I'm aware that several people are asking about Gap4's suitability for the newer sequencing technologies. Currently gap4 gets pretty memory hungry and slow when dealing with millions of small fragments, but intrinsically it doesn't have any soon-to-be-reached hard limitations. Also the current (or maybe CVS tree version? Sorry for being vague!) gap4 release does support viewing of 454 and solexa trace data too.
The speed/memory is a significant issue though and so this is my current focus on Gap4 work. I now have some new algorithms which are MUCH faster and more memory efficient. As a standalone test outside of gap I can open a "database" consisting of a 2Mb genome comprising ~2 million solexa reads in 1 second and scroll, insert and delete almost instantly, albeit in text-mode "curses" style. I expect some overhead in moving all of this into the gap4 environment, but hopefully we'll see a new super fast Gap4 sometime. Or maybe Gap5... :-)