 Hi all, welcome to our today's talk on the Akamo in Notso standard. I hope I'm pronouncing it right but Ashok will do it again, tell you how to pronounce it the right way. So the Akamo Notso is open document standard for judicial and legislative documents essentially to store these documents in an XML format primarily designed for parliamentary legislative and judicial documents. While this has been designed under the Africa Eye Parliament Action Plan, we have with us Ashok who was involved with the making of the Akamo Notso standard and he has been in the legislative legal informatic space for the past 15 years and Ashok will walk you through the specification and why courts across the world are using this standard for recording legal documents. Ashok, over to you. Hi everybody, so my name is Ashok Harihara and I'm going to be talking about the Akoma Notso open standard. So it's not Akoma Notso, it's Akoma Notso just a clarification there. So just a little bit about me before we you know before we start. So I've been in the legal informatic space since 2005. I was a member of the UN project from where Akoma Notso originated. I am particularly interested in the XML stack of technologies, you know, Xquery, XSLT things like that. Also process workflow design, mix of passes in front of frameworks. I've done multiple projects internationally that make use of Akoma Notso. So I'm a sort of practitioner of Akoma Notso also and I run a company in this space called Bongani Consulting. So what exactly is Akoma Notso? So I could give you the technical explanation but let me just start with a story so you understand the context of what exactly Akoma Notso is. So right now you know we are all doing this conference from home because there's a pandemic out there and one of the laws that has kicked in because of the pandemic at least in India is this Epidemic Diseases Act you know of 1897. So mind you this was a law passed back in 1897 and that is the law that has been used by the government you know to implement various protocols. So this is the pertinent law for the moment but it was amended. So they took this law of 1897 but they passed an ordinance in 2020. It was called the Epidemic Diseases Amendment Ordinance of 2020 and this is what it exactly does. So they had the law of 1897 they passed a new ordinance for various reasons which we'll talk about and what it exactly says is on on line three in section three it says that you know we are changing the act as it was in 1897 and we are adding various other conditions to it related to harassment of doctors and so on and so on. So and this is what it looks like. So are these just words on paper so you have the law that was passed in 1897 you had the modern one with changed the old law of 1897 but are these really just words on paper? Not really. So just to give some background this was 1897 you know that's a doctor giving some kind of treatment to somebody who got the bubonic plague in Bombay. That was the reason why the law was passed the Epidemic Diseases Act of 1897 and apparently they didn't have to use it until 2020 and this is the this is a picture from Bombay now you know all these doctors dressed in PPE gear and taking the temperature of someone. So what happened was there were doctors who were being harassed by people because they thought the doctors would carry us themselves so they have to change the Epidemic Diseases Act because of social condition. So they passed the ordinance which I showed you just now and if we if we consider the original Epidemic Diseases Act as I've labeled it original this is called the amending ordinance or the amending act so to speak. What it has is merely instruction. This document just says that go to section 3 of the Epidemic Diseases Act of 1897 and add these new lines. That's exactly what it's doing. So it has instruction which allow an amended Epidemic Diseases Act to be created. So what they publish is this and what legal drafters go and do is take these changes like the way you do track changes in words they go and take the original 1897 act they go and type in these changes and they apply it and that creates the new amended Epidemic Diseases Act of 1897 as changed in 2020. That is roughly you know to put it crudely that is how things happen that is how acts get changed. So just to take a longer term perspective you know I've added a picture of 2040 there hopefully you know Mumbai is gonna look like that. Let's all think positively and so until March 2020 so I'm just revinding back to March 2020 when this amended ordinance had not been passed. So until March 2020 the original 1897 act was applicable. It was not the old 18 you know it was not this amendment had not been passed. Then what happened in 1820 in 2020 we passed the amended the diseases act and from April 2020 onwards the amended 2020 act version is applicable until any future amendment. So if we assume that there is a future Epidemic Diseases Act 1897 with further amendments passed in 2040 you'll have the original 1897 act you'll have the amended 1897 act as amended in 2020 and then you'll have the 2040 version. So this is this whole chain is called the Epidemic Diseases Act. So the Epidemic Diseases Act is not just one document but it is all of these you know the complete history of these documents and why is that? Because all three of the laws can be applicable in 2040. I mean if you look at the Indian legal system sometimes cases get stuck in court. So let's say somebody went to court in January 2020 and somebody else went to court in May 2020 and the cases were you know they were related to the application of this act but they went on for decades and it'll turn out that in 2040 the new act will be applicable but the old cases for the old cases that happened after April 2020 this amended 1897 act is going to be applicable while the one prior to April 2020 the the original Epidemic Diseases Act is going to be applicable. So all three versions of the law can be applicable in 2040. Many people don't realize this, many attack people who approach the legal space they don't realize these nuances until it's too late so that's why I'm highlighting that and because this idea is part of Acomantoso this core idea is part of Acomantoso. The law is applicable based on its point in time that is a key thing to keep in mind. So the key ideas and concepts are that I'm just summarizing it here so legislative and judicial data is very different from financial or sales data. So number one you know legal data is all interconnected across time. You saw in the previous example all three versions could be applicable at a particular time. It's also very long lasting so in our case it was like a hundred-year-old act but it could go on to hundreds of years depending on how old the legal canon is in your country. The law is about documents it's not records in a database because that is how tech people approach everything you know trying to normalize stuff putting it into a database and so on. The process is legal data so the process itself the way the law is created is actually pertinent data content metadata and presentation are important I'll show you what that literally means. So all these ideas together are part of the key ideas and concepts of Acomantoso. So what is Acomantoso? So you had some background of laws and how they change over time and how they're impacted by different things. So Acomantoso is an XML document format much like other XML document formats which are out there like data or doc book or open document format or open XML you know all these are popular formats they're used in Word they're used in open office they're used in archival systems and so on. Acomantoso is not just a format to store laws it's also an approach to manage legal docs and associated information. So it is a standard I say it's a standard because it's not created by a company or you know or by a vendor trying to sell software it's under a body called the OSS standard body. If you if you google up OSS and standards you will find that this is a kind of industry organization much like IEEE and you know other industrial bodies who create standards for different sectors. So how do we pronounce Acomantoso? It's Acoma in Toso that is the right way to pronounce it. It is means link heart in the Akhan language that goes back to where this project began because we were you know based out of Kenya in Nairobi and that is where the project came out of so we decided to take something from Africa. So it began with the UN it is now with a standard body. How was this created? So it was not created out of Kenya. People then just sit down and you know come up with a brand new idea on their own. No you know this was built on the back of a lot of existing work done by other people like the European Parliament you know like this UN Pan-African Interoperability Framework the legal XML community. So in 2006 the UN essentially collaborated with the University of Bologna which is the oldest legal faculty on the planet. They go back to 1088 I believe so they're like a thousand-year-old legal faculty. So you know as the UN you know we collaborated with them and we basically started building this legal standard so it went through multiple iterations it adopted different ideas from around the world you know I've mentioned took some ideas from Brazil it took some ideas from this other project in Netherlands finally in 2012 the UN decided it was not the right place for a standard so it got handed over to this OSS standards body where over a five-year period it went through something called a standard ratification process. So it went through committees it went with involvement of different parliaments and you know people in the commercial legal space and then finally in 2017 they released Acomantoso as a legal document standard it's the only legal document standard out there so what are legislative and judicial docs so they cover a whole gamut of documents these are just things I've taken out of the top of my head and put them there but there are a couple of dozen more or even more because every time I see a parliament or I see some court documents there's always a new one in you know depending on the national or cultural tradition there's a new one so this is just a bunch of you know documents that came to my mind you know bills acts debates amendment amendments things that typically pass through a legislature and which are required to produce a law and these are on the right side you will see they are legal documents coming out of course so who's using it just a quick view it's you can see it's being used around the world not by small bit players but it's being used by the data archives in the UK it's being used by the US House of Representatives it's being used by the European Parliament it's there's ongoing projects in FAO and also in Spain you know they have the office of the president who are in charge of legislation they have a big project going on there they have the LexxML project in Brazil they have the library of Congress in Chile unfortunately nothing out there in India you know in India the whole legal informatics space is really I think it's very it's at least 10 years behind the rest of the world and not many other projects in Asia though I've heard of initiatives in Hong Kong and in Japan so what are all these people using Acomantoso so I'm just going to give you a quick view you know these are all the things that people are using Acomantoso for drafting legislation publishing laws building search engines on laws consolidating amendments into laws you know you know so you know anything surrounding the law or judgments or legality that goes into the Acomantoso bucket and they're doing a whole variety of things and who are these people so there are different entities who find views for Acomantoso so the most typical use case are national and regional parliament so you know for example in India we have the Lok Sabha then we have the state legislatures so those are like typical candidates for this stuff then there are the you know different judiciary supreme court the high courts you know even the smaller courts in your countries or places you know they're all applicable candidates international organizations and agencies people who pass standards these international organizations have legislative bodies with member countries who participate and they produce standards or you know they produce communicates and resolutions and so on those are candidates organizations publishing laws online you know legislative and judicial monitoring organizations they consume they know they're consumers of the data but I've seen people even creating data in Acomantoso from national archives digital libraries commercial law data so that's the whole spectrum of players in this space why are these organizations using Acomantoso so just to give some reasons as to why some of these organizations are using Acomantoso so in many countries there are transparency requirements under law so they have access to information laws so they have to publish their laws in machine readable formats you know and to be to allow people and other parties to do things with it to to basically query the data so by this I don't mean just laws but things like debates you know parliamentary debates who voted for what that is actually very interesting and crucial data in many countries to see who voted for what law and so because then they can build a thread in terms of why are why are certain members of parliament voting for certain things are there loving groups involved so there's there's all manner of information that is extractable once you once you have it in an open standard there's also shared knowledge you can provide a it's essentially provides a common vocabulary to build upon distillation of best practices so you know basically Acomantoso has been built on the back of best practices so as the standard was developed over a ten year period it has adopted best practices and standards from parliaments from courts and from other archaegers you know people who have been doing this for dozens of years and it has distilled a lot of the best practices in the standard itself and you can build skills around the standard so if you look at different legislatures and courts around the world many of them have tried to build their own system they think their problem is unique they are the only people in the world to have this problem so they went around building their own system but then you know systems need to be updated it takes money and time but if you have a standard then you can build around that and you can kind of use the common pool of you know of of like knowledge and developers and people who are around it to build it so using a standard also lowers risk for many people so because the standard means it has worked somewhere else for someone else if they can google up they'll find a case probably and they'll find out that other people have used it it's also risky to build systems and design your own document formats from scratch because you know laws tend to survive for very long periods of time so you want to have a design which is robust and tested and has been built with longer periods in time so building on top of a standard automatically reduces risk it also lowers cost because using a standard allows you to reuse stuff other people might have built systems other people might have built a library to work on the standard so you can take that from the open source space and reuse it it also provides a sharing of best practices you can read about how somebody else implemented a whole system based on this technology and then you can reapply you can take their best you know you can take their context and adapt it to your own context and save time and cost there's also a marketplace of vendors and consultants are understanding okay so how are laws made this is a graphic i'm told off the internet it's talks about how laws are made in america but the context is applicable everywhere because this is what happened in very many countries including ours and it's talking about a bell being introduced in parliament the bell being you know going through parliament being debated committee work debate amended finally it becomes law so this is the entire process that the law goes through before getting published and this entire process is important legal data is interconnected so if you saw the previous graphic i've just distilled that into a simpler form here you know as a cycle so laws are drafted you know they are drafted in parliament by people called legal drafters who talk to members of parliament and then they go and type in their little computer in there and they produce bills they could be amendments also these get discussed in parliament they get debated by members of parliament those debates are recorded then the laws are passed they become acts they get gazetted those are more official documents time passes you know there's social changes you know new pandemics and so on and then the laws get either challenged in court or you know there's an ordinance passed or sometimes the court itself is making a judgment which is saying you know we want to strike down section three of that particular act and so on so in court or in parliament when they want to change the law they usually debate you know and they look at why did the law you know why did the law say what it meant you know did it mean uh is is what they meant actually applicable in the current context so we need to change it so and then the law gets changed and then the whole cycle goes on so that is why we have laws from 1897 which are still applicable now because they can be amended and they keep going through this cycle legal dogs are also very long lasting so this is something again many tech people they don't get a handle on because they're trying they're used to solving problems in the now and they don't think about the past so when we were designing the coban to so uh this was you know so i you know i'm from the tech background originally and so when i first saw the laws you know i thought hey this is the computer science problem you know putting things into databases no no no it's not like that so the laws are very old you know for example the oldest extant law now is the statute of malboro which goes back to 1267 it's a uk law and i believe four chapters from this law are still applicable i think the oldest indian law is probably the indian penal code of 1860 uh you know maybe there are some local laws in goa which are older than that which are still applicable from the portuguese era but you know that is at least what i could google up and find so in terms of longevity uh there are two aspects here you know when people deal with legal data they have shorted me so this is the this is the problem i mentioned about tech people trying to solve problems now so they want to use the legal data in different ways they want to put it on an app they want to index it with the search engine they want to run it through document workflows they want to consolidate the legislation or maybe they want to classify something so this picture here on top on the right is from uh you know from the us house of congress search engine which allows you to search laws you know and the one below is from the ua e uh you know data archive of their laws where they've done some kind of analytics with the laws i can't read arabic so i don't know what it was about but just to give you an idea so these are short term names and the problem is many tech people tend to approach legal data from this point of view like i want to solve a problem now so let me just build a system or let me try to store the data in the present context and see uh and solve the problem now but it does not work like that because uh you know are your laws and judgments going to be readable in 10 20 30 40 or 100 years time so if you're going to put it into a little database in is that database going to exist you know those are questions that you really have to think and those are questions that are being answered by a comment also so uh and you can't rely on specific software architecture or frameworks or even the existence of documentation and 40 years time so people have to be able to take your data understand what it meant and uh you know build systems on top of that just on the basis of data so that is the orientation that a comment also takes it's got a very long-term perspective and in terms of long-term it looks to the future and it also looks to the past so docs have to be self-expressive and self-contained uh so in terms of longevity so this is the space that it falls in so there's short-term need that are important but they're also long-term challenging so a comment also falls into that space in the middle it tries to enable those short-term needs and also tries to reduce those long-term challenges uh it's always there in the back of its mind so legal docs are not records in the dv so this is the classical problem i mentioned you know so i'm a tech person i see the law and i think you know this can be normalized act title uh act date long title reamble section one so i'll put all those into columns in a in a database record and i'll stuff it into a dv sounds great right so that's a great design this is very bad in the legal space this is not good at all it's not predictable because you saw one law and you thought that that is how the data structure is but then in two months time there's some legal drafters going to come up with another design or there's going to be a law that you had not seen yet but in the past which will completely defy that structure that you imagine so putting it into a database like that normalized there's no room perfect the other aspect is you have to always have some kind of software to recompose the law and show it but the problem is then if you want to render your law uh in 20 years time you're still stuck to the middle ground you don't have to call the consultant when he's 80 years old to pull out data from the assistant because just because that's the only way you can recompose your legal document based on the middleware that the guy wrote 30 years ago so so these are problems that are common to suit addresses it's about self containment so if a law is a document it's always stored as a document it looks at separation of concerns which is which tech people will understand very clearly but i'll show that with an example the content and metadata live together but they only reference but only to reference each other so i'll give you an example of what's happening so for example content metadata and presentation so i said akumantosu can look into the future but here i'm looking into the past this is the nalanda inscription from 860 so this is from the nalanda from the ruins of nalanda some archaeologists pulled out this copper plate uh you know which is a land grant given by the king to some nobleman and it is actually a legal document so if you look at it closely you can break it down into a preamble a body and a conclusion so this is in sanskrit i can't read it but i actually read the translation so it has clear segmentation like that uh just like a law has uh it has type metadata so there's a royal decree there right on top uh you know that the king's emblem which says it's a royal decree it has an authority character saying you know this is under the authority of king uh you know deva paladeva it has placed saying where was this law passed so budgagiri is the place it has language metadata saying it's sanskrit so in that period they were you know they were publishing stuff in uh sanskrit and pali so i suspect the pali was written on paper our family if that got destroyed it wouldn't be the proper date for it survive and it has a date right at the end which says uh year 39 date 21 years always began with the birth of the king so i i presume that is what it's about uh and then it has uh so this is authentic content because it is authenticated by the king's royal seal much like how we put rubber stamps or we put wax seals nowadays that is how it was authenticated it has inline metadata it has metadata within the content it has names of places it has names of people it has designations of who these people are to who this land is being planted to and so on so this is a legal document from 860 and this can be represented in a comment also so there is uh this is the content it has presentation so it's the presentation is basically copper plate but the same law can be transcribed into a computer and then the presentation becomes hdml so so you get the idea of how content is separated from presentation you can take the content and render it into a digital view which looks like a copper plate or you can put it into your page or your mobile and so on and there's metadata around it and then there's the presentation aspect so this is a modern law this is a you know the us code you know one of their laws from 2011 uh so i've given it the same treatment you can see you know what is applicable for the 860 law is also applicable for a law from 2011 there's metadata there's a preface there's a body there are sections there are short titles so a comment also aims to identify and tag specific parts of the document using a standard vocabulary in a way that in this information can be extracted and made sense so the process is also legal data so i mentioned this earlier that the process is very important uh so the process in this case is what are the circumstances surrounding uh a judgment or a law that was passed we saw the circumstances that went around the epidemic act where i showed you that the law was passed in 1897 there was the bubonic plague in Mumbai and then you know time passed and then there was a need to amend the law uh in 2020 and then the amendment was passed but it meant that the whole chain of laws starting in 1897 until 2020 is one work you know that is when we talk about the epidemic act it is it means the whole body of work surrounding that because that provides the full legal process and the full legal context surrounding that so that would include debates which went into why was that epidemic act passed and so on so here i've given an example of a judgment uh you know which was a very interesting and pertinent judgment part in the you know in the international court of justice in the 60s uh so basically Ethiopia took South Africa to court uh to the international court of justice saying uh that the apartheid policy uh was against human rights and uh they ended up losing the case uh incidentally in 1966 and you can see there's the history of the case going you know with different opinions dissenting opinions declarations and so on and there was one particular dissenting opinion from a guy called Judge Tanaka who voted against uh this particular judgment so the process uh so essentially Acomantoso allows you to track uh this kind of context to the law so if you want to know who dissented uh you can you can essentially query and know who dissented who was the person who dissented it was a guy called Tanaka what was his role he was a judge and uh what was his opinion you can pull out even the structure of the opinion so you can query different contexts you can pull out all the opinions you can find out what was the type of opinion was dissenting uh was it an agreement and so on and so that's what i mean by the entire process uh is part of essentially legal data uh there is an Acomantoso community out there uh if you google up you will find plenty of links uh but essentially just to give you a summary so there is an annual Acomantoso conference which is held every september in italy where uh which is typically hosted by the department of legal informatics in the university of polonia and uh this is the place where people working in the Acomantoso space be they parliamentarians uh you know people in the data archiving space uh people in the tech space you know vendors you know uh companies operating in this space like mine we normally go to this conference and we present what we did in the past year we you know we basically talked to other people who are dealing with them there's the os's technical committee uh which is dealing with the standard uh there are also associated standards which are kind of related to Acomantoso so one is called legal rule ml uh which is about uh putting machine readable rules so you know rules and regulations are also covered by Acomantoso so for example uh you know your uh uh tax notifications you know customs tax notifications there are literally thousands of them and they and they also follow the same model that you have a customs notification issued by the uh by the indian customs body which changes the previous modification and that notification changes the prior one there's always this chain of a notification changing the prior one and that prior one changing the prior one so that is the typical use case for Acomantoso that's what it addresses now people have to go through paper to find out what was applicable in my case which is going on since 2006 so you know I have to pull out the paper and see what is the connected notification so on so all those cases are being uh addressed by Acomantoso and some of these associated sister standards so there's another one called legal sitem which I think is still going through uh essentially a committee process or which looks at standardizing legal citation uh there's a public mailing list uh there's Acomantoso XML mailing list on Google groups uh where which has a membership of around 500 or 600 everybody uh and there are people asking questions they have to use it uh people doing things in Hebrew and then uh and in arabic and in japanese so it's all there so there's a vendor supported systems out there they're commercial vendors I know who are supporting Acomantoso uh building solutions on it uh there's a growing community of open source applications and tools so even our company you know we have published open source applications and tools there's uh there's a site called schema.com which allows you to query the schema as a learning tool uh I've got a blog called Acomantoso.io where I infrequently write about Acomantoso so that is essentially what Acomantoso is about that was the last slide uh do you want to show us any websites that are already using it for people to have a look at it? Sure I believe the UK data archive is using it uh hang on I don't have the link on me right now uh hang on let me just get it for you thank you but people if you have any questions uh that is the question and answer tab you can just ask the questions down there if you're watching it over youtube you can even pass past your questions in the youtube chat and we can take it over from there while we wait for questions uh uh let me let me field mine first for sure. Sure go ahead. Hello all this is uh this is Prasanna I'm a practicing lawyer in Delhi basically qualified as an advocate on record in the Supreme Court and I have a background in technology as well so I wrote the enterprise software for more than a decade so I'm already sold on both the need for and the desirability of a standard in a space like this uh when I first took to law coming from a technology background two things that struck me for that first law was so much like a code so a project like this a standard like this was certainly something that I imagined would exist would probably evolve uh something that must be there. The second thing that that struck me was that there was nothing like code right so there was I was really torn between these two uh two two different kinds of thoughts crews of thought when we say uh so I started both as an optimist and a skeptic at the same time for this rule right in fact to give you examples on why it is like code in fact there was a seven judge bent to the Supreme Court of India in the early 70s that tried to interpret what the lists of constitution meant so there were the there were three lists under schedule seven which kind of delineates the powers between the state and the center list one of uh list one of schedule seven in the constitution refers to all powers of parliament as the union legislature and the list two refer to powers of the state legislature and the list three refer to the concurrent powers that both state and union legislatures can legislate on so there was this question so there was one particular entry in the list one that said taxes on everything except agricultural land there is another entry in list one that said that's a reasonable provision it's a catch all saying everything that is not enumerated in either of the lists is a power that belongs to parliament and this question of taxes on agricultural land it was not enumerated on either of the three lists so the question that arose in the night before the seven judges of that how do we interpret this so whether we interpret this entry to say everything except agricultural land whether this refers to an intention of the constitutional makers to exclude this legislative field from parliament's domain or whether we should interpret the residual clause to mean that it's a catch all anyway and therefore there is it's not enumerated anywhere it will be caught all by the residual provisions so that in fact they went out and and said it is so all of these three lists need to be interpreted as though they are definition they're not transactions they don't confer powers so it's very similar to court where we have separate we know what is an invocation and what is a definition for instance you know one on your environment programming language for instance right so this particular and even in tax law we have water courts charging provisions and what are called definitional provisions so in many ways our laws aren't like courts in fact our penal code is called the penal code Indian penal code we have a court of similar procedure court of similar criminal procedure with insolvency and back-up cheat codes several of them are court codes on the other end of the spectrum where I feel laws are anything but anything like nothing like court is is the industry that is that is developed around interpreting laws so as you said in your presentation also when you said the whole point is to make the laws more readable right readable like I presume is more understandable it is for the general public to know what the law is say the ignorance of law is no excuse and so it is for the general public to know what the law is but the question of what is the law is what is actually that's the industry that is that is developed all our lawyers and the judges all of them try to discern it for the general public as to what the law is and what is stated in the statute is just a starting point of what the law is so I'll give you an example under a rent control legislation so a landlord has a right to evict a tenant if the landlord has bona fide requirement for any residential property that he has missed out so this place less residential property was was interpreted by the Supreme Court to include commercial property as well so by a stroke of pen the entire gamut of meaning of the legislation got changed as to whether because there was a specific intent only to include residential property within the purview of that particular right to evict on bona fide requirement that will expand it to include commercial property as well this is something that you will never know until you know the judgment that actually interpreted this place and that is so many of our laws in fact in fact there are many communal jurisdictions where there are volumes produce what are called restatement of laws so what is written is is just a starting point and as you said it is organic growth in fact sometimes it is it is growth where we don't even realize it's completely unrecognizable from what the text is and there are certain other problems that compounders will give another example of a recent case that I am currently litigating in fact we have a question of whether indentation in your provision in a in a statute itself means anything at all so there's an example that is that is not a surprising thing to hear you know things like indentation or drop caps you know I've seen cases like that where the drop caps has a particular meaning that's because somebody interpreted it to mean something in the past you see that also with side note requirements you know which are technically very challenging to render but they become part of like standing orders or tradition of how the law is written or they are part of a legal drafter style guide just because they decided to write notes by hand historically on the side carrying that over into the text space becomes a talent but those are you know kind of the presentation issues that are also being addressed by the standard so one one whole part of the standard is how to enable publishing on paper so you know one you know some of the projects I've been involved in there was one I did with FAO which was all about being able to draft meeting documents you know using a command to so online being able to do communal amendments you know being able to do it together as a group and then being able to render it on paper in six different languages in the way they have been rendered for the past 30 years so these all seem like conflicting requirements as I said you know there is a requirement from the past and there's a requirement from the future so it's trying to bridge both those requirements so you'll find other cases for example in the US code they refer to line number not just in the US but you'll find that in different countries you know where for historical reasons they refer to line number and the line number only has meaning on paper that too on specific sizes of paper but it does not carry over into the online space so how do you cater to a past a requirement from the past that requires referring to line numbers on an A4 sheet or a letter size sheet and then also support rendering that same law on iPad or being able to compose the law on an iPad or you know some kind of modern browser based device so those are the kind of tech challenges it tries to balance out you know you want to support the past because there's a whole legal canon going back a thousand years which needs to be supported and then you also want to support needs of the future so not just from the presentation point of view or you know from the drafting point of view but also in terms of searchability like you you know the case you mentioned in terms of interpreting the law itself is is this a definition so what Acomantoso allows you to do is take a law take segments of it and say this is a definition so you define it semantically you put a tag in there saying this is a legal definition and so it is clear even for somebody looking at it 20 years in the future as long as they can read English or they you know they can understand that this tag says definition so this is what they meant they meant it as a definition so and that's what it means and even there there is a whole context of local jurisdiction so if you you know for example in India there are parts of the laws called articles you know you have you know within articles you have sections I believe if I'm not wrong so in other countries they could have sections and as a higher level structural entity then within that they can have articles so you know it's not rigid in those ways so it tries to fit your local jurisdiction if article is your main structural entity in your country you can use article and you can put sections inside it so because you know 50 years in the future the people in your country know what the law means and you know they know what the article means they know what the section means so they're going to interpret it appropriately so that is the whole context of it so you know because a lot of times understanding what the future might be is just you know how the past was so that is you know that is the kind of space it sits in okay we have some questions Prasanna do you have a follow-up yeah my quick follow-up was that in India at least on the question of indentation we've not had case law on whether indentation place has any role at all in interpreting what a particular proviso means so that's something we hope to find out in the next few years but in the mean right there are a few questions that I take yeah so there's one by Mr Javi who's an advocate the government doesn't adopt the standard what is the path to make it workable in India which has a huge volume of legal data generated on a daily basis does it depend on a body or volunteers constantly tagging and updating data now I really think this is this is not something that can be done by volunteers for the simple reason that it's the government that is producing the data you know the data you know the government is passing the laws and it can be an infinite effort for the you know for the community to go and try to be encode the law in this form when the producer is actually the government the most efficient way is for the producer to produce it in this format that is what they are doing in most other countries they're producing it in this format you know instead of trying to take because I've done projects where I've taken PDF documents like I did a project with the eco was body which is the economic which is the economic community of west african states it's like the you know it's kind of like the european union but for west african countries and they had like 30 years worth of meeting documents and resolution and so on so and they had it in PDF so I had to convert you know it was like a seven month project where we sat and converted you know 30 years worth of their these legal documents into akumantoso xml you know we had to write a whole bunch of dsls domain specific you know language processes just to take the pdf text and convert it to akumantoso xml it is a lot of work so instead if they had been drafting their laws and producing it in this format it makes things very easy and that is the root what it should be but where I believe you know the community plays a part is essentially putting pressure on government to adopt standards and you know provide us data in open standards you know that is what it should be maximum what the community can do is take specific acts and you know as a civic you know to make things easier let's say things like the contract act or traffic act you know produce like I think that is the only level that is feasible for you know for you know for the wider community to do it to take specific acts and provide I don't know apps or sites that allow a historical rendering of that in the context of you know let's say the contract act you take 10 years of the contract act and you put all the consolidated legislation in context so you know but I think that also has commercial implications for people trying to sell services as you mentioned there's a whole body of lawyers and law companies who are there to interpret the law and tell you that you know we have all the versions of the law and we know what that means it's very hard to find like all the you know all the historical amendment chains will say the contract law in India or the company act it's not there you know you have to pay a vendor to get it because on the government sites it's hidden in like dozens and dozens of PDFs you can't find it so that is the source of the problem and I think the only solution is to make them do it they have the money they have the resources and they are the ones who should be adopting standards you know they they need to have legal drafting departments who understand tech and you know who have some vision you know if you if you have a vision for a country then you should be having the laws also as part of the vision so that's that's what I think you know and it's not just law that applies to also judgments we are not an English speaking country so we have regional languages so and you know it also applies in the regional context very much the state laws because for most people it's the state laws that apply to them you know the health health and education laws are at the state level most of them so I think the linguistic context is very important here there are two quick follow-ups here so one is by J.D. Pintzelf where he asks if because there is going to be a huge amount of manual tagging involved if we have to do the coating of some of the earlier laws that we have to the AKN standard the other follow-up that I have for you Ashok is whether the standard itself supports this flag of whether a particular legislation has been produced in the AKN format from the source or is it something that is back up sorry I didn't understand the second part of the question no so the standard itself supports a flag or some kind of implication whether the particular document has been introduced from the source or whether it is a back-up version of an earlier law that is produced in another format okay yeah yeah of course you know so as I mentioned you know let me just take you back to this slide I didn't go too technical because I was not sure what you know whether people have been exposed to this at all but I'll just take you back to this slide for example can I just share my screen so in this slide hang on let me just change the yeah so you can see that you know this is the act and there are three versions of it you know there's the one of 1897 there's the one produced in 2020 and there's the future assuming we are in 2040 so you know so what Akumant you know what Akumant also does is it follows the IFLA standard so IFLA is like a lively body and they have a thing called you know functional requirements for bibliographic records which basically classifies any publication in terms of you know it's a work so this entire body of three versions you know the full historical chain of the act is a concept you know it's not a physical document it's a concept called the called the epidemics act that's one level and then it looks at it at the expression level which is in terms of if it's produced in English and Hindi you know there are two expressions of it one is in English and one is in Hindi so you know again that's a conceptual idea and then there's the manifestation which is the physical form itself so if it was produced as XML it would have the primary manifestation as XML if it was produced in PDF then the you know the primary manifestation would be expressed as PDF and that is fundamental metadata in an Akumant also talking well the other thing you know that it's about manual typing that is involved and is there any other way to work around that on any or trying to minimize manual tagging as much as possible so so there's there's different approaches to that you know so so you know one is manual tagging you know that's you know that's going to work if you it's going to work from if you start from the current let's say you you know you want to say I have a you know all the laws from now they're going to be you know they're going to be drafted in the system and all the laws from now would be drafted so even when these projects are done typically the now is one project of how you're going to handle the how you're going to handle the current problem you know how are the laws going to be tagged from now and the past is handled as a different project because typically you have an older canon of legislation which needs to be tagged and that requires different technology and a different set of skills so the output format is the same Akumant also but the technology that you apply is different so in the first case you know you will have a legal drafting editor which supports Akumant also and you would start tagging you know acts which are being passed now and build that are being drafted now you know you would tag them for the past I've done a project as I mentioned you know which took like 30 years of documents and converted that I so there's no one size fits all you know it depends on the context and the structure and what what probably works for you so what I ended up doing I looked at different technologies when I did the project I looked at natural language processing and you know different things but in the end what really worked was just writing a DSM at least in my case you know writing a domain specific language parser because the thing with laws is as you mentioned the laws are code so you'll find if you if you know if it's there in many countries they have a style guide as to how laws are drafted you'll find that the preamble nine times out of ten will always have the same syntax whereas whereas whereas it'll start with specific phrases you know having uh it'll it'll usually be in a particular case maybe uppercase and then when the preamble ends it will have a particular syntax there could be 10 different syntaxes but you will always notice that there is a particular syntax so if you analyze if you're able to analyze the body of documents you can possibly cover 90 percent or 95 percent of it using automated conversion you know by by writing custom dsls uh and natural language processing you know or integration of the tool you can do automated conversion of a lot of the past you know and then the remaining five percent is what you handle uh in the regular you either uh tag it manually or you partially tag it manually or you write custom scripts and so on so there's no one size fits all recipe for this because uh it really depends on the nature of the laws and what format they are in so if you look at uh government notifications for example if you look at tax notification uh you know like if you go to the customs notifications uh in the Indian tax portal you'll find that they come in very weird shapes and forms they don't have a stranded form because they are tied by some clerk sitting in an office but they're locked you know because they refer to customs regulation and they get amended and they have weird things like tables tables with call spans and uh you know those kind of things so those those those things are the real technical challenges how do you how do you convert those so what i see uh people do in many countries is not they decide not to convert things like those very complex tables because in some cases those tables were not even done on something like word they were done on a typeset machine you know they were done in an old typeset machine with cyclo styling or whatever and uh they are almost impossible i mean it's a cost-benefit analysis if your project has enough budget then you convert those parts of the document into you know uh into an electronic form or you retain it as the scanned image you know so like apple one too so allows you to do that you can have a whole part of the document as uh excellent uh you know in the in the radio form and then you can have parts which are just packed images yeah why do you know that that i have a quick follow Ashok so i'm so Delhi High Court has been in fact for the last about two three years has been discussing allowing judgments to have multimedia content in them because for several of these copyright judgments design judgments patent judgments that they give out it is much better for instance if there is super super cassettes that are sued copyright infringer it's better to actually have the multimedia uh uh clip in the judgment so in fact it is being exploring in fact in very likely in two or three years we may have judgments that may have these binary law the blobs in them so whether uh agn actually administers that whether it supports it whether yes yes it supports that it's uh you know it allows you to embed any kind of media content in there there's no there's there's no limitations and you know okay you know the standard has a set of tags you know the standard has about 200 plus tags to support different legal contexts and things but there's always exceptions you know so because the law is always about exceptions there's always a surprise there so what it allows you to do is also use your own custom namespaces you know so typically that's a presentation thing you know so you have some kind of weird logo or you have mathematical formulas for example there are laws which have mathematical formulas and equations and stuff like that so it allows you to use any other standard if you want so you can use mat ML and embed mathematical formulas in there uh there's no restrictions on that or you can define your own tags and do that you know you can define your own proprietary tags and support that things like that so thanks we have two more questions i think i don't think we have any more um so one is by laurence d uh he asked sam just quoting his question asses do you see any connection between this kind of formalization and the increasing use of machine learning in legal prediction is this a helpful connection that was foreseen by the development of akm or do you think the ends intended by oasis are being co-opted by commercial legal prediction firms now you know that's a pretty loaded question it depends on uh so you you can implement machine learning systems with akumato sorry i mean it you know it it depends on how you are tagging it uh but in terms of uh legal prediction i know there are ethical concerns and uh i mean there was a you know i've seen a case where there was one particular court uh you know what you know one particular country's court who are embarking on uh an akumato conversion project and one concern they had was that people would be able to you know that people would be able to take the judgment open data and be able to query the performance of judges to be able to query and uh you know come up with stacks on performance of the judges and then subsequently the judges would be could be influenced by that statistical comparison you know so you know let's say uh you know you want to monitor the quality of judgments in the country you could have an external uh monitoring organization take this whole body of judgments in akumato so and you know do machine learning on it and you know extract meaningful information like this particular judge consistently takes two months more than other judges or this particular judge uh issues stricter judgments for the same you know for the same kind of case you know somebody uh stole 10 000 dollars this judge sentenced the person for 10 years and he always seems to do that while another judge is far more lenient or you know you could do some kind of analysis just based on the you know just based on the affluence and respondents and say you know uh extract other kind of uh statistical information like you know many of the appellants were uh you know were of a particular race or a particular religion and their case failed i mean i'm just this is a hypothetical uh you know scenario but these are real questions you know these are uh real ethical concerns that uh i've heard you know over the years and unfortunately uh the you know the standard cannot address that the standard is about representing information semantically and what people want to do with it is really not you know under what the standard is about it because the judgment is public information so it's up to the publisher do they want to publish do they want to strip out some data and then publish it those are you know it allows you to do that for example uh you know there are specific tags available in the comment also where you want to omit names of appellants or uh you know in the case of cases involving minors or whistleblowers you can have specific tags where you want to strip names out uh you know where when the law gets published there's an indication to the software that is publishing it that these names need to be omitted you know so you can do things like that and you can have your you can decide not to publish information which allows you know things you know cases where judges can be compared you know i think that's an implementation specific detail uh which the standard really cannot address it it's a bit like trying to blame http for fake news i would say it is that it's really up to the implementer and the service that is uh providing a kumantoso or using the kumantoso it's an ethical responsibility at that time but but generally this is ethical concerns around judges being uh sort of under the standard i think we've had in law we've had answers for it for instance there we so we've got a resistant sort of live streaming cases for for some time now but that i mean those decks have been cleared by government last year perhaps the year before last although we've not had uh anything i mean live streams comments here but i think we will see that soon um yeah but that's that's a digression there's one last question Ashok of dviji guru um so that is what would a non-machine dependent form of a k and b essentially the compilation of the history of each act as recorded archive to the imprint or is it something else um i i i don't get this question for me dviji if you're still there can you if you want to uh ask this or clarify this question a bit more i think i kind of got the gist of what he's saying so right if you if you look at laws historically so you know the law is always printed on you know like for example um you know i gave the case of the nalanda or uh you know the other case uh of the statute of marlboro so in the olden days laws were written on parchment uh or on copper plates you know things which could last time so i think the question is uh in the case of apaman too so it's all digital you know it's tags xml tags in a digital document what is the long-term form of that so the long-term form of that is digital probably so that is the idea and uh you and the really long-term form would be print you know would be putting the xml on parchment maybe you know i think that is you know that is one way if you want to do it semantically probably you would uh edge that on parchment uh long-term archival paper something like that you would put the xml on there so that you would be able to extract it you know in the absence of any digital there you know to be able to get the same semantic information out i think that that is the only way to do it thank you very much ashok this is a fascinating topic i'm sure this is not something you've heard last off thank you very much for the time thank you very much everybody else for tuning in suniva's the last review i have one question i guess uh so in terms of this whole process of standardization and in terms of uh converting laws whether it's from the parliament or even the judicial orders uh what's missing i guess in india is that whole process of a standards body for digital right uh like we don't know for example the standard is being ratified at oasis in india you have beauty of indian standards for all forms of standards uh but somehow uh these standard bodies are like like mighty itself has one more uh the ministry of electronics and information technology has one more project called the national e standardization project under which is what we got a lot of e governance standardization some of it i won't say it applies to every sector uh so the question i have essentially for you is uh sure there are standards international standards some of which are being adopted by the bit of indian standards directly in case of isa standards uh how would one how should one go about like getting say something like a common or not so or or even a modified version of it for indian uh quotes to be uh say pushed through a standards body now you know i uh so i have i have a uh you know very commercial view of this so you know akumant also was made to uh you know to provide legislation and uh judgments uh you know in terms of open data and access but behind all that there is also a cost motivation so fundamentally uh it's about reducing costs and being able to deliver more at the lower cost so uh you know the way i look at it is not even from approaching a standards body but this is this should be more cto driven you know like uh chief technology officer driven unfortunately bodies like uh the courts in india or even the parliaments and even the national archives and people like that you know who are the custodians of these laws they don't have that figure i believe i don't think they have a role uh called a cto or somebody who has a vision about the data and how it's going to reduce costs and how it's going to make publication cheaper and how it's going to make access to data cheaper so and from what i've seen in other countries it's always cost and function driven it's never it's rarely standards driven that's a requirement by law in some places but the ultimate driver is always what can i deliver more at lower cost you know that that that is the that is the fundamental driver so in the case of the courts uh if you see the you know the uh you know what is the site where you can access judgment it's probably indian kanu and that is done by one person doing a project you know if you think of it a whole country like that all the courts they have not done anything you know they uh they could have even made money you know to literally put it like that they could have had you know they have the data they could have uh you know uh taken the data published it for the public free of charge and also sold it to vendors who can do other things with it like pay sir yeah you know something like that you know some kind of model like that so but unfortunately i don't know that vision is not there you know i that uh you know that data is essential for a country like ours to grow you know that legal data if it's available to everybody it's available to more people you don't need so many lawyers you don't need to spend a couple of thousand dollars just to fight your little case because the law is accessible it's online and you can get it and more lawyers can get access to it so the price has come down so i mean i've been involved in litigation myself and it cost me more to do litigation in india than in uh and say europe you know so uh it's uh it's crazy you know just because you are at the mercy of the law and the law is known only to a few because the access to the law is available only to a few and what is the background of the law what is the history of the law is not available nobody knows so that is the tragedy of it you know so i think it has to be function and purpose driven rather than a standard because uh at least that is my belief you know so okay the economics of it needs to be looked into and also the interest which could be by the courts or the judiciary itself uh that's a follow-up i just promised him that he can ask his question sanna you want to ask that follow-up i just asked asked him to uh do an oral follow-up are you are you able to hear us he can unmute himself and talk yeah hi uh this is dviji here hi hi yeah um i actually did type it out um so i actually i had to read the what i typed out yeah so uh the explanation to my question was that i was thinking beyond machines now i can understand uh akn makes a lot of sense when it comes to machines but if you think beyond machines and are trying to see if it's it's it's basically an extrapolation which is essentially about archiving am i right in understanding like that and the second part to that as an explanation note is that i'm trying to see if it is more about managing the you know the various things that are possible in the digital space which would not be possible in the print space or in the you know the traditional space so um this is more to understand the standard and some sense map out the possibilities rather than uh you know uh like i'm trying to understand what akn is so uh in to you know to put it simply akn is a format to represent your laws so so far you've been typing laws out in word and pdf document uh and while the laws are not words on paper the laws mean something they have a particular structure they have a preface they have a preamble they have sections which have numbers and they have and the sections possibly have articles within them which have numbers and the articles have paragraphs within them which are also numbered and uh they they refer to other legislation so uh and they refer to specific versions of other legislation so what akumant also aims to do is to capture that to capture the essence of that the essence and the meaning of that this structure and what these references are to other legislation in terms of citations what the numbers are what do these numbers mean so when you reference a legislation it's possible to reference you know when you write out article three section one paragraph two you should be able to resolve that to the exact point in the document you should be able to have a computer uh you should be able to write a program that can resolve that uh if you are provided an akumant also document by using very simple logic perfect make sure yeah that does that does and in some sense by practice i think you just answered that it's primarily about the digital space and that's fine i i think i'll read up more and then it will be a little more clear to me thank you okay i think we are at the end of the talk it's almost a 20 um and i hope we had lots of questions and i think uh shashan sinna uh who runs india kan indian kanun is was in the audience and he shared a link of his project i posted it in the chat uh which was on what he has built on tracking laws in india it's in the chat for anybody who wants to access but we'll try to get him for the next talk as well again thanks ashok for explaining us with the whole akumant standard and if there is anyone else who want to reach out to you post it off we'll get you connected to them and i guess we'll end this talk here