsored by:
> Sybase ASE Linux Express Edition - download now for FREE
> LinuxWorld Reader's Choice Award Winner for best database on Linux.
> Skychart-devel mailing list