We have a server task running on Domino Server 6.0.3. This server task was coded and compiled using C++ API 2.01 and is still running. We are thinking of recompiling it using C++ API 3.0, which is the latest release. We are having some database corruption issues. Lotus has recommended to upgrade to latest release of the API.
We are also thinking about re-writting it in Java and run servelets on domino instead.
This server task manipulates notes rich text fields, compares rich text fields for differences and highlights them. It removes different styles like italics, bold, colours etc. It also read a number of response document compares them and create a summary table in the parent document. Processes about 10 documents per minute.
My questions are:
Is it possible to use new R6 LS/Java classes to accomplish this? RichTextNavigator class is only available in Release 6.5 and we are running 6.0.3 on server.
Re-writting the C++ code in Java - would it be slower using Domino Servelets? If we decide to migrate the application to J2EE, is there any benefit in re-writing in Java?
Are there any issues with C++ API 3.0? If we decide to re-compile in C++ API 3.0, are there many functions need to be coded again?
I’m sorry, because maybe I am misunderstanding, but that sounds fairly slow. I have done a number of projects with our Midas Rich Text LSX which sound like they do the same level of manipulation of rich text and responses and all, and the speed has been more on the lines of 10 documents a second.
Again, I don’t know what you are doing, but manipulating attributes, rich text comparison and building tables shouldn’t be that slow with the API’s. I know we are using the C API and you are using the C++ API, but I still think I would question the design some. This may be a bit off-topic, since I don’t know all the answers to your questions, but I couldn’t help but jump in.
As for some specific answers, the LotusScript/Java classes are probably not quite robust enough for what you want. I would normally say that Java would be slower than C++, but it is hard to imagine it much slower than that. Finally, my guess is that the re-compile would work fine, but why not just try it. If it re-compiles fine, you shouldn’t run into many problems.
Oh, and if you want to try either our Midas Rich Text LSX or Midas Rich Text C++ API, I’d be happy to help you see if you could create a faster solution. If not, I’d be happy to take a look and see if there is something I could suggest to speed up your current code.