 I'm going to describe a few of the extensions that we did in the OpenAir Connect project. So to support research objects we already had a rich selection of different types that you could upload and for some of them they already implemented the concept of subtypes. So for instance publications where the types of different articles and other subtypes are pretty well standardised across many communities. We already have that in there. For instance also in images here where you're selecting an image and you're uploading something even you're presented with the different image subtypes that are generally common across all the different communities. So what we did for OpenAir Connect was extend this to other products like how we said specifically the software and the other category which is the catch-all for all other research objects. So to do that, well before I get there, once you've uploaded it in the search interface you also see these subtypes. You can see a small plus next to the type and if you expand on that when searching or when doing faceted searching you can see specific products, types and subtypes. So extending that in the Connect to each of the specific communities You now see in the upload form and as I said we've put this in production already in the upload form if you select a specific community to upload to then when you select one of the specific types here software then you get an OpenAir specific subtype list that drops down. So the community has defined what are the subtypes it wants to use for this specific community represented and you can select already on the upload and this will also be displayed in the search capacity. So that's for software, for other this is the catch-all, it's much more complicated and each of the communities here is defining a rich set of different subtype products and it's both there. The other thing that we've been developing in parallel is the statistics recording the gathering of statistics and the adding of statistics to the records pages here and I think it's been mentioned already the problems that we have are collecting the right statistics grouped by the right community by the right versions and things like that so working within OpenAir Connect we're getting experience with how we should be grouping this stuff together so already in releasing the first version of the statistics we started to group by versions of the records so you can see the downloads and views on all versions this particular version of the record but we have to extend that to make it more flexible what information has up to the dashboard and you can select more fine-brained what you're exposing. We also developed and released during this work schedule the versioning and as I said this was linked to how we do the reporting so you can see each of the versions and in the downloads and the views as I said you can now see the statistics for the versioning but all of these concepts are important to expose to make open science work but the more and more we expose these the more that we're running into the problems of how to handle them in the truly open science world and I'll just give you an illustrative example here so very recently last month the Astro Community Astro Physics Community released a paper which tried to summarize the Astro Pi project which is the Python code base for analyzing astrophysical objects it's a very complex rich code base and in their reference section here you see 50 or so different products so you have references which is very good to DUI so 23 references to Zinodo objects there where software has been stored in there but some of them are to specific versions some of them are to the concepts and basically none of this is reflected back on the record pages of those things so the interlinking the panel was talking about is very rudimentary at the moment and it's not a common black practice so the fact that we're now giving the building blocks means that a whole new code of conduct code of practice has to happen in the research communities but we're still, if you look here, specifically at one of the very popular, most used packages it's not given as a link to a DUI it's not given as a link to the software itself it's given as a link to a 2007 publication that described it and that publication predated all of the other publications all of the other facilities that we've just now made available which means that if you go to that publication in fact it does have a DUI but that was never used in the paper which unfortunately is another citation on there loss but if you go to the paper of course it doesn't refer to the software but we're now making it available within the open science framework so if you go to the software which is interlinking and does have a DUI there's no link this way there's no link that way and as I said there was no link this way so we're making building blocks available but it's far from common practice to use them correctly even in a community that is very advanced is talking about it's open software putting orchids on everything putting links wherever they can it's still not perfect so finally a few other developments that we're still working on behind the scenes are we coming out within the open air framework open air events and open air connect so specifically the the scholar explorer scholar explorer that Harold was describing in other presentations this is going to be gathering all of those missing links and we're going to be putting them back into things like Solodo which can then expose them out through the communities so this will be automatically completing the loop for at least the stuff that we can mine and discover and we'll be releasing new versions of this statistics where we'll be aggregating our communities as well and then that will be fed back up through and we'll be adding some more community extensions we did a lot of work in aligning the concepts between dashboards what they needed to expose on the dashboard and what we have in Solodo and belief but we'll be giving functionality so more people can manage the same communities within Solodo and within a community you can define sub-communities so you can already do some aggregation of the products at the visitor level as well so that's in a pipeline