this is a follow up on a post on the Open Office thread about a “feasibility study” to port Open Workbench to Linux. I started to look at the source code and I think it is feasible, especially if OWB is not going to evolve any more. Porting OWB to other platforms can be the only one to have this product becoming more popular from a user perspective as well as from a developer perspective.
After a quick look, three aspects of the software need to be addressed in this study :
1)Find a way to replace all the commercial tools currently used, mainly in the GUI layer. This may introduce substantial changes to the presentation layer (for the better and the worse).
2)Replace all the platform specific code and more specifically such as MFC goodies such as CArray, CString CMap with platform independent substitutes.
3)Finally, although no changes (nor a recompile) on the java part is needed as the current assumption, the current JNI implementation may need some tweaking / adjustment ...
Currently, as mentioned in some of the post about a port to Linux, WxWidget layer is, to my knowledge, the best candidate to take care of the presentation layer as well as the MFC classes. Some work on the JNI implementation will identify if any work is required.
To my knowledge (I don't want to start a heated discussion around this), Code::Blocks is the best IDE supporting WxWidget project on the Linux platform.
Now, I asked the maintainer on SourceForge if they wanted to host a potential branch of the project. Otherwise, a new project will need to be setup.
From a time frame perspective, I guess we are all a little bit concerned with that considering the product at stake, the more volunteers we will be, the faster tangible results can be produced. I assume that we all already have a full time job, therefore, no tight project follow up will be performed.
If you feel you want to contribute, feel free to reply to this post or email me directly.
I will post the same message on the SourceForge forum.
From my perspective it would be nice not to fracture the development of OWB related projects into different SourceForge projects. We'd be happy to have any source for other platforms to live in this project.
thanks for the answer. I think it makes perfect sense to host the source file under the same project.
But, there is one thing that need clarification considering your answer on the "Reverse Engineering" thread : How can there be a project for another plateform ? Should we understand that it would be acceptable that a Linux port can create from scratch a Scheduling Algorithm using the same header structure ? That the later would not be considered "Reverse Engineering" ?
Sorry for insisting ...
Creating a new source module from scratch that just used the same header interface would be OK. Using disassembled/decompiled code, directly or as a model, from the existing scheduling module would not be acceptable.
Ok thanks, this is perfectly clear.
I'd like to help and I've been watching this project for a long time but until now it never looked like it had much community support.
If we are going to undertake the task of porting to a new platform and rewriting so much of the C++ layer isn't it worth considering converting the presentation layer to Java? I don't see the benefit of trying to keep a C++/Java hybrid especially since so much of the C++ would need to change.
Just my 2 cents worth.
I understand the problem and would like to help.
Openworkbench is a very powerful application, both for its cumputational libraries in C++ and in its GUI.
The question is not bringing the application to Linux as much as brininging the application to a living community support system.
Let's do it !
Log in to post a comment.