I know it sounds wacky at first, and may even provoke garbage truck collision analogies, but with Borland moving out of IDEs in favor of ALM and SOA ecology support and Sun Microsystems puckering up with wet lips to corporate developers, why not solve several problems at once?
If Sun were to buy/barter for Borland's IDEs at, I imagine, a fairly attractive price point, they would instantly advance in their desire to build and cultivate deep relationships with developers. Sun could show a path to their Creator and Studio tool offerings, segue the legacy crowd from C and C++ to Java on Sun's terms, better keep Java on (Sun's) track, continue to grow NetBeans (per chance to dream), and even -- this is the clincher -- gain an Eclipse presence that it really needs (while saving face on NetBeans longterm).
Sun already has designs on converting, in the field, the Borland tools shops to use the Sun/NetBeans tools. That will be very rough hoeing, friends. Microsoft and Eclipse will be massive bodies of gravitational pull to those Borland shops. Sun would do better to acquire, admire, and assimilate than try and grind down each developer site with ROI from JES pledges.
Sun already knows that tools are a loss-leader to the sale of middleware, infrastructure, platforms, and hardware. That's why they give the tools away, right? So what's a little more loss for a lot more leader upside? That's what the Borland IDEs could do for Sun.
What's more, by buying Borland's IDEs, Sun would make Borland sans tools a key partner for the ALM and testing solutions that they lack and cannot afford to buy or develop. It would drive a deeper wedge between Borland and IBM. Even more, with control over Borland's Windows-focused tools, Sun could make more hay from its lackluster Microsoft relationship (or else), and perhaps meaningfully reduce the pain for those managing both Java and .NET/VS applications under the same roof. That should sell.
So if Sun wants to steal the thunder from next week's EclipseCon love-fest right in its Santa Clara backyard, they will make some agreements over the weekend and announce an intention to take over (nay, rescue) Borland's tools, promise years of ongoing support, provide the IT shops imbued with Borland's IDEs a compelling path for Sun deployment trajectories, embrace Eclipse support already undertaken by Borland (while still talking-up NetBeans), and drive more JES into the major accounts -- all in one fell swoop.