I attended a dMIL-sponsored session on web resources presented by Cathy Hartman and Kathleen Murray, from the University of North Texas. They have been working on a National Digital Information Infrastructure and Preservation Program (NDIIPP) sponsored research project called "The Web-at-Risk", and presented some of their findings so far.
The main focus was capturing and preserving web-only information for the long term (i.e. the Cyber Cemetary, a digital repository of dead government division web sites). The differences between traditional and electronic processes were discussed (collection development, selection, acquisition, evaluation, presentation, maintenance, and deselection amongst others). The "unit of selection" is a big tool for them; defining where a digital resource begins and ends can be a challenge. Version control is also a huge issue.
Another thing facing us all is storage: not just space (a digital repository can quickly and easily run into terabytes of information) but usability. Information needs to be kept in a usable way, not just for now but for the future (we all have seen digital files that are inaccessible because the software is no longer around!).
The speakers discussed their findings to date on collection development and selection and capture issues, noting that the biggest hurdle is staff - not just in expertise, but in available time. (Curators need to work with IT staff, for example, and they don't speak the same languages.) Hardware and software issues are also problematic. As they progress, they are trying to find a middle ground of stewardship to deal with these problems.
More and future information can be found at their web site: web2.unt.edu/webatrisk/