Crosscompile cuneiform to run on Analog Devices Blackfin processors

Asked by daniel on 2009-06-04

Hi all,
I would like to port cuneiform to run on Analog Devices Blackfin processors. In a first step using the uClinux port for the Blackfin and later a standalone version.

To be able to guess the amount of work that would arise may I ask some questions?

Is there already support for crosscompiling?

Does someone have additional documentation (in english) that gives an overall picture of the OCR engine? Description of datastructures?

Is someone interested in such a port too?

Kind regards
Daniel

Question information

Language:
English Edit question
Status:
Solved
For:
Cuneiform for Linux Edit question
Assignee:
No assignee Edit question
Solved by:
Yury V. Zaytsev
Solved:
2009-06-04
Last query:
2009-06-04
Last reply:
2009-06-04
Best Yury V. Zaytsev (zyv) said : #1

Hello!

1) As far as I know, there is currently no support for the cross-compiling in the build system right away, BUT it's CMake based and currently compiles under virtually every major OS (Linux, OS X, FreeBSD, Windows) with both GCC and MS VC++ (on Windows). Also it has been recently ported to the x86_64 systems, so I guess the code is already is portable enough. There is a page about cross-compiling in the CMake's wiki:

http://www.vtk.org/Wiki/CMake_Cross_Compiling

I think that if you're familiar enough with cross-compiling, the current build system will make it easy for you to add the cross-compiling support.

2) We have to admit that the documentation is really a weak point of this project. It turned out, that Cognitive didn't have a comprehensive OCR kernel & UI layer documentation and strict file format specifications. The knowledge was disseminated between the programmers and many of them doesn't work on this project anymore. So essentially they have to write it from scratch. The problem is that given the current financial situation they can't just contribute some of the devs fulltime to do the job.

So at the moment, the only documentation you will find is the code itself (which is somehow commented, but the comments are mostly in Russian, we can try to spare some time translate chunks of them for you). Also, feel free to ask on the mailing list if you have questions; Cognitive reps promised, that if there are some specific questions on how the thing work they will do their best to answer them. ATM they are working on the description of the binary file formats used in the kernel.

3) Not that I know of. I am not a dev myself, just trying to contribute what I can.

Thanks Yury V. Zaytsev, that solved my question.