Using JMS with legacy systems ???....

Hi. If anyone of you guys have any idea I'd appreciate it. What I'd like to know is how you can use JMS to connect to legacy systems, and viceversa.
Does anyone know how to achieve this? I've read somewhere that you can use sockets to send messages with EBCDIC format (

). Most of the legacy systems usually expect messages in a rigid format.
any clues? experiences??

thanks

What happens is you end up using the facilities of the underlying MOM (Message Oriented Middleware) to do the kind of ASCII-EBCDIC translation you're speaking of. I've personally done this kind of work with MQ Series and Java (but not with the JMS layer that sits on top of the raw Java-MQSeries interface, but I understand that this is possible).
Kyle

Hi Kyle. Thanks for your response.

I've personally done this kind of work with MQ Series and Java (but not with the JMS layer that sits on top of the raw Java-MQSeries interface, but I understand that this is possible).

hhmm... i'm still a bit lost

. What I understood is that this ASCII-EBCDIC is a valid option to connect to legacy systems. Then, I will have to create a consumer that will receive the messages, format them appropiately (using EBCDIC) so the legacy system can understand it and send them using sockets. Am I right?
if you have any resources of information I can look at it'd be great.
thanks again
cheers

You may find that there are instances where this is not an issue. I may be wrong here, since because the only legacy apps I've had to link to were PC-based, but IBM has been big on Java for CICS. If that were your entry point, you could theoretically write a JMS client in mainframe Java (which should be Unicode-based internally), and if the result were to be placed in a database or text file, the code translation is supposedly done by the JDBC or text stream services automatically.
On the other hand, if you are writing a socket client in C, for example, you can easily do the translation yourself - on the System/360++ platforms, a single machine-language instruction handles that. Depending on whether or not IBM has enhanced the instruction set for the modern-day problem of going to and from double-byte character sets, you might need a minor additional amount of code of the "for ( int i = 0; i < nchars; i++ ) out[i] = in[i+i+1];" stripe.