With regards to specific directions, I think when it comes to curating/maintaining data sets etc, that we should keep in mind that the internet doesn’t tend to encourage/allow single sited data/information. Melbourne University may well ‘house’ certain data sets but that doesn’t necessarily preclude, or make less valid/important, there being available elsewhere. The institution’s status and integrity may be the deciding factor in whether people choose to access it here or elsewhere. (This article may be of interest for understading the values of the network economy http://www.edge.org/3rd_culture/kelly08/kelly08_index.html
)
In regards to the university ensuring the quality of papers given open access, I think quality is better policed at a departmental level, through engagement with researchers and admissions/appointments. I think a centralised system could turn into a bureaucratic nightmare. That said, I think the quality is important, though high profile successes will not be undervalued by a few stray sub-standard papers.