Why are wavelengths shorter than visible light neglected by new telescopes?
The diagram below, which I stole from this post by @HDE226868, shows that angular resolution as a function of wavelength suddenly drops by three orders of magnitudes from visible to UV-light. The resolution of wavelengths shorter than what the Very Large Telescope Interferometer or the European Extremely Large Telescope detect, in the near UV, suddenly cuts off to a factor of a thousand.
This is obviously because of the properties of Earth's atmosphere. But major space telescopes like the JWST and WFIRST will fill in the far infrared gap. Why aren't there any as ambitious space telescopes planned for UV and shorter wavelengths? (Or is the sudden cut off in that diagram misleading?)
Is it because it is more difficult, even from in-space observatories, or is it because the angular resolution of UV and shorter wavelengths are of lesser scientific value?
There are some technological issues to solve with putting any large telescope into space - and a space telescope is required at UV wavelengths. It is not possible to optimise such an instrument to work at both UV and IR wavelengths because of issues like cooling, mirror coatings and such-like. The simple angular resolution limit of a telescope goes $\lambda/D$, so on the face of it, to get equivalent resolution to an optical telescope, a UV telescope can be smaller. However, you also have to have optics that are good to a small fraction of a wavelength, so much better than the visible/IR. At even shorter wavelengths then conventional "optics" doesn't work because photons get absorbed and you move to the grazing incidence technologies of X-ray telescopes, which is a whole different game and much harder to achieve a given angular resolution.
Given all that, back in the 80s/90s I would guess that a decision was taken about the wavelength range to be covered by the successor to HST (ie JWST at a cost of approximately 10 billion USD) The real reason that no major UV successor to HST or IUE is ready to go now is simply that it is considered that the most important science priorities are achievable at near and mid-IR wavelengths. These are: observing the high redshift universe (essentially no UV light is detected from galaxies beyond a redshift of 3), observing star and planet formation (mostly in dusty environments where UV light cannot emerge and protoplanetary discs emit mostly at IR wavelengths) and doing exoplanetary science (planets are cooler than stars and emit mostly in the IR).
Thus, I don't think there are any technological showstoppers to a big UV telescope (at least the equivalent of JWST), it just comes down to science prorities.
Interferometry is harder at shorter wavelengths, especially in space, is that a factor? And is there an interferometry frontier being pushed towards UV by technological development?
You're correct in that the sharp dropoff is simply because there are very few planned major telescopes operating in the UV range, whereas there are a substantial number planned in the infrared range. As I mentioned in my answer you linked to, CHARA and the EELT, two of the top planned infrared/visible projects, will use new adaptive optics technology, making them far superior to previous telescopes - even though they're ground-based.
Obviously, UV telescopes cannot be ground-based, because Earth's atmosphere blocks a substantial amount of UV radiation. Therefore, any substantial improvement in ultraviolet astronomy will require a new space-based mission. The problem is that estimates for even modest increases require much larger mirrors. Proponents of the Advanced Technology Large-Aperture Space Telescope (ATLAST) proposals say that an 8-meter telescope, at the least, is needed to get good results at 0.11- to 2.5- $\mu$m wavelengths. That's much larger than HST or JWST - and ATLAST could grow to 16 meters!
If ATLAST or a similar project is pursued, the angular resolution at UV wavelengths could be on the order of 0.1 arcseconds or, hopefully, lower. That would match and then beat Hubble. But early estimates put the cost at $4.5 billion for the 8-m version, and Hubble and other space-based telescopes have been famously hurt by unforeseen cost increases. Smaller strides may be needed before we can get to 8 meters, and certainly before we can get anywhere near 16. That's going to take a while, probably a decade or more from now.
But an IR telescope needs cumbersome super cooling in order to observe anything but itself. I suppose a UV telescope wouldn't need that as much. The short wavelength space telescopes thus far have had rather small apertures. Chandra is about one meter, right? Wouldn't a Hubble sized UV telescope work? *(And I find it amazing that any ONE THING could cost $4.5 billion to make. It is like in a Saga of what the goblins in their caves demand from the emperor in order to create his crown of gold and diamonds stolen from the Devil. "Mirrors mirrors in the skies, tell us what the darkness hides!")*
@LocalFluff That's not so much the case any more -- especially in orbit with a good solar shield. And both building and aligning a UV 'scope's optics is very difficult simply because the wavlength is 2 to 5x shorter than for near-IR, and surfaces have to be smooth to a fraction of lambda.