how an evolutionary framework can help us to understand what a domain ontology is (or should be) and how to build one 12th International Workshop on Formal Ontologies meet Industry (FOMI22) 12th to 15th September, Tarbes, France Chris Partridge, Chief Ontologist, BORO Solutions – University of Westminster structure focus: domain ontologies exploring the past to understand the present and shape the future exploring the past to identify patterns of information evolution mapping the patterns onto the present – to understand it better to give us insights into ways of realising the potential (in the future) summary 2 focus: domain ontologies a gap between aspirations and reality aspiration what is a domain ontology? : A common vocabulary for a shared domain of discourse. “To support the sharing and reuse of formally represented knowledge among AI systems, it is useful to define the common vocabulary in which shared knowledge is represented. A specification of a representational vocabulary for a shared domain of discourse —definitions of classes, relations, functions, and other objects—is called an ontology . This paper describes a mechanism for defining ontologies that are portable over representation systems. … In the context of multiple agents (including programs and knowledge bases), a common ontology can serve as a knowledge-level specification of the ontological commitments of a set of participating agents. A common ontology defines the vocabulary with which queries and assertions are exchanged among agents .” Gruber, A translation approach to portable ontology specifications, 1993. 4 enterprise aspiration in an enterprise: Gruber’s agents are application systems – and sharing translates to interoperability. In this context, with the current concern of achieving semantic interoperability, it is natural to position this concern in maturity models these typically define a level of semantic maturity that involves a common model – for example: NC3TA reference model for interoperability (NMI) “The NC3TA reference model for interoperability (NMI) establishes interoperability degrees … Interoperability degrees define a maturity model that captures interoperability sophistication. … Degree 3: Seamless Sharing of Data This level involves automated data sharing within systems based on a common exchange model . Degree 4: Seamless Sharing of Information An extension of degree 3, this level establishes universal interpretation of information through cooperative data processing.” Tolk , Coalition Interoperability: Beyond Technical Interoperability – Introducing a Reference Model for Measures of Merit for Coalition Interoperability, 2003. In this context, we assume that the ‘common exchange models’ and ‘domain ontologies’ share a common purpose 5 interoperability involves two ‘topics’ in the NC3TA reference model for interoperability (NMI) there is a statement “This level involves automated data sharing within systems …” this suggests two ‘topics’ and a relationship between them: systems, and (automated) data sharing if we take a diachronic look at these ‘topics’ it becomes apparent that they are at different stages of evolution 6 systems evolving: (expanding) islands of automation “Islands of automation was a popular term used largely during the 1980s … the … usage is [now] defunct” https://en.wikipedia.org/wiki/Islands_of_automation see also: Bjork, B. (1987) The integrated use of computers in construction - The Finnish experience, ARECCAD 87, Barcelona, TIME the typical enterprise is now so automated that it no longer makes sense to talk about ‘expanding islands’ of automation … maybe, given the size, ‘ continents of automation’. However, it does still make sense to look at the bridges between the islands/continents. 7 data sharing – starting to evolve: alleyways of semantic interoperability when one tries to trace semantic interoperability, to map the journey of data shared between systems, it often travels through narrow APIs. continuing the metaphor: this leads to a cramped structure more akin to alleyways than islands – alleyways that are not always connected. this is an indication that data sharing (semantic interoperability) is not as evolved as automation (it still has a way to go). 8 two evolutionary stages domain ontologies aspire to expand the alleyways – to make the bridges as wide as the continents – enabling seamless data sharing data sharing emerges as a requirement out of the successful evolution of data processing systems. From another perspective, data sharing (interoperability) depends upon data processing (automation) 9 {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} second stage data sharing interoperability alleyways less evolved {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} first stage data processing (systems) automation islands ð continents more evolved reality? building domain ontologies to provide ‘automated data sharing within systems based on a common exchange model’ is challenging, typically requiring manual intervention “... Since domain ontologies are written by different people, they represent concepts in very specific and unique ways, and are often incompatible within the same project. As systems that rely on domain ontologies expand, they often need to merge domain ontologies by hand-tuning each entity or using a combination of software merging and hand-tuning . This presents a challenge to the ontology designer. ...” https://en.wikipedia.org/wiki/Ontology_(information_science)#Domain_ontology 10 exploring the past to identify patterns of information evolution exploring the past to identify patterns of information evolution mapping the patterns onto the present – to understand it better to give us insights into ways of realising the potential (in the future) evolution classification facet framework brain ó brain brain ó external external ó external storage processing communication brain external owner process path types 12 literate oral pre-oral stages identifying disruptions: key evolutionary events this minimal set of facets broadly characterises information technology’s evolutionary past through classical major disruptions {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} pre-oral oral literate brain storage YES YES YES processing YES YES YES external storage NO NO YES processing NO NO NO communication brain-brain NO YES YES brain-external NO NO YES external-external NO NO NO information’s classical disruptive shifts legend disruption 13 evolution classification mapping the evolutionary stages onto the classifications brings out the boundaries that mark the classical disruptions a more refined view of literacy what is literacy? “Literacy in Western cultures is not just learning the abc's ; it is learning to use the resources of writing for a culturally defined set of tasks and procedures. All writers agree on this point.” Olson, The world on paper, 1994. in other words, literacy is the communal cultural practice of producing and exploiting textual resources – more specifically it involves both collating and creating common access to historical textual information as well as creating new textual information . It is a property of societies rather than individuals. it seems a reasonable assumption that once writing technology becomes available, that literacy (in this refined sense) would naturally follow. But examining history (the past) reveals this is not the case. 14 pre-printing textual practices were oral “ We know a good deal about the actual procedures that Thomas Aquinas followed in composing his works, thanks ... to the full accounts we have from the hearings held for his canonization. … ... Still stronger is the testimony of Reginald his socius and of his pupils and of those who wrote to his dictation, who all declare that he used to dictate in his cell to three secretaries, and even occasionally to four, on different subjects at the same time . . . No one could dictate simultaneously so much various material without a special grace. Nor did he seem to be searching for things as yet unknown to him; he seemed simply to let his memory pour out its treasures ...” Mary Carruthers, The Book of Memory, 1992. “… composing a text was not writing at all but composing mentally and performing orally and, on occasion, dictating from memory. Carruthers (Carruthers, 1990, p. 6) argues that Aquinas' multi-volume Summa Theologica was produced in just this way: … these highly literate medieval churchmen did their work orally, relying on memory for examining, criticizing and developing ideas rather than relying, as is usually assumed, on the written text. Sermons were composed in the mind and sometimes written down later. Texts were not scrutinized so much as used as a record against which to check memory. Reading was not so much a matter of studying a text as ingesting or internalizing it. Once ingested, it could become the object of meditation and reflection. The scrutinized object was in the mind not in the text.” Olson, The world on paper, 1994. from our literacy-as-a-practice perspective, t he western medieval world was oral 15 evolution classification extending the facets to handle literacy 16 writing printing text stages text oral pre-oral stages brain ð external external ð brain brain ó brain brain ó external external ó external storage processing communication process directions of communication we firstly need to recognise that text evolved (in the western world) in two stages It is also useful to distinguish writing and reading – the directions of brain-external communication {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} text writing printing brain storage YES YES processing YES YES external storage YES YES processing NO NO communication brain-brain YES YES brain-external brain-to-external YES YES external (text)-to-brain YES YES external-external NO NO adding writing and printing to the framework we map these facets below and mark the disruptions from a storage perspective, if one adopts a narrow sense of text as letters and words arranged upon a page and reading as the ability to recognize them, then on this basis there seems to be no real change legend disruption 17 Lorem ipsum dolor sit amet , consectetur adipiscing elit , sed do Lorem ipsum dolor sit amet , consectetur adipiscing elit , sed do eiusmod tempor incididunt ut labore et looking at directions of communication 18 with the communication aspect divided into two directions, one can ask how practices changed in these two directions. the differences between the physical technologies of writing and printing (of brain-to-external (text)) are easy to see – as they involve different kinds of artefacts (quill pens and printing presses). however, if one takes a broader literate view, and looks at the cultural practices surrounding reading and writing, clear differences emerge between the writing and printing stages. {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} text directions of communication writing printing brain-to-external (text) writing printing external (text)-to-brain reading (pre-printing) reading (post-printing) legend disruption Lorem ipsum dolor sit amet , consectetur adipiscing elit , sed do Lorem ipsum dolor sit amet , consectetur adipiscing elit , sed do eiusmod tempor incididunt ut labore et printing texts “… the new tables may have been historically significant not so much as a ‘replacement’ for the old ones but rather as an alternative set that encouraged further checking against the writing in the sky. At all events, they were put into use very quickly - as the case of the young Tycho Brahe shows: On 17 August, 1563 at the age of seventeen while Vedel was asleep he noticed that Saturn and Jupiter were so close together as to be almost indistinguishable. He looked up his planetary tables and discovered that the Alphonsine tables were a whole month in error ... and the Copernican tables by several days. [Koestler, Sleepwalkers, p. 287.]” Eisenstein, The printing press as an agent of change, 1979. printing texts enabled new practices that improved scalability and hence the ability to compare “Handwritten letters could keep correspondents informed about their colleagues' activities only up to a point. Handwriting was less helpful when it came to conveying actual research results. Doubtless news that Kepler was working on a new set of tables could be conveyed quite efficiently by handwritten. But this point does not apply to the Rudolphine Tables itself. When it came to distributing hundreds of copies of a work containing long lists of numbers, or diagrams, maps and charts, or even precise detailed verbal reports, hand-copying was vastly inferior to print.” Eisenstein, The printing press as an agent of change, 1979. [Note: The Rudolphine Tables (Latin: Tabulae Rudolphinae ) consist of a star catalogue and planetary tables published by Johannes Kepler in 1627.] 19 Francis Bacon and Robert Hook on the new writing “… Bacon offered an account of how to make language transparent to the world it was to represent. He offered a discourse of things to replace the old, and as he thought barren, discourse of words. Old knowledge, even logical proof, he claimed, was merely verbal and may not correspond to ideas or to things: … In fact, Bacon discussed the new kind of discourse that he advocated as a kind of writing, what he called “literate experience” (Works, VIII.133: NO). Ordinary reasoning from raw experience is flawed, he thought, leading to anticipations and to imaginations and thus to allegory - relations based on resemblance and similarity. Literate or methodological experience, on the other hand, involves “the art or plan for an honest interpretation of nature, a true path from sense to intellect” (Works, vii.78: RP). In the past invention was done by [allegorical] thinking rather than by writing, but “ Now no course of invention can be satisfactory unless it be carried on in writing ” (VIII. 136: NO, I.ci).” Olson, The world on paper, 1994. “Robert Hooke… [ i ]n the Preface to his major work Micrographia (1665/1961), he sets out his goal as describing nature “as it is” … “... I indeavoured (as far as I was able) first to discover the true appearance, and next to make a plain representation of it .” (The Preface, p. 24) … [This] was sufficient to divide … the pseudo-sciences from the genuine sciences: astronomy from astrology, chemistry from alchemy, mathematics from numerology, predicting from foretelling and the like.” Olson, The world on paper, 1994. if a text was going to be copied exactly and circulated widely – and become part of a trusted literate resource - then it needed to be written in a new way – one that encouraged accurate representation. Cultural practices evolved to support this. 20 accuracy – one possible reason (among many) “The best maps, indeed, were often carefully hidden from view - like the map made for a fourteenth century Florentine merchant which was placed in a warehouse ‘secretly and well wrapped so that no man could see it.’ To make multiple copies would not lead to improvement but to corruption of data; all fresh increments of information when copied were subject to distortion and decay. This same point also applies to numbers and figures, words and names. Observational science throughout the age of scribes was perpetually enfeebled by the way words drifted apart from pictures, and labels became detached from things. Uncertainty as to which star, plant, or human organ was being designated by a given diagram or treatise - like the question of which coastline was being sighted from a vessel at sea - plagued investigators throughout the age of scribes.” Eisenstein, The printing press as an agent of change, 1979. “The passage of time … presented a counterfeit problem which ironically was even more effective than the real motion of the planets in fostering recognition of errors in the Ptolemaic method. Many of the data inherited by Copernicus … were bad data … Some had been collected by poor observers; And others had … been … miscopied or misconstrued during the process of transmission … The complexity of the problem presented by Renaissance data transcended that of the heavens themselves.” Kuhn, The Copernican Revolution, 1992. pre-printing text writing of copies was error prone 21 summary – the printing revolution: the emergence of literate practices from an information perspective, the emergence of printing resulted in little change at one basic level – the printed texts were still letters arranged upon a page. however, it enabled literate practices to emerge – allowing the production, checking and revising of historic and new texts to scale. A key element of this was significantly reducing the cost of producing multiple ‘exact’ copies at scale. this literacy-as-practices example shows how new practices as well as new technologies drive disruptive shifts. (looking forward, I will suggest that we are currently in a similar evolutionary pattern, where new practices are needed to realise the potential to move to the next stage of evolution. That there is a computeracy-as-practices pattern analogous to the literacy-as-practices pattern) 22 {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} text directions of communication writing printing brain-to-external (text) writing printing external (text)-to-brain reading (pre-printing) reading (post-printing) legend disruption Lorem ipsum dolor sit amet , consectetur adipiscing elit , sed do Lorem ipsum dolor sit amet , consectetur adipiscing elit , sed do eiusmod tempor incididunt ut labore et mapping the patterns onto the present: to understand it better exploring the past to identify patterns of information evolution mapping the patterns onto the present – to understand it better to give us insights into ways of realising the potential (in the future) evolution adding data to the facet framework for a contemporary perspective, we need to add the ‘data’ evolutionary stage to the facet framework earlier we made a distinction between data processing (islands – continents) and data sharing (alleyways) stages – we add these two as well 24 text oral pre-oral stages data writing printing text stages processing sharing data stages present (broadly) distinguishing data from text even though the distinction here between text and data is intended to be fuzzy at the borders, it is useful to articulate it typically, data is structured in a way that allows it to be processed. In our case, it is structured for processing by a computer a simple litmus test would be to try to process the data using a computer semi-structured forms or tables can be seen as limit cases – even when these are produced digitally whereas structured database tables or objects would be more obviously data 25 distinguishing data from text: one example of cultural practice early forms of writing are now seen as directed to creating a functioning system of visual communication and not an attempt to represent language. Cooper (2008, p. 83), for example, concluded that “no writing system was invented, or used early on, to mimic spoken language.” Representing language was a slow and largely unintended achievement: “ For the Egyptian and Mesopotamian scripts … a relatively full notation of language, including its grammar … was reached perhaps half a millennium after their first appearance ” (Baines, 2008, p. 150). Olson, The Mind on Paper: Reading, Consciousness and Rationality, 2016. among other differences (computer) data returns to information’s (representational) roots. as Aristotle in on ‘On Interpretation’ said: “Spoken words are the symbols of mental experience and written words are the symbols of spoken words.” In other words, language is composed of signs for things and that writing is a set of signs for these language signs. but writing was originally composed of signs for things – it was composed of thingologies rather than wordologies . 26 These three ways of writing correspond almost exactly to three different stages according to which one can consider men gathered into a nation. The depicting of objects is appropriate to a savage people ; signs of words and of propositions, to a barbaric people, and the alphabet to civilized peoples. Rousseau, Essay on the Origin of Language, 1754. and computer data is similarly composed of signs for things – in other words, the data refers to the domain, not a language for describing the domain. evolutionarily, a kind of paedomorphosis: phylogenetic change that involves retention of juvenile characters by the adult. computeracy as an analogue of literacy 27 what is computeracy (aka computeracy -as-practice)? computeracy is not just the ability to process data; it is the communal cultural practice of producing and exploiting data resources – more specifically it involves both collating and creating access to historical information as well as creating new information as shareable data. It is a property of societies rather than individuals. in a similar situation to writing and printing, the mere availability of computing technology is necessary, but not sufficient to produce computeracy . This requires the development of data sharing techniques and their deployment in cultural practices where both historical and new information is produced as shareable data. {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} text data writing printing processing sharing brain storage YES YES YES YES processing YES YES YES YES external storage YES YES YES YES ? processing NO NO YES YES ? communication brain-brain YES YES YES YES brain-external YES YES YES YES ? external-external NO NO NO YES ? adding (computing) data to the framework literate computerate we map computing/data disruptions below: we are only starting the shift to sharing, so this leaves open interesting questions about how wide the impact of the data sharing disruption – the transition to computerate – will be legend disruption 28 (electronic) data processing as formal data processing where data (formalized information) is transformed in a system in some useful way modern electronic data processing is characterized by algorithmic specifications of rules (in program code) for the structure of the data and carrying out the transformation process essentially making the system formal developing this formality for enterprise applications is still currently an error-prone, largely manual, process humans, used to working with text which does not need to be formal, find it tricky to get the specification right this formality can be seen as another major differentiator between text and data (in the sense we are using here) 29 (electronic) data sharing leading to semantic transparency data sharing where data (formalized information) is exchanged between systems – and ‘understood’ – and used different systems will typically have different (data processing) formalisations that need to be exchanged successful sharing/exchange translates (semantically) losslessly between these formalisations if the exchange uses an interlingua (or domain ontology), then this needs a common formalisation which accommodates the formalisations of the exchanging systems in other words , interlinguas (or domain ontologies) have a requirement for formalisation based upon the formalisation of the sharing systems they depend upon 30 to give us insights into ways of realising the potential (in the future) exploring the past to identify patterns of information evolution mapping the patterns onto the present – to understand it better to give us insights into ways of realising the potential (in the future) look at two data sharing practices let’s look broadly at possible future trajectories for two data sharing practices data architecture choices design approaches 32 data sharing architecture choices 33 point-to-point hub-and-spoke “a process of taking data structured under a source schema and transforming it into a target schema, so that the target data is an accurate representation of the source data.” Doan, Principles of data integration, 2012. transformation has insufficient space for a domain ontology a process of taking data structured under a source schema and transforming it into an interlingua schema and then into a target schema, so that, with each transformation, the data is an accurate representation of the source data. interlingua is essentially a domain ontology (at least in the sense of a common vocabulary) that the connections in a point-to-point architecture expands non-linearly with the number of systems, is often used as an argument for hub-and-spoke architectures. As hub-and-spoke architectures assume a common model (a kind of domain ontology), one can also take this as an argument for domain ontologies. for an early version of this architecture see: Gio Wiederhold , Mediators in the Architecture of Future Information System, 1992 data sharing design approaches 34 {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} point-to-point hub-and-spoke empiricist typically empiricist-lite less common approach rationalist not compatible common approach: attractive as lower initial cost both empiricist and rationalist approaches have IT pedigrees: {616DA210-FB5B-4158-B5E0-FEB733F419BA} empiricist empiricist rationalist empiricist-full empiricist-lite the existing data and schema the existing schema expert human advice design approach bases current practices Department Table Schema Data empiricist rationalist design approaches for a domain ontology 35 stages of design approaches for a domain ontology rationalist: empiricist-lite: extract the domain ontology from the systems’ schemas (probably in stages) - this will automatically generate a (schema) mapping then map the data into the domain ontology based upon the schema mapping - amend the schema mapping and domain ontology as required empiricist-full: extract the domain ontology from the systems’ data (and schemas) (probably in stages) - this will automatically generate a (data and schema) mapping note: in practice budget constraints and poor quality procedures often mean that amendments to existing artefacts are not sufficiently high quality empiricist: use experts to develop the domain ontology then map the systems’ schemas into the ontology (probably in stages) – amend the domain ontology if required then map the data into the domain ontology based upon the schema mapping - amend the schema mapping and domain ontology as required comparing stages choosing the domain ontology design approach 36 build a new formalisation then align with existing salvage formalisation refactor and align tested formalisations in existing systems build a new (untested) formalisation from experts’ advice creates a need to align new formalisation with existing proven formalisations a key difference between the approaches is the way formalisation (and the ensuing alignment) is handled one way of characterising the difference would be to say: the empiricist approach focuses on enhancing the existing formalisations with semantic transparency the rationalist approach focuses on building anew a semantically transparent formalisation – and then aligning it with the existing formalisations. there seems to be not a little hubris in assuming humans can accurately specify a working formalisation prima facie, the rationalist approach appears to incur substantial extra work – and risks. it is well-recognized that building a formalisation is difficult, and even more difficult to build from a blank sheet of paper the rationalist builds a new formalisation, when these already exist, and also incurs the work and risk of subsequently aligning the new formalisation with the existing ones, this seems unnecessary work and risk rationalist empiricist epistemic-full - parallels with writing 37 Writing Perhaps to say that is to speak adequately for poetic purposes — for metaphor is poetic — but it is not adequate for understanding the nature [of a thing]. Aristotle, Meteorology 357a24ff Where there is a disruption, there is a pattern of disparaging the inaccurate ‘representation’ practices of the previous stage. Printing all depends upon keeping the eye steadily fixed upon the facts of nature and so receiving their images simply as they are. For God forbid that we should give out a dream of our own imagination for a pattern of the world. Bacon, The great instauration, 1620 changing practices: adopting empiricist-full approaches as noted earlier, the shift towards a literate culture involved fundamental changes in practices we gave the example of Tycho Brahe comparing astronomical charts we have identified an empiricist design approach as a new practice one can see new practices emerging with this empiricist shift, for example: the focus on salvaging investment in current systems is relatively novel 38 bCLEARer – an empiricist-full process – stages {21E4AEA4-8DFA-4A89-87EB-49C32662AFE0} stages Collect Collect the datasets in scope in order to establish the broad scope of the process – establishing a bCLEARer master dataset Load Define the detailed scope by selecting from the Collect dataset the data in scope Translate the dataset into the cell-based format – the table paradigm Evolve Reveal the underlying semantics of the Load Dataset – ‘entification’ – in an ‘ entified ’ dataset Mine the ontology from the ‘ entified ’ dataset – the EVOLVE ontology dataset Assimilate Merge the EVOLVE ontology dataset into the full ontology model Reuse Publish dataset in a format suitable for the reuse context 39 collect load evolve assimilate reuse one concern might be how one might salvage the formalisation. there are examples, here is one: the bCLEARer process bCLEARer process – repeated sequence perspective 40 a repeated sequence of automated processes: a scalable way to systematically improve semantic maturity data collect collect collect increasing semantic maturity reuse reuse Foundational ontology example salvaging process 41 one way one might salvage the formalisation changing practices: extreme shift-left testing shift-left testing (Smith, L., 2001) ( Bahrs , 2014) ( Firesmith , 2015) is a known and respected approach. it is an agile approach much favoured in DevOps, in which one aims to test earlier than usual in the lifecycle. there, it is often described as based upon the first half of the maxim "test early and often”. 42 the empiricist-full consumption of data leads to a different way of working which creates opportunities for further changes in working practices one such opportunity is what considered an extreme version of shift-left testing: this is data-based testing (at scale) from the start of the development checking the data - parallels with writing 43 … Even when they were aware that they had inherited bad data, astronomers could do little to arrest scribal drift, and drifting texts were uncertain indicators of shifting stars. Moreover awareness of error was kept at a low level given necessary reliance on one seemingly authoritative corpus of texts. … Tycho soon mastered the use of these tables and perceived that the computed places of the planets differed from the actual places in the sky ... He even found out that Stadius had not computed his places correctly from Reinhold's tables. And already while Tycho was a youth only sixteen years of age his eyes were opened to the great fact which seems to us so simple to grasp but which escaped the attention of all European astronomers before him that only through a steadily pursued course of observations would it be possible to obtain a better insight into the motions of the planets. [68 - Dreyer, Tycho, pp. 18-19.] Tycho's 'eyes were opened' to the need for fresh data partly because he had on hand more old data than young students in astronomy had had before. … Tycho was himself a new kind of observer. … As long as accounts of separate stellar events were transmitted by scribes; as long as separate observers lacked uniform methods for placing and recording what they saw; as long as the most careful observers lacked mastery of trigonometry; how could falling stars be permanently located beyond the moon's sphere? pp. 596-8, Eisenstein. The printing press as an agent of change. 1979 summary 44 summary – past and present the evolutionary framework can be used to highlight: past: the shift between two stages of evolution: writing => printing this shift was driven by changing practices that evolved into literacy-as-practice present: point to similarities with two current stages in evolution: data processing => data sharing that the sharing stage is only just emerging the opportunity to shape an analogue to literacy-as-practice – computeracy-as-practice how domain ontologies are an obvious facilitator of this computeracy 45 summary – future looked at two possible ways of practicing data sharing architectures, and design approaches suggested a hub-and-spoke architecture practice is compatible with domain ontologies (and point-to-point less so) suggested that an empiricist (more specifically an empiricist-full) design approach practice had substantial benefits provided an example of such a design approach shown how this can lead to further positive changes to working practices (shift-left testing) 46 questions 47