If you're a technology firm and you're acquired by another company, you must have been doing something right. In BEA's case, an offer to be acquired by Oracle suggests the company might have been doing things a little too right.

The database giant's offer on Friday was something close Oracle watchers have been expecting for years, even more than its decision to buy up Siebel Systems. I was in a room with Oracle chief executive Larry Ellison around the time 10g came out, when a reporter asked him about BEA point blank. He said Oracle wasn't buying BEA, but his "no" went on long enough that you knew he had been thinking about it a lot. He suggested that the company valued itself too highly. Based on BEA's response Friday -- a letter calling Oracle's offer too low -- it still does.

Despite the initial back-and-forth, there's a good chance this deal will go through. Despite its success in the market, a stand-alone application server/middleware company isn't going to have a long shelf life. At least not with companies like IBM (and Oracle) offering those things and so much more. JBoss learned that lesson, and sooner or later BEA is going to learn it to. In this case, there's also none of the grudges that existed between Larry Ellison and former PeopleSoft chief executive Craig Conway, a drama that reached near Shakespearian proportions before the whole thing wrapped up. BEA will probably do a deal, but like any vendor it's holding out for as much as it can get. Oracle will not easily let its reputation as the industry"consolidator" be tarnished easily, so if BEA holds out you can expect Ellison (or more likely his deputy Chuck Phillips) to do everything short of picketing the HQ until they give in.

What Oracle really gains through BEA is a little harder to discern. This is a company that has spent the last three years -- basically ever since the PeopleSoft deal -- talking about how it's trying to bring together its various applications suites, a project it calls Fusion. The big highlight of that project so far has been its Fusion middleware. If it's so great, though,why would Oracle need BEA's middleware? Of course, Oracle was perfectly happy to buy out PeopleSoft in order to simply eliminate a competitor, but in this case BEA may offer more.

Besides its expertise in service oriented architectures -- it bought a popular repository product which is being used by Canadian firms -- BEA has also been far more focused on the Web 2.0 market than Oracle has. Perhaps because it's smaller BEA can be more nimble, but its expertise in helping customers make the most of their SOA may be what Oracle needs to sell its own middleware, among other products. What's interesting is that the day before Oracle's offer was made, a BEA customer told me he was bugging BEA to support databases other than Oracle's. That scenario seems unlikely if this deal goes through, but if Oracle is really serious about taking chances, interoperability may be the biggest one of all.