Wednesday, April 10, 2013

Facilitating the discovery and uptake of software: EMI, DCore, SciencePad

Beyond European Middleware Initiative (EMI) was the subject of a popular EGI session yesterday afternoon. Robert Harakaly spoke about some of the developments with DCore data and knowledge platforms and software, which are rooted in EMI technologies. One solution addresses the need for match-making for scientists with similar issues building an automatic ontology collaboration platforms. Robert is also currently piloting a platform for internal secure enterprise storage system at his own campus at Safarik University Slovakia. There has also been interest outside the academic world from the banking sector in Switzerland, and a new technology project is in preparation examining security in medical data management again also based on EMI services. 

Another project from EMI, addresses another important issue. How do developers really know who is using their software? EMI is very good at developing software but perhaps not as good at advertising or finding out how their middleware has been used. This was a previous concern of EMI Project Director, Alberto Di Meglio.  Last year, EMI set up SciencePad (formerly known as ScienceSoft). This project will try to solve issues in identifying, evaluating and leveraging existing software through re-use.  "It's about following the value chain so the impact of software can be made more apparent,"said Alberto. SciencePad is currently preparing for the operational phase. iSGTW reported on their last workshop on 'Persistent IDs for software workshop' in January. Their next workshop will be on software registries and metadata later this month.  

It’s not only middleware that has that problem with gaining wider recognition and keeping track of its impact. Software is no longer easy to define, says Neil Hong. Neil is the Director for the Software Sustainability Institute (SSI) in the UK. What is required is a simple way of  recognising and monitoring use of a piece of research software.  Neil pointed out that a simple presentation can have a DOI, ORCHID and fig-share, it is more complicated for software because of four considerable problems of "Boundary, Granularity, Versioning and Authorship". 

There were several take-home messages from Neil's talk including the SSI five star of research software. Check out Neil's recent blog for more information. Neil also mentioned the SSI Journal of Open Research Software, which is helping with reusability for software.  SSI also has a top tips page for scientific software. Neil recommended the extensive work of ImpactStory which is developing a framework for measuring impact in non traditional ways by directly monitoring software citations. One day forking a github repository would be as recognised as formally publishing a paper.

Morris Reidel, who chaired the session emphasised the importance of "e-science application enabling".  Information is dense in specific communities, but luckily there is now more coordinated effort to find out what’s happening in the other infrastructure currently.  Towards the end of the session we heard from a community which is trying hard to facilitate software discovery in their domain. The biomedical field is complex covering genomes to systems to everything in between. Julie McMurry from BioMedBridges (a joint effort of ten biomedical sciences research infrastructures on the ESFRI roadmap) introduced a project she is working on building a software/ tools registry. They are collating together information on all types of biomedical software  including web services (REST-style), web UI, desktop GUI, grid enabled tools, command –line tools etc. The software can be filtered attributes and use cases. 

Slides from this session can be found on the EGI website, here

No comments: