NASA launched Hubble in 1990, so it was designed and built with mid-1980s technology. There was a strong case against repairing the Hubble rather than building a new one from scratch and launching the Hubble 2 to replace it.
I remember that in a previous upgrade they put in a new motherboard for one of the computers that had a 33 mhz Intel 486 cpu to replace the 16 mhz 386. I am guessing that the Hubble upgrade mission might bump that up a bit and put a whole lot more memory on the motherboard.
Technology, including computer speed and capacity, digital imaging arrays and communications have come a long way since 1985.
There is a strong emotional reaction to letting the Hubble telescope burn up in the atmosphere. Hubble has given us some of the greatest space images in the history of astronomy. I suspect that we will pay for that emotional attachment. I think that upgrading will be more expensive than launching a Hubble 2 telescope.
Even if all of the electronics are replaced, new high density CCD arrays replace the old ones, lenses are replaced, and the power system completely upgraded, the best outcome will be that Hubble 2 is built in the old Hubble chassis. We will still have a 1985 basic design and all that we have learned from operating the Hubble telescope will not be incorporated into the basic structure of the telescope.
That being said, I look forward to the new images that the improved 2005 design will bring us starting around 2008. The Universe is a beautiful place and the Hubble has brought some of that beauty home to us.