 Well, good afternoon and for some good morning everybody welcome to this this webinar on in unpacking interoperability So my name is Keith Russell I'm really interested in this discussion around what is interoperability mean in practice So my name is Keith Russell manager of engagements at the Australian Research Data Commons last year we started a series of work around fair and what does fair mean and As you went through each of the topics interoperability definitely was one of the more complex Issues and that was also out of the survey survey we conducted We also noticed that it was viewed by a lot of people as quite a complicated topic So we thought we'd put a little bit of an extra effort into that into unpacking What does he actually mean? What does interoperability mean and what levels are there around interoperabilities in ways of thinking about that and For that we have two two guest speakers today We have Bruce Simons and we have Jonathan you I'll present I'll introduce both of them in a second So first of all I'll just provide a little bit of background a little bit of context To key it off and then then I'll hand over first to Bruce. So first of all Just a bit of context around interoperability and from the perspective of the fair data principles So if you look at the fair data principles the third third letter I is around interoperability and There they've listed three principles Around interoperability the first one is that meta data and data use a formal Accessibles shared and broadly applicable language for knowledge Representation so we've had a lot of questions. What does that really mean and what does that look like and Things you can think about and there are sort of using controlled vocabularies ontologies, etc. And also using a sort of a common a good data model Yes, so well-defined framework to describe and structure data and metadata and today We'll hear a bit here quite a bit more about what that means in practice from from Bruce and Bruce and Jonathan Another aspect to that second principle there is that meta data and data use vocabularies that follow the fair principles So that the met the vocabularies themselves are also made fair findable accessible interoperable reusable So that means that the vocabularies need to be documented and resolvable using a globally unique and positive position Identifier so you can point off to the vocabulary and know that you actually have the right version of the vocabulary and The last principle they have there is meta data and data Should include qualified references to other related pieces of metadata and data So just having a reference off to another data set and saying this data set is related to that data set is It's a starting point, but it's not a qualified reference so actually having a bit more context around what does that relationship to that other data set or Related metadata look like adds value. So that it is part of or derived from or a subset of That actually is a much richer reference between the two pieces of two elements so one of the challenges in this space if you look at interoperability is is having standards and using standards and Referring to those so if you are looking for standards and looking for places where you can find standards in specific Disciplines or specific areas one of the places you can go to is fair sharing.org It's a global So website portal that collects all sorts of standards databases policies guidelines, etc And I think for the discussion today, especially the pages around standards are useful It's got more than 1200 standards in there at present and you can actually add your own and deposit your own standard and add it to that list And point off to those the standard there That's one place where you might be able to deposit your standards. There are other places, too and If you are looking at Setting either using a vocabulary or setting up your own vocabulary building one Have a look at research vocabularies Australia It's one of the places you could you can use to either find vocabularies or set up your own vocabulary So I hope that this is these are two two pieces two pieces of the puzzle that might be useful If you are working yourself on on making your data interoperable So that was all the slides from me at this point So now I'd like to hand over to Bruce And just hear his perspective. So Bruce is a research associate at Federation University's Surrey the Center for e-research and digital innovation and Bruce has been involved with designing information management systems and Research into data exchange mechanisms and interoperability Including that geology and groundwater open geospatial consortium data exchange standard So that's sorry the geology. That's a geosci-ml and the groundwater groundwater ml Open geospatial consortium that most people just refer to as OGC OGC data exchange standards. So Bruce, I'd like to hand over to you And if you'd like to give us a bit of an overview of the work you've been doing there, that'd be great That's okay, so it's pretty much Said everything I need to say. I don't think I need to say much here. It's been a wonderful introduction there What I'd like to do is just give a bit of background on why interoperability What means and then perhaps some of the challenges for for implementing it So, you know, I saw it from a long time ago, you know turn of the century stuff There's lots of problems and you know trying to standardize data across organizations And there are very good reasons why we don't have standardized data across different organizations But at the end of the day, you know, this makes it difficult for the client that the user and I guess What interoperability is a bad is about looking at your data from a user perspective and that's A change that most organizations perhaps don't think about So one of the challenges What we'll do is we'll just make it we'll have a standard database. We just want the same database and then we all Everything will be up be hunky-dory Well, back in the 90s The North Americans spent a lot of effort 10 years of coming up with a standard way of storing geological data I was mandated by the USGS that anyone would do it So I think we needed four of those Databases Australia, we we we took them model and implemented as well And by the end of those five different implementations, none of them could talk to each other Basically the variations that each organization placed on top of that made it That there was no Interoperability between the data mates. There's no standardization between the data mates So silent cops from CSIRO came along and said look, let's stop worrying about the persistence land Don't worry about your database instead Establish an exchange protocol and I guess that's sort of the start of the interoperability journey that that certainly was for me There's different ways of your making data available and and this is From a paper that that pull box and code put together sometime ago now But yeah, traditionally what would happen would have a whole lot of different data sources Different databases and and we usually would have to find the right databases Work out how to access the data work out extract it interpret it Um, we have a be just last month Thank you Yeah, transform it and I know the effort is on the user side And I'll turn with this is he has a centralized system where You know a single data custodian makes their data available and uses all abstract they don't from that single data source Like you kind of see what this could work for something like, you know The weather data or everyone just gets it from the bureau of me But even in practice that tends not to be the case that that we've got Other viewers from other countries that you might want to get the data from There's no crowdsourcing weather data. You might want to get to so even a centralized data system tends not to not to Be sustainable Another approach and it's been common for for some of the Encrust facilities is to have an intermediary So we'll aggregate data from these various sources and make that available for all the users No, this is a good system that the users only have to worry about going to the aggregator it becomes a bit of a challenge for the With the business model for the aggregator I mean, we're not a provider of data and so they're not a vested interest in maintaining that data More they take necessarily a use of the data. So there's kind of a business model challenge ahead with this particular model A model has been quite successful. So particularly in the groundwater world in Canada was where the National Groundwork organization the national resources of Canada Well, if each of the provinces put together put puts up a service in this case They used a word feature service, but the technology doesn't really matter Then what we'll do is we'll Access those services and then we format it and put put a community application schema over the top of that So that the users only have to access our Services and they'll be able to access all these other Provincial survey services Works fine if you've got a mandate to do that, but it still requires some kind of agreed schema to do it and finally, there's I guess The gold star and clearly therefore the most difficult is where each data provider puts up some kind of service according to Standard that that is all agreed on and and then users doesn't matter where they're getting their data from Accessing it via the same mechanisms in the same format with the same content Ultimately, that's that's sort of the interoperability gold star So what what the heck is your company? Well, there's all sorts of different definitions for it around the place Into I quite happy with Leslie why borns up When she says, you know, look, it's my stuff from when I saw that stuff to that computers programs data Accessory all that other stuff for your stuff, and I don't give a damn where it is how it works or what the format is Keep talking about fair. This is yeah, I guess, you know, we took the interoperability in the early days now fair is is the guiding principle There's governments around the world and moving towards it Research funders require data to be published that way and research journals, you know what data we made available So it kind of adds this nice Words that we can understand that you have to find the accessible and reusable to this horrible interoperable word So what does it mean? Well, you know, you probably part of fair is that? Yeah, the computer can interpret data so that can be automated Automatically combined valid data as Keith sort of said, you know, there's metadata and data used community agreed by tabularies There's links to all the information and that's all done through persistent identifiers And the data is accessible by community agreed formats in community agreed languages and using community agreed vocabularies So Keith's told us about that But we look at the other parts of fair you know the English the bits we can understand the findable accessible reusable We notice that actually there's a lot of interoperability built into those that we need globally unique and persistent identifiers We need to have searchable resources There need to be data and metadata needs to be retrievable identified Identifiers using some standardized protocol. The protocols need to be universal There needs to be machine-readable license of provenance information and domain specific data metadata standards So these are all actually parts of interoperability So what are the requirements? Well, this is a slide from from Berwick and Gigan a long time ago And I talked about database systems, but yes, you can just think of them as data systems Yeah, what does it mean? Well Putting a bit more understanding of this effectively this area. We need to have systems agreed systems We need to have an agreed syntax We need to have agreed data structures the schematics and we need to have agreed data content for interoperability to occur I would add an extra one on top of that and say we need Organizational and probability we need a social commitment for data providers to make their data available via standards for to occur Maybe interoperability in terms of driving analogy like this two systems in the world rather drive on the left Do we drive on the right? But if it doesn't really go in the world, we actually recognize roads and we recognize whether they're farm tracks or four-lane highways And if you recognize vehicles the things that drive on these roads That's kind of the systems that we have in common In terms of vehicle operation, you know the steering wheel may be on the other side But but fundamentally, you know, we get into a car anywhere in the world and it kind of makes sense We know how to operate it Um You can imagine how difficult it would be if for instance the brake and accelerator pedals were the other way around I mean, I mean, there'd be a major barrier to interoperability of driving if that was the case There may be some differences like, you know, whether the differences More perhaps between vehicles within any of these systems rather than between the system so that you know the indicator or the water Controls may be on different sides of the steering column, but that's within each system not not not Between systems and that's kind of like in interoperability languages, you know Like whether you're using xml or jason in in in the syntax that they it doesn't really matter When we move a little bit further up, we see you know that the road wheels And the signs and stop signs give a size tend to be the same we can go to these different parts of the world and and It's kind of All makes sense. We've got some kind of of standards happening there um A little bit further up the chain The driver behavior might be slightly different and that's going to be more challenging that you know that that to really, uh The first treatment of pedestrian crossings or the italian treatment of speed limits or the indians use of Of headlight. So these are kind of local changes that that if you're a stranger coming into it You find it might be find it quite difficult to understand But but ultimately for interoperability to occur. It's that final organisational requirement that basically you Were dependent that every driver wants to get from point a to point b safely and and they will drive according to those standards um What what the issues with interoperability is As we move up this chain it becomes more and more difficult because more and more social Um, the technical aspects at the bottom are far easier to deal with So just looking at some of those working way up that chain Yeah, the the system interoperability. So that's so Leslie would say that's that we are connected via standard protocols Um, it's all about the you know, that that wonderful it world that there's appropriate operating systems Whether they're they're microsoft or apple or android or whatever that will have established network protocols hcps and And hcpses of the world We have web services whether they're west with the web feature services Of the jason and I should say in this particular world. There's lots and lots of acronyms I could Explain these acronyms and they wouldn't make any more sense to to to expand on them um I guess they're just there so that if you actually do recognize any of the acronym it kinds of puts it into a context for you And and we need appropriate governance of all these systems So whether it's the the the web consortium or the open space or consortium or iso or whatever that that these systems are Undovernment appropriate and we can therefore rely on them Moving up the chain There's syntactic interoperability. So the data languages Leslie my machine talks the same gibberish as yours This is about the packaging and transmission mechanisms. So whether the languages are xml gml jason Whether the format to shape files or g adjacent or let's cdf What there is also identifiers whether there's the htp uris or or urls or something we would call them urin's doris handles, whatever Again important. There's appropriate whatever it is whichever one of these ones we're using there's appropriate governance for it So these systems and syntax interoperability we've been kind of thing as foundational They're the building blocks that we need for the data exchange between systems However, they aren't sufficient They don't tell us how the receiving system can interpret the data without further In the past mostly human intervention So the schematic in club million, this is the area I guess that that I've mostly been involved in This is where the ml's come in It's an urgent term basically we're agreeing on what what the attribute names and types are for the kind of data In the past we've taken some kind of conceptual model This has been done in ml, but Do it on whiteboards. Doesn't matter how it's done Moving across the some logical physical model In in database talk that's kind of a An er diagram if you'd like or you know, it's a structure of how you're structuring your data And then the final one here. We've actually moved it across to what's called a physical model This is a real world implementation of something a machine can understand that's obviously impoverable xml machine language So one of the issues with the schema though is Is a question that the community has to think about how widely understood they want their data to be So we can easily exchange data with communities we've been doing a lot of time And where the community understands what we're talking about we could just send them CSV file for instance we can say, you know, here's a column called temp And here's a few numbers and and here's some more numbers then it will make sense to someone we hope Once we want it to be used in a wider communities, we need more precise definitions To explain the complexity is the real world. So here that previous example never say well, the observed property is temperature And importantly when we say not even that it's just temperature, but we're going to use a a link to A meaning for temperature So that's the content and I'll talk about that a bit later on the feature of interest is a is a platypus The procedure is some thermometer process the time and here we've got it in an iso time Format here's a result where the minimum value is this and the unit of measure is this and the maximum value so you have provided far more information And so agreement at an international level enables data to be used by other domains. I don't need to know anything about um platypuses to to be able to to read what this this information is I can use it straight away another thing is what Is important is that is to be able to reuse established patterns from other domains rather than then coming up with Your own particular set of patterns that's specific for your domain that then can't be used by others And for the majority of particularly science based work Why is this about observations so that a pattern of observation and measurements can be reused? So now I think you probably yeah This is this is the one that We often get Struggle over so in years term Yeah, this is where my concept of lake is exactly the same as your concept of water body And nothing that he is a if you wouldn't resolve this uri It would tell you more about lake water body those kinds of things So rather than the English term lake or things term water body if we use this uri It provides that some anti-contrapunity tells us the data content So computer systems can can exchange that data unambiguously Um There's a question of what kind of control of vocabulary is required is a is a simple term this sufficient Do we want an authority file? taxonomies the source Or or a full-blown on apologies So it's because more and more difficult as we move up that chain to to establish those things What's the value representation we're going to use and what's is it scarce as a double core? I'm just going to use codes Ah Again languages falling to this we that's computer languages do we use rdf your xml l Well, how are we going to represent that this um, you know, Keith mentioned the research vocabularies Australia That's one one way of of accessing these vocabularies. There's sorrow. Have I leaked data registry? The sysmark as a tool that the labels the labels you to access vocabularies services But to me that the big challenge in the semantic world is is the governance aspect um You know units of measure is is a classic example, but this is crucial to all all our observations anything we're doing yet The the governance of units of measure Ontologies is is appalling. You know, there's there's so many out there. There's The ucum and qdt Both nasa ones as others there's rba has a unit of measure that's governed by cheers so much australia It's none of these are complete That some of them have got Not very transparent governance processes How do you you know, how do you add to it? Have you had you manage it? So to me There's a there's a real challenge in this this governance and and and setting up of of vocabularies The main specific ones tend to be pretty good. You know the ugly rock one run by the the fao Um, the keby the chemistry entities for biological investigation. I think um, the cgi do simul geology ones and you know Sorry, I've got a west came out. So when we get to the main ones, it's not so bad But but across the main particularly a particularly challenging And I think governance is a is a real challenge for the semantic drop the lead Finally up the organization drop the lead and they are literally says look this says I will share my services I I'll share my data with you And it covers, you know, the policy the social the organizational and and this is really difficult I mean, what's the value proposition? So what's that one of the benefits of Carson and We talked about that earlier on with with from a research point of view where Um journals are requesting data to be provided made available through fair We've got governments who are sort of adopting open policies That sort of says that we need to start looking at at how users are going to access your data Um, but what about private data? I mean, that's a real challenge to be getting getting a lot of private data available And and moving into the agricultural field. This is this is a particular big challenge There's a person in funding. I mean, how do you how do we we we convince organizations to put the resources into it Establishing agreements what what's the world's responsibilities and you know the stewardship of of the whole community And and finally a lot of risks and responsibilities of participating in any community that that agrees on on interoperability So, yeah, I guess I will be probably requirements are fundamentally a desire to participate Having decided you will participate that a commitment to use whatever the community agreed standards are They need to cover common data content. They need to cover common data structure uh, they need to be available with appropriate technologies and that's mostly the case now and Don't want to be knocking organizations into into particular types of software or or software vendors So that's not sustainable into the future So I guess some of the challenges and there's lots and lots of challenges There's the organization drivers. I see as a real Real big one. You know, we've been doing this for a long time And it's really really challenging to get organizations to change their their behavior To look at things from a user data point of view Having done that, you know, one of the alignment processes and policies within organizations and across organizations Security is becoming more and more of an issue. One of the services that I've dealt with of course, we're all open Um, that presents problems for for access to the services and potential service attacks um, I think Fundamentally the governance of standards and online resources They require all levels of interoperability, but but the vocabaries up are particularly fraught A lot of data standards are immature There's challenges in terms of of different modeling practices The standards are complex, but you know size domains are really complex domains and that leads to standards that are both difficult to develop And from a developer's point of view really difficult to implement So we keep getting this pushback and gate standards to try and simplify things Um standards keep changing Uh, you know persistent uri's is a good example that that when I started urn's were all ago Then then htp uri's become URLs and and then maybe do i's and handles um And an xml and jason is a good example that xml is really good because we can have an xml schema and that can Check to see that what you're delivering validates whereas jason's nice and simple so jason's Developers like that, but of course now jason they're going well actually that's a bit too flexible Maybe we need a jason schema to be able to check to see if the jason validates um We're cutting edge of some of these solutions. I'm wfs 3. I haven't played around with that. That's coming out Finding a linking services. So if I've got content, which is coming from different different services How do I embed 85? I've got a I A web service which requires one kind of format and now I'm getting a vocabulary service coming into there Which is a different format. How do we how do we link all those together? How do I run filter queries across some of these things? And and we've talked about these protocols and so I can say look is here going to be and a A gsiml protocol that's going to be in gml and it's going to be an xml But if I want a Australian version of gsiml, how do I profile that? How do I how do I know which particular profile that these protocols are coming in? um, and and you know finally that the software that there's lots of challenges in in Developing software that's usable and does what we want it to do Both open source and proprietary that there's challenges there But probably the the biggest challenges is our human skill capabilities. You know, let me think about what we're trying to do here we're moving from Right down the grassroots Right down in the in the weeds of of it world down the systems right up into The domain specific areas of vocabaries up into organizational management areas and you know, who has the skills to be able to Move up and down that that whole whole range of of of domains So that's it for me. I'll hand back to you, Keith Thank you very much bruce that was really interesting in a broad overview of Of the different layers of interoperability and also the challenges in this in this space so, um, Jonathan Jonathan is a data scientist with the environment environmental informatics group at CSIRO And his particular expertise in information and web architectures data integration data analytics and visualization He currently leads a number of initiatives to develop new approaches and tools for connecting information flows across the environmental domain and the broader digital economy within australian internationally That includes the CSIRO knowledge network, which is enabling greater discovering use of government scientific academic and spatial data So, um, without further ado, I'd like to hand over to uh, to Jonathan to talk about WESC, I think you pronounce it the water and energy supply and consumption data standard and what the the steps you needed in developing that Okay, thanks for that. Um, so I'll just be talking, um About a domain specific example, I think bruce gave a really good overview of Interoperability and some of the issues there. I'll be diving into specific detail about a data standard that we've been developing that bruce was also involved in at CSIRO around, um Enabling interoperability for water energy supply and consumption Research data So this this project, um, was commissioned by the oran So oran is the australian urban research, um infrastructure network And at that time they had a challenge around, um, gathering research data for the water energy supply and consumption, particularly in um In cities and our regional areas And the context of that is that this sort of data is Scattered across different organizations our utilities abs Energy companies and each of them have maintained their own databases around this sort of data So just to give you an example I've got some here So this is data that um when you request data from these organizations, you'll get something like this. So this is Water consumption data They'll tell you the quarter the billing quarter the locality or lga What kind of you know aggregator premises there were how many sites and You know killer leaders of built usage So this is just from one provider You ask another provider they'll give you this format, which is lga and then the different quarters and then some number Which is fine. You just have to ask them what the numbers mean and in this case, uh, I think these are Killer leaders And then you go and ask another provider and they'll give you this sort of format. So instead of having Uh one set of columns you have multiple set of columns going across the sheet um and so on and so forth so The challenge is While the data's all there. How do you actually interpret this in a sensible way and how do you actually? um Do some research over it so What the team did was um given this challenge We defined a information model. So i'm just walking you through the weska mouse site, which documents all of this um, and we've tried to make this um As clear as possible on the website. So if you've got any feedback around um, or any Things that you don't understand, please give us feedback. Um, but what we did in the first instance is to start to conceptualize what um water energy supply and consumption Data should be represented as in a common format And therefore develop this information model Through different um classes of things. So this is just showing the aggregated consumption and the different um properties for it Then we can also drill down into specific consumption And look at meter readings for water gas and electricity And then look at supply side things. So this is a information model for Water energy supply so you can get um represent things that are common across supply data as well as specific for electricity water and um Even down to the specific zone substation electricity supply Um, and that's just the common model and there's a more detailed model But basically this allows us to represent things in a common format a common representation Common semantics. So all those levels that bruce mentioned around syntax schema and semantics Um, and then we can then start to define the semantics of different definitions. So here we've got um a section describing the vocabularies They're maintained at our link data registry here So if you get it's linked off to this side. So these are all the resources defining um, the commodities The energy types the land use types Let's drill down to one specific one around Electricity consumption and so you've got a definite and then you can have you know narrow consumption descriptions and Get into more detail and typically this is where the researcher will come in to Help define that and we'll encode that in this vocabulary We've also got ways to visualize this so they can be more easily accessible by people and understand the hierarchy of things so Using scos as the common representation here we can define All the different representations and nest them and visualize them for quick access Um So in this particular project what we wanted to do was standardize the language which this page represents But also standardize some tooling so we can do some rapid deployment of data infrastructure So if we have a common, um representation of the data, uh, we can build common data services Um leveraging that representation and then deliver that out via web services to researchers third-party applications And in this case, um, we've partnered with the oran platform So our data sets that we've harmonized can be accessed by oran This is their data catalog showing if you search for west You get nine data sets and these are all different providers providing water energy consumption And then in the portal what you can do is pull them in and get the tabular representation And you can see here that they've been This is for yara valley water and it's been Harmonized according to the different fields that we define in our information model So there's no ambiguity around, you know, what specific numbers mean We even give uris for the property types and the units, I believe. Yep units here Um So yeah, you can get the definition of of those units from from the data itself. It's linked to external vocabularies um, and then When this in this portal we can start to visualize that so pull in the data and do some porpoise map visualization, so We can pull in the yara valley water Um by localities as well as the city west water and start to see If there are any patterns of consumption that you can see here We've also been working with another project called nia so nia The nia program is the national energy analytics research program that sarah is working with the department of Energy and environment and environment and energy Commonwealth And that's also trying to do the same thing that Which we we did in the rm project, which was to aggregate data from across different organizations And in this particular case west is being applied to zone substation data and so Again multiple utilities represent their zone substation data in multiple different formats And there's been a lot of work in cleaning that up and providing fine grained Zone substation data using the west and our format So so you can go here and basically download The osgrid zone substation load data And be confident that the fields are described And you can compare that data set with other utilities So, you know, for example osgrid with energex and do that 30 minute load data comparison So I guess That's really in a nutshell the west ml format and how it's being used and applied In these two different projects both in oran and nia um And it's really an example of how we can gain into operability by applying the syntactic Into operability up to the semantic Some of the challenges around this is more on the social side of things. So while we have these technologies and Formalisms for describing the data The challenge is actually getting adoption. I guess Down at the supply side at the data provider side In adopting these sort of standards to deliver the data out These ways and this these representations currently the drivers for this particular data standard come from the The consumption side so as in the researcher side. So researchers or policy analysts who want to do The analysis at you know this scale where you want to drill down into a locality a suburb or a postcode or a sa2 You need the data in a good easy way to do your analysis and visualize for example However for the data providers, there's really You know little incentive for them to provide that out in in a way that you know can be accessed in this manner So projects such as oran and nia really catalysts for those Transformations to happen that the interoperability work But I guess the challenge is how do we get that in a more sustainable way where That sort of interoperability is shared across different actors The other challenge is also On this side around the vocabularies, which I started discussing. So here where we have different terms and different definitions of things Each one of these terms and definitions needs to be curated by somebody Or a community so As bruce mentioned, you know scaling up that the human skill levels That can understand how to formalize these vocabularies in a way that's Discrete but also linked to different definitions and external other vocabularies But still be scientifically Accurate and applicable in this in for research data like this Is is quite a you know a skill to be able to cultivate. So, you know, how can we scale that? across different communities But once we do have those vocabularies, I think, you know We can we can basically reuse them over and over again. So Getting good quality vocabularies In the first place is a challenge, but once we do have them, you know, that helps the interoperability story across different projects and research So I'll just conclude there This was really an applied Presentation on how we use information modeling vocabularies data transformations In this domain of water energy supply and consumption Thanks, Jonathan So, uh, thank you very much That was a really interesting especially as a use case to see what it looks like in practice and how you can work on interoperability in practice I think it was very interesting to hear these different perspectives and the different angles required around making data interoperable And one of the one of the things that sort of just came to mind looking at these presentations and looking at the the way you're tackling this and addressing it is With all the data that's out there Around australia or maybe even more even globally Is it is it actually feasible to make data? everywhere Interoperable is it actually Possible is that something we should aim for? What is the scope you should aim for when you are thinking about making your data interoperable? You have Bruce and Jonathan do you have any thoughts on? How far can we realistically go? Okay, can you hear me? Yes. Yes, I can hear you now. Thanks Yeah, a good question. I mean The the example of the databases from from north america was you know, we can't have one system But everyone uses it's just that's just not possible. So to me it's about minimizing the barriers to interoperability That that You know, if every if you look at the the railway gauge issues, you know, I can say Yes, we can overcome the fact that we've got different gauge railway tracks australia, but it requires Having dual gauge tracks that requires having bogie exchange systems It requires getting people to change trains Platform so so we can get around that but obviously if we can minimize that then that makes it better So so it's about communities going one of the One of the ways that we can minimize those barriers What's to use is using our data now some communities will will go down a Perhaps they're quite a rigorous path and others might say and that'll depend on kind of the the level of governance I guess of that particular community and the importance to be able to exchange data Others might have a more flexible and and I you know I you've got to enter somewhere and I'd already mentioned that that's sort of a early Not so interoperable, but at least you know some level of interoperability and then gradually over time that would move towards a A gold star kind of interoperability. It's kind of a immaturity index sort of mature of interoperability if you like Jonathan Yeah, um, it's it's I don't know if we can gain 100 percent interoperability across all communities, but yeah, as Bruce mentioned lowering the barriers Providing tools where people can more easily share their data in in standardized ways And to a certain extent it's it also is to a large extent. It's a social contract with multiple communities, isn't it like um People come together to agree that we're going to exchange data in this way for these purposes, so Getting that agreement That social commitment in the first place will help the way for that sort of interoperability You'll be it Keith Sorry So is it actually working from that organizational interoperability down to to sort of Well, maybe not down, but at least making sure that The organizations have an agreement around a shared goal that you're working towards Yeah, and the organizations have to have a Be on board um, how you get organizations on board is is quite challenging. Um That in you know, north america was tended to be a a carrot that they said What we'll pay for each organization to to be involved in in in this uh In europe with the inspire program It was kind of a legislative approach that every country will deliver their data according to these standards and you know We'll use a big stick to to make you do it Both of those mechanisms have problems uh The the canadian example where they you know, basically made it easy for each province to put up their data via Whatever step services they had and then federally pulled them together was kind of meant that each of the Data providers could see advantages for them in in that their data was made available along with everyone else's But it wasn't a wasn't a huge barrier for them. So there's kind of Each each community has to look at at ways to do that. Um, you know the the ar I guess to me it's It's kind of like the ar dc saying well from a research point of view These are the benefits that a particular community could have by making their data available So what can we do to to to to assist that and and making You know, so research vocab is Australia's and look here's a place that that people can put their vocab is up quite easily I make them available that that kind of makes it is a low barrier to the to the semantic interoperability Um, if we sort of go well There's two groups in Australia that have gotten the same kind of data if we get them together and so well What's a what's a common structure for your data? That that you're happy with it may not be internationally standard But it's a you know, it works well between you two and that can kind of help get over that bit And and so it's trying to get over those little bits But but it usually requires some some other external driver, whether it's So the ar dc or federal governments or whatever Yeah, I think I think getting buy-in from different stakeholders and organization is key And having the infrastructure to support that collaboration is also a key thing So if you come together and say we're all going to do You know have a standard for you know, this sort of data But don't have any way of sharing don't have any infrastructure don't have any services don't have any Formalisms for describing the vocabularies then that's going to fall over So I think it's both the social and the technical coming together as well Part of the challenge of Is the persistence of these things the users expect to be able to If if they're building applications to access Data from from various data providers and they don't want them falling over or changing on a regular basis so We've got this kind of tension between we want it to be easy for data providers to make their data available But we need it to be standardized, which means it's not going to be easy to make it available and How we find a sweet spot there that That allows persistence through standardization, but doesn't put big barriers into people joining the community Hmm And I guess there there's also challenges if you're talking about using vocabularies or standards The longevity or sustain Sustenance or keeping them going sustaining those standards and sustaining those vocabularies You've done so far if there's maybe already things there that you you are you've seen or Possible solutions in that regard too Yeah, well the more those Standards don't really exist Independently of people using the standards So the more people that use the standards the more I guess risk management That you manage the risk of it being you know dying off, right? So um having them having Standards that are well defined well scoped and easy to use across Jurisdictions across organizations globally You know like the ogc and double 3c I could you know organizations where you do have processes for standards adopt adoption and defining You know having them there having those organizations take care of the publishing and maintenance of those standards It takes it away from a project takes it away from any Individual actor to maintain that standard. I think we had an instance where An important vocabulary was locked up the earth sciences because you know unfortunately somebody passed away and that was the That person was the sole Maintainer of vocabulary So, you know the sweet the sweet vocabularies That's that's a risk in and of itself. So having processes and organizations and users Provides that longevity I think Yes, the sustainability of vocabularies is is I think a real challenge Because we're up in the domain specific area here largely that's but So individual Scientists so this is a vocabulary that I require so I will publish this one And and so we start to get a proliferation of of overlapping vocabularies without any real connection between them because that requires extra work to try and build those those inter connections and then Unless the organizers there's organizations that are prepared to take on board the the ongoing governance of those those particular vocabularies show If I'm going to build a system I don't want it to be dependent on on you know one person maintaining maintaining a vocabulary I want it to be a persistent sustainable and well-managed vocabulary and and how do I judge that how do I know? Apart from as Jonathan says, you know, maybe by use if everyone's using it then maybe you'll persist Yeah Let's go on Jonathan I'm sorry. So I was going to say or Yeah, I'm not just users using yet, but also having the backing of You know organizations You know OTC And other, you know, maybe government agencies as well. Yeah It's actually I asked that question to fair sharing I said, well, can anybody just put their standard into fair sharing? Because that would just mean it becomes a big bucket of all sorts of standards Clarity on how how useful that standard actually is and they said yes Anybody in theory can deposit their standard But it only becomes more interesting when you see that that standard has actually been used in policies or in databases or in other approaches or supported by organizations So if you look at the way fair sharing setup, you can actually see that this is a standard that Doesn't have any ties Or this is a standard that's actually already been adopted by these organizations Is referencing these policies and is implemented in these databases, for example So I thought that was an interesting way of of providing a bit more depth on top of what just the standard itself Yeah, so There's a question for both of you really for the from the perspective of the people on the call and that if you If a researcher comes up to you and says I want to make my date interoperable Where do you start because you've presented a whole array of considerations and array of Perspectives and things you probably need to tackle down the track to fully make it interoperable But if somebody's right at the start of that journey and thinking I should make my date interoperable What would be a good starting point or a good few first steps to to move them along that that So from my perspective, I'd be working down the the interoperable the arrow from from the top You know Just the organization that you represent want to make your data available If you've got a commitment for some resourcing to actually make this happen um That comes kind of questions about This data that you want to make available. Is it actually your dad? Do you have the authority to to make this this data available? uh and then Are there other in your domain in your community? Are there other? practitioners With similar data making theirs available so If so use there, you know Don't don't invent things yourself reuse stuff If if there isn't but there are other practitioners there What's their feeling? Do you know there's there's no point having one telephone? We need two So so, you know, can can you both um work together towards toward Something that you ever can agree on if there isn't one already out there Um, and then you know your final last because okay. Yeah, we're doing this. This is the way These are the standards available or we're going to develop our own standards to do this uh, what technology do you need to implement your organization will have certain rules and practices about the way It works in their in their organization. So how are you going to? um meld what you want to achieve with what the what what your organization Wants to achieve And that sort of goes back right back to the first question again. Okay. Does the organization really want to be a a need to operate interoperational organization Okay, thanks Jonathan Uh, yeah, just I could um bruises Comments around don't reinvent or if you don't have to um And just look at what others are doing. Um, there's a lot of resources available um, you know in the vocabulary space And in this standards the schema space, um, both OTC WVC Um, and others are are formalizing standards in those spaces. So you'd want to be checking them out before Determining that you want a new standard Um, but there's also lots of you know experience with people Developing standards. So, you know, there's there's probably worthwhile talking to people who Have done it before as well Just I just say in terms of the reuse there We all kind of think about domains as being very special and and and unique therefore we need our own standard but if you think about what a lot of Particularly science, but other domains as well. It's about making observations, you know In a broader sense of observations that an interpretation is an observation as well so if you take a generic pattern and observations and measurements is a is a pattern is LISO standards no gc standard that Simon Cox was was main author of You'll find that you can take that pattern and apply it to your domain and it doesn't matter whether it's vegetation It doesn't matter whether it's marine Yeah We're basically we're observing things at certain times for certain people using certain procedures and we get a result And and that pattern can be reused And it and it just makes it so much easier. So okay. Well, I don't have to worry about you know learning how to model things What I need to do is learn learn how to have vocabularies for my procedures and vocabularies for my results and and those kind of things Okay Sorry, I just realized that I Was looking in the wrong place for the questions and I found the question But I'm also aware that we're actually already running over time I think there's a series of questions there some of them around the other slides available Yes, the slides will be available. So that would be great. This has also been recorded So people can watch it back and for those that weren't able to attend. Please pass it on Um, I think one question here that's probably worth having very briefly tackling I think it's an interesting perspective as question is around. Is there a role for funders here to require standardized data in funded projects? So applicants can be asked to justify the standards and instruments they use I thought it's an interesting angle. I don't know Bruce Jonathan. Do you have any perspectives on that? It's it's It's tricky that um The we can lock things down too tightly and then that becomes a barrier uh, I guess in terms of you know publications say look data needs to be made available is kind of a a financial Process of saying it Make your data intropable. Um, I think it's certainly government organizations need to be be looking at that It's their role to manage their data and then and I think part of the whole managing and then curating your data is about making it available and intropable It it's it's it is tricky to work at what that that balance is between I guess forcing Using a financial driver to to make data available Okay, Jonathan um, I Mean we I Bruce and I work in the earth and environmental sciences space but looking to the other domains like the biomedical and the um, the biosciences, um, they've had a long tradition of Mandating that the data needs to be in a certain format and the advantages of that is then it makes them interoperable across different projects programs So I think there are merits of um mandating either through funders or journals that you know, certain data sets be Published using certain existing data standards if that's you know, if it provides those advantages Yeah, I guess if you know there are advantages, but but you know With the standards not there Then who's going to sponsor that standard if you mean that? Hmm Okay Well, well, thank you. Thank you, Bruce. Thank you, Jonathan for your time It's been really interesting to unpack unpack a little bit. What does interoperability mean in practice? And I think especially those perspectives around the organization and the commitment behind it I think I've really interesting questions and things that are certainly of interest to Well to the community out there and also to us the research research data commons when thinking about If you want to bring data together into one Virtual commons, how can you actually do that in an interoperable fashion? And what does that mean and what are the aspects you need to consider for that? So thank you again very much for your time and Looking forward to ongoing conversations and developing this thinking further around interoperability Thanks, Keith. Thank you. Thank you