This week, PeopleSoft has been schmoozing financial analysts and commentators from around the world.

Top priority has been given to convincing everybody that its purchase of JD Edwards will bring in wads of new business, and help the combined firms to slash operating costs.

As for the 'O' word, PeopleSoft's stance is that Oracle's takeover plans are dead, and the company has moved on.

Meanwhile, Oracle is readying itself to launch the latest version of its database and application server technology. In San Francisco, Oracle is set to layout its vision for grid computing.

This is no minor release, but don't expect Oracle to go quiet on PeopleSoft: it has set time aside to keep banging that particular drum.

All of which leaves me with some unanswered questions. Firstly, PeopleSoft might act nonchalantly about Oracle, but it seems spooked. It has already reintroduced guarantees designed to assuage customer fears about a takeover.

And it's got good reason to worry: the Oracle offer forced PeopleSoft to rush through its JD Edwards bid, handing over a greater proportion of cash than it initially intended.

So can it deliver the greater cost savings it has now promised?

I've never been entirely persuaded of the logic of the JD Edwards acquisition. Sure, it makes PeopleSoft number two in the business software market. There was never much overlap of customers, so its install base has bloomed.

But PeopleSoft specialised in selling software to large services firms, JD Edwards to mid-sized manufacturers. I'm not convinced that you can just bundle the companies together and produce one company good at selling to the entire market.

While Oracle's offer to its shareholders still stands, PeopleSoft needs to keep producing good results and deliver on the promises it has made. Failure to do so will make Oracle's offer begin to look increasingly attractive.

But Oracle's position is hardly any clearer. If it clears the legal obstacles in its way, and persuades enough PeopleSoft shareholders to back the offer, it will be faced with enormous development issues.

When it revamped its client-server database to incorporate internet technology, Oracle followed up with a revamped version of its business applications.

Were it to do the same for its next-generation software, then a grid computing version of its business applications would be on the horizon. But that takes effort. What, then, will happen to PeopleSoft?

Oracle has promised PeopleSoft customers a pain-free migration. This is no mean feat and, if Oracle is really intending to make it a cost-neutral move, it will need to invest significantly in developing tools to make this happen.

Whichever company wins out in this ongoing bun-fight, we can expect the victor to emerge with more than a few battle scars.