Re: Rules Engine [OT]

> They basically say that application integration is a
> myth. Since, this is somewhat extreme statement, I would relax it to:
> application integration products are myth. The ones that I saw are simply making
> too much assumtions, burden integration engineer with ridiculous programming
> model, and just fail to work.

THANK YOU !!!! Just wasted 3 hours of my life today having to cool down a little brawl that broke out over an EAI project ..

I especially like the story of the integration server that allows you to run 2PC transactions over an array of systems for legacy integration <ROTFL> anybody ever seen an assembler legacy that conforms to the X/OPEN transaction model just as an example :-)

> They ignore the fact that the effort put into
> coding the integration bridge in their tool plus writing adapter code is bigger
> than just writing the bridge code without any framework.

85% of the effort is redocumenting legacy interfaces anyway - garbage offered by the legacy means garbage at the integrated interface - unless you reengineer and normalize it ..

> In short, hire knowledgeable consultants, buy no tool, make sure the you have
> connectivity drivers to the applications, design the bridge yourself, write
> end-to-end translation code in a popular language.

what about hiring clever in-house people .. :-)

Anyway - you made my evening ...

Wolfgang
http://www.objectarchitects.de/ Received on Wed Aug 22 2001 - 22:59:29 CEST