archive-org.com » ORG » J » JEFFSUTHERLAND.ORG

Total: 379

Choose link from "Titles, links and description words view":

Or switch to "Titles and links view".
  • Online HTML Help
    slide Back to first slide View graphic version Notes Class attribute help in Javadoc format sweetened for end users Help and other forms of online hypertext documentation are created directly

    Original URL path: http://jeffsutherland.org/oopsla97/marshall/tsld031.htm (2016-04-27)
    Open archived version from archive

  • Views of Business Object
    created automatically from a Java class which is logical in that the class represents the knowledge that we have of that aspect of the business By centralizing this information in the class one eliminates the need for separate repositories for metadata specifications documentation help and so on Java reflection and introspection enables software tools to inspect and update classes for almost any purpose For example a business process and workflow

    Original URL path: http://jeffsutherland.org/oopsla97/marshall/tsld032.htm (2016-04-27)
    Open archived version from archive

  • OOPSLA'96 Integrating Java, Objects, Databases, and the Web
    embedded many rules and procedures for product delivery in computer systems the software applications that run the business must undergo significant change To gain the strategic advantages of speed and flexibility corporations must remodel their business processes then rapidly translate that model into software implementations Business Process Reengineering BPR sets the stage for continuous evolution of business processes to meet rapidly evolving business requirements Implementation of software systems that support BPR requires Business Objects that can both simulate corporate procedures and translate smoothly into software objects Well designed Business Object implementations can be easily modified as the business changes Business Objects as Reusable Components Early adopters of object technology asserted that packaging software in object classes would allow software to obtain some of the benefits of Moore s Law seen in IC chip fabrication Cox 1986 Some projects have achieved major productivity benefits For example a Maintenance Management System at General Motors originally written in PL I was rewritten under EDS contract in Smalltalk and achieved a 14 1 increase in productivity of design coding and testing Taylor 1992 Detailed analysis of this project showed 92 fewer lines of code 93 fewer staff months of effort 82 less development time 92 less memory needed to run and no performance degradation While there are many isolated projects that used object technology to achieve dramatic productivity gains during the past decade this success has not translated into broad improvements across the software industry In 1995 META Group reported that despite the promise of reusable objects most IT organizations have realized a scant 10 30 productivity improvement from object technology OT Failure to achieve larger productivity gains was attributed to Data centric task oriented application development Methodologies and cultures that do not promote reusability Few linkages between BPR defined business processes and IT support initiatives Meta Group 1992 While productivity gains from object technology in recent years have been limited some companies have been able to achieve dramatic returns on investment by bringing products to market sooner with the flexibility necessary for rapid tuning of the products to meet changing market conditions For example a recent analysis of return on investment ROI from object oriented development of robotics software by Marcam Corporation showed a 56 5M return on a 6M investment Return was calculated by multiplying the value of an improvement by the estimated probability of its occurrence and dividing by the cost of the improvement The following spreadsheet was generated Software Magazine 1996 Perceived Advantage Advantage Improvement Probability of Occurrence Incremental Cost ROI Time to Market 100M 4 3M 37M Flexibility 100M 2 2M 18M Productivity 2M 8 300K 1 3M Quality 1M 9 200K 700K Other Costs 0 500K 500K Code Size 0 0 0 Reuse Requirements 0 9 10K 10K TOTAL 6M 56 5M Table 2 MARCAM Return on Investment for OO Project corrected Business Objects are designed to support a clearly defined relationship between BPR defined business processes and software implementation of these components Using an object oriented development methodology yields quick time to market and good object oriented design allows for rapid evolution of Business Objects in response to market conditions The bottom line is that object technology is a necessary but not sufficient condition for large returns on investment It must be combined with focus on delivering Business Object Components that enable fast and flexible delivery of new or enhanced products in the marketplace The Need for a Business Object Architecture As business models are renewed software architectures must be transformed A Business Object Architecture BOA is an effective solution for dynamic automation of a rapidly evolving business environment Dynamic change requires reuse of chunks of business functionality A BOA must support reusable plug compatible business components The two primary strategies now being used for implementing client server systems to support reengineering of business processes are visual 4th Generation Languages and classical object technology While both of these approaches are better than COBOL neither of them can effectively implement Business Objects Building Business Object Components A group of objects is the ideal unit of reuse These groups of objects should behave as a higher level business process and have a clearly specified business language interface Business components are encapsulated with a protocol that allows efficient communication with other objects on the network Consider a typical client server application like an order entry system This system takes a Purchase Order as input and produces a validated order as output The internals of this component should be a black box to the external world The resulting order is input for another subsystem or alternatively an exception condition is raised if the Purchase Order is not valid for processing Figure 1 An Order Entry Business Object To support plug compatible reuse a business component must be encapsulated in two directions The external world must not know anything about component internals and the internals must not know anything about external components other than allowing interested objects to register for notification of specific events or exception conditions The internals of a business component are made of other encapsulated business components For example when a Purchase Order passes through the membrane of the Order Entry business object an internal component must see it validate it look up customer information inventory availability and catalogue pricing and build an order that is consistent with business rules and procedures Each of these tasks is accomplished by embedded components many of them communicating with external data sources External databases must be encapsulated as Business Objects or reuse will not be easily achievable There must be a database access component that causes values from any kind of database to materialize as objects inside the business component Whether object oriented relational or other database access is required a set of class libraries designed to automate this interface will result in a major savings in development resources Sutherland et al 1993 Vendors producing products in this area include Persistence and Ontos for C systems Riverton Software and Vigor Technologies have

    Original URL path: http://jeffsutherland.org/oopsla96/webapps.html (2016-04-27)
    Open archived version from archive

  • Part 2 -- Simplification (Product and Project)
    are evidently of variable granularity some can be composed from others That gives us the first kind of interrelationships between REs View composition is carried out by RE methods We can also define interrelationships such as ordering Balance isLessThan CreditLimit or mathematical ones Y isSineOf X using RE method algorithms that do indeed hide some basic realities which we may happily regard as real complexities which we may not wish to express further in our information systems We accept ignoring totally any further complexity that may be lurking there But the entities themselves X Y isSineOf typically also have many abstract properties of their own Those are all in the layered zone RE methods validly hide complexity But where there is even a possibility of interconnection it is better to go first into the layered zone and model the detail there Thus RE methods composing views should be compilable after composition from more logical rules in models in the layered zone in terms of application aspects such as relatednesses or relevancies e g unique names are useful for visually verfiying codes and more generic aspects such as styles and widget natures On the agate model once compiled and executable under appropriate circumstances they then belong to the crystal zone but their meaning is derived from the particular boundary and related layers which define and model the application or execution context Rather similarly state machines handling i o and other events are almost always best kept in the layered zone with only the thinnest or most elemental interrupt handling kept on the rough surface that is logically on the rough surface though bound in as inner crystals In that way MACK enables a decomposition of the total i o function into naïve realtime physical i o events triggering the application of operations imposing the corresponding relativistic realtime perspectives at the appropriate respective times Similarly again legacy file assimilation into MACK compliant form like canonical version migration is extensively modelled before being assembled into the appropriate user meaning handlers encapsulating user semantics in efficient RE method shaped code There is also much scope for refinement and elaboration around the outer boundary For example what are the relationships between Balance isLessThan CreditLimit and X isLessThan Y And how do the storage aspects integrate with those abstract relationships But now we are getting uncomfortably close to the MACK trade secret core so I pass on In Metaset there is at present no strict control over the good behaviour of the C coded RE methods That is controllable in the short term For example an untrusted RE method can be run by a stub in another process with pre and post conditions added automatically In the medium term we can be sure appropriate alternative coding methods will arise perhaps analogous to Java s sandbox There is no great problem here as RE method functionality is typically highly refined by the abstraction out of generic layered zone functionality For example there is no room for any OO here as all abstraction and refinement is handled via the agate layers openly and reflectively accessible However we generate the inner crystals that in run time represent our formal RE meanings that is whether manually or by binding and code generation they always operate within a well defined topological zone of the layered semantic net that is semantically adjacent to the outer reality boundary concerned That is conventional thin coupling The art of model design including its associated RE and RE method design is to ensure that such semantic topological coupling is maximally invariant under likely model manipulations and transformations That is the key to the stable reusability of RE methods and their defining layered zone models That might sound complicated and difficult but all of it really does have plain and familiar counterparts It requires the same skill sets and habits that any OO designer finds natural though not the structures or tools Except that with MACK the realities of constraints and facilities work out manageably with many familiar problems naturally resolved and further opportunities offered as we shall now see in some further examples Most particularly it is evident that distributed function calls or operations on typical coarse naïve realtime Classical Object Model class encapsulated methods are highly unlikely to exhibit such semantic topological invariance whether with refinement or change and are therefore the very opposite of prima facie candidates for casting in crystal form That explains the failure of RPC as the basis of distributed architectures They simply try to encapsulate far too much and inevitably oversimplify too soon That is particularly evident of course when the respective agate models are changing That explains why the OMG s OMA based architecture has repeatedly failed to provide for versioning and migration Evidently the dynamic multi perspective creative value adding that is continually taking place is an effective MI Multiple Inheritance One immediate and important example is in resource contention management Every entity may dynamically become an instance of a resource type and will do so when the reflective activity management detects possible conflict Since everything that happens in the agate layers is open and transparent to the kernel with only the irrelevant hidden details of RE methods not being so there is automatic reflective variable granularity predicate locking Because of the integrated database there is no problem in having multiple mutually independent predicates operative at any one time There are clearly a number of tricky aspects to it but no intrinsic complexity that cannot be expressed in user meaningful terms by the application of further reflective patterns The result is an efficient and natural resource contention management with minimum conflict and requiring no user programming at all Generic design patterns can provide for any eventuality in open and reflectively manageable ways without undue user confusion Another application of dynamic MI is the insertion and removal of measurement systems for the tuning of any operational facet Since there is no need whatever for any tiering in this integrated db environment all storage management including such matters as garbage collection physical contiguity optimization locality index management full text inversion the representation and management of distributed entities and the creation and management of redundancy for access speed or resilience purposes is likewise handled in an MI enabled reflective integration with the relevant RE methods in the appropriate models with extremely high reuse of simple generic RE methods and in terms of parameters presented to the user in appropriately simple policy terms Inter database messages physically handled as transport level i o carrying well defined contexts and goals do not need any RPC like namespace concept Hence there is no problem of misinterpreting the meaning in a name as all such matters are isolated in terms of simple REs The key here is semantic specification managed common knowledge of a fully canonical and far reaching kind particularly well integrating and leveraging MACK s inheritance equivalent still trade secret It also integrates tightly with the entire relativity aspect and is further the basis of interfacing with legacy data Plain information systems such as the groupware that is our immediate marketing infrastructure target have the fewest RE methods That means that application development in that environment will generally be able to reuse those available in quite small MACK realizations such as Metaset v 1 and will not require any further RE method coding hence no procedural programming at all and only simple non procedural model specification How does MackWeb document or message preparation contrast with the Old Web s In one word finer grained That gives far more useful prompting or suggestion of content hence better reuse greater control over the details and far more dynamic reconstitution of the full message in terms of perspectives and information relevant to the reader s HTML XML s Unix text stream roots are a quite unnecessary entanglement most unsuited to a more interconnected world As application size increases how much functionality and intricacy will be handled by the automatic logic within the layers and how much will require RE method code to be written The agate gives us a lead As a sphere s size increases its volume grows with the cube of the radius while the surface area only grows with the square That simplistically quantifies what is obviously the case with increasing application size and integration the interconnection aspect would far outgrow the atomic reality contact aspect as ever more REs would be reused But I don t think it is wise of me to try to sketch too much of these applied architectural aspects considering that the fine structuring techniques and most particularly its precise Beyond OO mechanisms are still MACK trade secret I shall merely refer you again to the relevant section of my 1997 paper and its link re the trade secret aspect Remember once more that the typology of my earlier papers is here just model Also see the 1996 faq question 7 reply on how the model is the medium for the application of patterns There is the important message that the MACK compliant world will be very dependent on good pattern people for the creation of suitable reflective models of the kinds so often referred to above It further points out how they will find the MACK environment a very gratifying one in which to hone and ply their talents Their product cycle time in the groupware assisted open market will often be measurable in hours Finally it is useful to summarize by pointing out that there is no room for any Divine Programmer in such dispensation He occupies merely the thinnest boundary zone while his products exist in and are managed by a strict and transparent framework of far more easily and even reflectively manageable logic representing the relativistic world in which his compositions are to operate And that big link represents the closure of that theme and our escape from the Divine Programmer Syndrome And while it would be nice to hear some applause penetrating through the fog of my own words that is an unreasonable expectation but at least we there have a closure of my most immediate invitation to you But all that is still theory though it does all seem very obvious to my programmer s mind Now comes the real practice where the proof of the pudding is becoming ever more evident as already stated I shall assume for the rest of this Part 2 that all the above together with Part 1 is at least a consistent and coherent indication that MACK has a good epistemological basis particularly regarding the position and hiding of complexity has coherent and consistent practical consequences down to the finest technical details of implementation seems to have taken at least many of the most notorious difficulties in application and system design into account and most importantly can be implemented in a basically simple event driven kernel on top of a plain conventional operating system host with good efficiency thanks to managed building in of fixable specifications and to indefinite user extensibility in canonical ways I am asking quite a lot of you there of course but in the light of the trade secret requirements I can merely trust that the whole credibility argument in Part 1 including its link to my own IT biography will support me enough for you to be able to regard the remainder of this Part 2 with a sufficient degree of suspension of disbelief The status of Metaset programming The programming is all in C programmed using Microsoft s Visual C development environment on Windows NT 4 0 but using the very minimum of the features of the development and runtime environments in rough terms define WIN32 LEAN AND MEAN and include WinSock The APIs used and their evidently underlying entities and properties are easily modelled canonically to facilitate transformation to different windowing systems object models and porting in the eventual canonical bind and code generate way Metaset is at present a program roughly an internal version 3 which represents a good portion of the inner quartz crystals of a basic boot or seed agate though in the form of manually bound RE methods That is they are not all formally organized in the layered zone yet where they will be logically situated at the outer boundary though they are structured in anticipation of it At the same time the inner models of the layered zone of the agate are being built up and stored in the integrated database all very MACK canonically again and being retrieved and printed out in various interrelated views The design especially after a rewrite that took place this year is very tight yet provides for smooth canonical distributed independent supplier market driven growth towards all the dynamic multi user distributed transformable scalable etc qualities that are required for the full groupware and market infrastructure application environment targeted with 24x365¼ operation The runtime efficiency is excellent so far All the C is structured with a view to eventual formalization as nice simple RE methods As a result the programming has several interesting qualities There are no non canonical work fields hence no problem of hidden conflict or mismatches and no memory leaks In fact all memory allocation is done by Metaset in very big quanta themselves managed canonically and then managed internally in a canonical way As indicated in the boot product description and further details in my 1997 paper as far as the user is concerned the environment is fileless with user data management integrated in application appropriate ways That enables very extensive and reliable use of C indirection facilities partly explaining the speed of processing as well as the virtually complete absence of lost pointers during the programming itself The present focus is on building up the inner models mentioned and it is anticipated that after not much more work there the long anticipated self booting nature of the MACK model will begin to manifest itself in an increasingly spectacular exponential way But that work is quite difficult at this stage involving painstaking assembly of the kernel models which are the closest thing in MACK to meta metamodels and the very ones that will make that kind of work easy for everyone ever after So I am reluctant to give estimated launch dates in the continued absence of a synergetic team of the kind of which I have had many most gratifying experiences funded and managed by a competent and committed organization The latter is something which at this stage I myself can neither afford nor manage without removing myself from the programming which I am still reluctant to do As a result as indicated under 1995 in my IT story I have been working in solo mode with some signal success as related there but the time is now ripe for some synergy considering the increasingly clear nature of the entire design in itself for me a highly convincing proof of the pudding But before we can consider how the project might be taken further with a view to such synergy we have to consider how MACK and the MackWeb will transform the software market Commercial incentives for developers in the MackWeb The key commercial opportunities are twofold the creation safeguarding and transfer of intellectual assets and the provision of related services The MACK agate model implies an enormous scope for the latter but relatively a shrinking scope for the former Why is that The question is key as the consequences to our industry are evidently going to be very radical It all derives from the relative proportions of surface to interior as size grows as already commented Interconnectedness starts to predominate ever more However that does not mean that the surface will shrink The surface is of course where the asset of mastered complexity is appropriately hidden in a conventional way in RE methods Quite the contrary it too will explode as total application volume grows exponentially I O device drivers multimedia handlers mathematical techniques everywhere standalone external processes being managed by and being fed data by MackWeb applications all that conventional coding and commerce will still be there Furthermore all that business will even be greatly facilitated by the MackWeb data management and market infrastructures The MackWeb itself will in effect become a distributed application operating system to a significant degree much as Old Web browsers are at present tending to become as especially Microsoft is insisting And e commerce will be implicit in the burgeoning MackWeb market infrastructure with its concomitant supply side and demand side stimulation So that business will be even greater The same surface argument will apply to services too as user contact is part of the reality interface Service growth too will be greatly facilitated and further transformed by the new medium No the real significance is in the growth of the interior of the agate and in its nature because that is where so called complexity is no longer hidden Transparent simplicity will be pervasive It will all be revealed or revealable because it will be more simple to manage see and understand but more importantly it is interconnection and the prime interconnections are between parties in the market There is even an existential quality to that primacy That is of course mere cliché from The Mainstream once more the open market requires transparency for most rational individual behaviour and optimal overall satisfaction and efficiency And a more infrastructure facilitated and automated market requires it even more It is after all already common cause that component reuse requires a full feature component marketplace Despite a possible surfeit of opportunities though the user will not be confused thanks to better view design and more needle sharp marketing in general thanks to the MackWeb s more powerful projection of The Mainstream trends towards personalized marketing thanks to MACK s more appropriate approach to complexity mainly thanks to its model based relativity with less obscuring of relevant details

    Original URL path: http://jeffsutherland.org/oopsla98/SpottSimplification.html (2016-04-27)
    Open archived version from archive

  • JSOT: Business Object Architecture Tiers
    object model component model development process use of common patterns and common tools is essential The target must be clear and the entire organization must be incented to migrate towards the target A core development team must mine existing code

    Original URL path: http://jeffsutherland.org/objwld98/reuse.html (2016-04-27)
    Open archived version from archive

  • Ride The Mainstream
    may make our escape as Odysseus did on his mast and keel But there is much more to it To see how the image and its details can be interpreted in many further interesting and relevant ways see at these points in my 1996 paper with its own further link in which also find Scylla or Charybdis and in my 1997 paper and later reflections in which find complexity for where the classifying in Scyllan and Charybdian terms starts The image was created by Homer I allege with that specific intent in bold above Why was the message was not spelt out more explicitly or directly Any epistemological or philosophical issues to the degree to which concepts could at the time identify any such matters could only be cast and experienced in metaphors that we like to call mythological We may recall that Homer whether a man or a lineage represents the story telling medium for an age old wisdom passing tradition That largely explains how his two full works have survived as the bible of their times while only the merest fragments from the much later pre Socratic Greek philosophers are known Demythologized however the interpretability of the allegory is uncannily detailed The image and its terms remain as modern even postmodern as one can find That is why the monsters the whirlpool and the figtree are depicted on the cover of my book introduced below But the Homeric image portrays mainly the human implications of dealing with complexity We still need one which more usefully represents the more formal or symbolic aspects Hence also from September 1966 a twin image that of the agate It seems difficult to credit and you can see why I did not dare mention it in any of my previous papers considering the well known alchemist phase in the history of what people have at various times liked to call philosophy but the model for the MACK knowledge model is actually a philosopher s stone The agate image portrays the structure and growth of symbolic knowledge to an amazing extent It has remained with me as a usefully elaborated portrayal of abstraction as applied to the real world and has much direct relevance to any computerized knowledge models It has subsequently even become a detailed model for MACK It is described in detail below and then applied in Part 2 But where did those images come from Under what circumstances would a 25 year old even have any such need Are such pictures not merely arbitrary or chance metaphors More importantly are they not themselves restrictive or misleading further Scyllan dangers lurking uncomfortably close to the unwary philosopher or knowledge architect or system designer To answer the first three of those sceptical questions I can safely assert that the images most definitely grew in the most relevant environment In 1963 aged 22 I arrived in Cambridge England to do research in pure mathematics But as an English speaking South African We typically saw ourselves on the centre left of the political spectrum who had just spent over 4 years at an Afrikaans medium university once the home of apartheid s creators and at the time attended by their children I also brought with me many close up experiences of a very different political outlook So in November of that year a number of individually commonplace matters came together in my own little mind 1 the lesson from pure and applied mathematics that pure abstract systems have a life of their own independent of reality and that real things once we admit them in symbolic terms when applying mathematics go the same inexorable but often useful though thereby simplified way 2 my own observations of how apartheid s so well intentioned theorists like any social engineers put people reality through the same kind of conceptual mill 3 the frustration of my own persistent attempts at explaining the latter behaviour in the naïve hope of thereby changing it and finally 4 the sudden realization that such explanations are themselves error prone for exactly the same potentially oversimplifying reason Those explanations were in evolutionary terms as intimated above and clearly beg their own questions Very obvious you might rightly say but that very meta realization as the prototype of seeing Scylla and with its rather intense history was of course my own small Damascus experience with the usual personally significant effects It immediately led to my dropping all thoughts of research in pure mathematics and concentrating on a book to be entitled The Phenomenon of Knowledge in attempted emulation of Teilhard de Chardin s The Phenomenon of Man But if you see Scylla can Charybdis be far away So that plan grew via the 1966 images and the course reported under 1975 in my own more detailed story until it finally appeared under the title Beyond Apartheid in 1986 See also 1984 and 1985 There is a further epistemologically and autobiographically relevant aspect of the simplification process as embodied in the Homeric injunction Hug Scylla s rock and related recommendations We may build on the extraordinary degree to which that episode like the others too may be read as an epistemological lesson Hug Scylla s rock means of course that we have to keep on the simplification side if we are to avoid the fearful Charybdian whirlpool of perplexity by complexity despite the tragic fates I have never had to witness a more pitiable sight than that of the inevitable oversimplifiers as Scylla with her six heads snatched and munched six of Odysseus crew and the ablest hands amongst them too Cliché once more Mere reductionism of one kind or another But further Call on Cratais Scylla s mother who brought her into the world to prey on men She will prevent her from making a second sally That is if we can try to understand the real nature of simplification and oversimplification we are better protected against its further excesses Ancient yet simple and wise advice for perennial problems And there is far more in the Scylla episode in that very connection that I shall not even mention here as it is not quite as immediately relevant The most basic kind of simplification and one that is incorporated in detail in its simple axiomatic terms into the MACK reality model and is presently taking programmatic shape in Metaset is the seeing as or interpreting in the context of process whereby any experience or thing may be seen or interpreted differently by applying a different conceptual schema to it See the vignette in the box That represents a creative and conceptually synergetic step and is the very basis of how MACK works in practice It is the operation in a semantic net given precise meaning by its relativistic context Creativity is a further application of the normal simplifying processes of perception or scientific explanation Seeing as is the key complexity filtering process throughout Thus an otherwise fleeting sense of colour and shape we see as an apple in a tree A response by a person we may see as a sign of love The frog sees a grasshopper as food though only when it moves Thus while we simplify by filtering out inessential noise we also add meaning thanks to the context in which the observation process places the experience or through which we interpret the input Our preconceptions may be supplemented by yet further ones Newton saw gravity when the apple fell Love may obscure what we also come to see as an undue element of self interest New conceptual interpretations may supersede old ones as in falling out of love or as when with a wholly new kind of perspective Einstein with his General Theory of Relativity saw the phenomena of gravity rather in terms of some larger conceptions of time and space and energy Or as did the Athenians when the citizenry rather than the ruler was regarded as the proper source of political authority Their vague inner sense or intuition of justice which the classical Greeks are well known to have felt about very strongly and very extensively came to be partly seen in democratic terms Any experience or any thing may be seen as one or more of a potentially infinite number of further things we need only liberate our vision and escape from our habitual contexts into new mental planes Such is creativity whether scientific artistic or humorous as presented by Arthur Koestler in his magnificent book The Act of Creation Hutchinson 1964 Koestler s picture was for me a signal lesson in that year as I acknowledged in my 1996 paper and 1997 reflections Find Koestler there When new conceptual schemas for all their abstraction thus add value to our real experience it is in a synergy that further enables our very lives and can indefinitely enrich That is also the very basis of the profound congeniality as distinct from mere user friendliness that MACK will help us cultivate in our natively ignorant computers as already much mentioned in my earlier papers But let us return to the other sceptical question about those images As metaphors they may be most highly relevant but have they not been taken too far Does their use not oversimplify in its turn Certainly those far back dates do situate me squarely in the silly Sixties that climactic period of many grand dreams which have since turned to dust So what considerations might exempt me from dismissal as a simplistic fool of one of the many kinds that have come to characterize that decade from the misled would be mystic through the ineffectually idealistic to the dangerously technocratic Maybe on the other hand the answer lies precisely in those 1966 events which have left me with all these admittedly simplifying images How plausible after all that the Homeric image of Scylla and Charybdis so graphically and fully capturing the Sixties temptations might have helped me survive the rest of the Sixties syndrome which is basically Scyllan or oversimplifying In another simplifying contrast if the subsequent decades of world history might be characterized by the resulting confusion and chaos on which the denizens of static since overcomplicating Charybdian figtree feed would it not make sense that such a widely applicable image from the objective phenomenon of knowledge might have proven a good mental stabilizer and star to navigate by It would not be surprising if hugging Scylla s rock and calling on Cratais have helped here too And so you may see how actual events have unfolded as a story of an ever growing convergence with what I am asserting has transpired to be The Mainstream of history which history of course includes that of software architectures As merely a few examples classical OO with its Divine Programmer and naïve realtime roots are Scyllan UML and Big Methodologies are on the Charybdian figtree with process still outstanding while MACK follows Odysseus mast and keel formula for eventual escape from both with process implicit The continuity and the coherence as the story and the metastory converge are massive The story relates actual events which as the metastory significantly extend the compelling objective evidence in favour of the proposed MACK standard with its implementation project The story also builds up to the same process pursued in the following section in respect of the already published MACK papers So in the separate document linked to below there is a rather blow by blow account of Metaset MACK s further history But why do I risk wasting your time with such subjective trivia I try there to show that many of the lessons learnt have been highly relevant to the design and implementation of something like Metaset with its underlying abstracted architecture I also hope to learn from you whether by direct communication on the gaps you spot or in due course by yourself moving into the MACK compliant market collaboratively or competitively And in the short term it also functions as a resume or CV in support of my application for some role in any MACK promoting or implementing project you might put together As we shall see in Part 2 The story also shows how the development has been guided by the original epistemological lessons and images with each aspect buttressing the other In addition the MACK background is of course not really just about South Africa as will hopefully be evident from the detail I recount But the South African aspect is relevant too Firstly MACK is about human knowledge generically and certainly aims to have global applicability Together with awareness we can better abstract the universal aspects into it Secondly when it later comes to the full marketing of compliant products especially for the needs that evidently drove the design at first South Africa is an excellent microcosm in which to test them In its groupings histories disparities and shared resources it is a globally relevant and almost universally eager laboratory for change and as a proving ground could quickly give results that might be scaled up or translated elsewhere We are in fact already used and have long been used as guinea pigs by a variety of global companies test marketing high tech products Therefore it might help for you to see from the separate document how few of the IT related lessons of MACK s history are after all specifically South African while yet not ignoring its South African taints and qualities But why is it necessary to bring South Africa or a South African into the picture at all Can t MACK be argued and pursued objectively on its own merits Unfortunately not yet very well and only partly because as this paper surely shows quite amply it is such a big and still poorly described scene Metaset in its present state cannot help out either in illustrating MACK s qualities and potential in a widely accessible way and certainly not without giving away many of its still trade secret details So it might help for you to know more about where MACK comes from I am not just being embarrassingly narcissistic there are so many opportunities and roles for you too The Mainstream would make no sense unless it can be widely ridden and the sooner other people can help make it so the better for us all That is my premise On the contrary therefore I have tried to draw objective rather than subjective lessons what I have described is a course that has long been converging on The Mainstream as it has become increasingly apparent in far more recent years The story might also help you spot more clearly those areas where I am still lacking as I am also asking you to help me Do please therefore read this seemingly irrelevant story with some care matching it against your own experience and abilities to comment on knowledge architectures In that connection do also take to heart the opening limerick of my 1996 paper Surely also if indeed there is any merit in MACK it is thanks to the evidently rather unusual course by an otherwise usual person in unusual circumstances which has made it possible A significant degree of confirmation from published MACK papers To start with I may point out that this paper my other ones linked to and their further links to my pages on this site are very consistent with one another Reading the earlier papers might also help consolidate in your mind the various Metaset MACK backgrounds and concepts Just one wording change to remember the typology of earlier papers is here simply model More importantly in the light of subsequent events they also constitute some significant confirmation of my thesis with only minor refutations The most evident major confirmation is the way that software architecture directions as I have conceived them have remained the same Even the most sceptical can regard that fact as significant in view of the rapidity of change as it is generally perceived in our industry So at the global level there is at least a good case to be made for my assertion that it shows that I do have my own meagre finger on the pulse of The Mainstream in our industry at least My papers all tell the same consistent story while the only things that have been changing are the current events to be commented on and the new ways I have been trying to get the whole big message across That comment applies even more to MACK as it has long been evolving but those autobiographical facts are not so easy for me to prove to you though the metastory in this paper does try to convince you of it through force of overall consistency and coherence However there are some specific aspects of my accuracy that are easily pointed out For example when my 1996 paper was published I was unaware of XML and I first learnt of it from another paper for Jeff Sutherland s 1997 BO Workshop Mark Baker s Revisiting Sims See also here My second introduction is built around the remarkable convergence between longstanding MACK and that particular current in The Mainstream Thank you Mark Thank you Oliver Sims too The current of user definable semantics has in fact run for quite long already Tracing it backwards one might mention the lineage of at least repository data dictionary self describing file db schema and cobol copylib Neither XML documents nor Sims semantic packets however come close to MACK s pervasive approach and framework In general there has also been a bluntening of industry expectations even a deepening of disillusionment in respect of all conventional OO such as CORBA and Java and indeed most of my documents have consistently poured scorn on that whole scene Hence my MACK BOO stance MACK is Beyond OO which I started adopting in 1992 After OO what else is left Components is the new life jacket However loudly though marketeers have been trying to push it strongly though market demand has been trying to pull it and much though it works not badly in view design it cannot escape the abysmal present OO and its poor inheritance that is meant to give flexible reusability That is perhaps the most fundamental of all the good things that are presently missing from conventional technologies and infrastructures and is consistent with my various criticisms of conventional OO Most particularly in my 1996 paper and its faq q 1 reply and my True Love play Unfortunately of course at this stage it is difficult for me to convince you of the better alternative as MACK s inheritance equivalent is part of the trade secret core of MACK that I am not yet disclosing On that aspect see further in Part 2 In due course though you will not be able to stop me from enthusing about how just right MACK s abstraction refinement feature is how it pervades core MACK and how responsible it is for the impending explosion of reuse in the MACK market It is even so simple But the fog of complexity hiding is at present obscuring its possibility In all my papers the easiest potential refutation of which I am aware concerns my quantified time predictions for Metaset MACK s arrival and penetration into wide use in the market Thus in my June 1997 Background document I had predicted that it will be between one and two years from now that MACK will convincingly replace the OMA and DCOM HTML Java Beans etc as the architecture of choice for Internet leveraging applications I believe there is still a fair chance that that one will come off within the longer period and not the shorter one seeing that my sooner and better tactic of that document has not worked But in my 1997 paper of August it was as if I had expected that sooner and better tactic to materialize During 1998 d v Metaset will start quickly bootstrapping the MACK compliant market and will begin hosting massive compliant application development by ISVs Well maybe all I can say there is d n v in respect of that tactic You and I still need a bit more time You might well say that the time prediction is the only kind that counts To that I can only beg that you put yourself in the shoes of the manager who knows that the programming project is behind but still has many good marketing and technical reasons for not being unduly concerned by such a situation rather normal as it is especially for pioneering applications On the other hand there is confirmation in the way the Metaset programming has been turning out While I am of course disappointed that my poor concentration on it has not allowed it to progress faster the shape it has been taking is the most beautifully tight one that I have long been predicting For example in the question 12 reply in my 1996 faq I had expressed a certain confidence The basic MACK model is really so simple and so general purpose that there isn t all that much programming to the Boot Product s kernel After this year s rewrite that aspect is stronger now that I had thought it would be But I am keeping those details under wraps still so in this paper I shalln t insist or count on that further However there have certainly been no refutations and only confirmations of my far more deep reaching and detailed predictions in the OMG finds True Love play published June 1997 concerning the OMG s BOF quest You may also go there via Jeff Sutherland s kind introductory warning on his OOPSLA 97 Business Object Workshop III index page I must preface my review with my own apology for seeming rather small mindedly to say I told you so to some individuals for whom I have the highest respect and admiration But this type of argument is far more accessible to the average reader than much of the rest of my writings so I beg your indulgence and particularly that of the said individuals for this different kind of attempt at conveying the vast coherence yet correctness of all my wide ranging argument I had also in anticipation prepared for this very situation with these words in that play under the signifying nothing subheading and referring to the widespread misconception of encapsulation and complexity hiding here characterized as the fog Though I hasten to add the idiot here is not any of our characters it is the situation with its constraints Such of course is the best dramatic tragedy marvellous people ineffective caught in a naturally misconceived situation In that play under the Some key details heading I referred to the OMG s current inertial and predictably tragic course Then under the subheading Thus all five BOF submissions must be dismissed I predicted that even if only for the secrecy reasons given the OMG would not choose the MACK alternative with the result clearly referring to the outcome of the BOF RFP that the OMG will have a tragically unnecessary and messy short term way ahead In the event that RFP was withdrawn at the OMG meeting in Helsinki in July 1998 and quite possibly at least for the very reasons which the True Love play also indicated as we now see In the play I had criticized all the BOF submissions for their lack of synergetic and exciting semantic substance inside their hollow shells and how could it be otherwise as they were all cast into forms of insipid and indigestible IDL Nonetheless I had praised the JBOF submission for failing less badly than the others so I had bet on JBOF In the event JBOF did become the front runner evolving partly into the BOCA mentioned below that was so very nearly adopted in Helsinki At the time that play was first e mailed early May 1997 the OMG s Analysis and Design Task Force s RFP had recently been answered by the initial Submission of Rational s UML There was clearly so much overlap with the very concept of a BOF that I could not resist this paragraph in the play where just in case you had wondered ArchiBoard is the character representing the Architecture Board of the OMG while the castle where the Grand BOF Banquet is being planned represents its dogmatically imposed OMA In addition there is both injury and insult our embattled ArchiBoard is apprehensively eyeing his brilliant and friendly neighbour Maestro Rational Frictions are looming over the Maestro s venerable UML tree Its deep state roots have long been opening cracks in the castle walls spoiling its ideal beauty and now its spreading branches are dropping leaves into the BOF dishes Later in August 1997 the final UML Submission appeared UML 1 1 with an entirely new addition OCL Object Constraint Language It not only significantly added to semantics external to IDL but in its extensive reliance on class invariants it is state based rather than behaviour or interface based The interface basis of the OMA and hence of the BOF were of course the play s big bête noire Even later at the OMG s Salt Lake City meeting in February 1998 the BODTF Minutes report this Jim Rumbaugh said that BOCA is at the same level as UML when you look at it You can map between them but you don t gain anything because you re not shifting levels I thought the idea of this BOF was to raise the semantic level And indeed the August 1998 replacements for the original BOF RFP the first being this all concern a requested UML for Business Objects So it is appropriate to end this episode of the story and metastory with this quote italics now added from the True Love play which immediately followed my already mentioned bet on JBOF rather than on MACK for the BOF That said as time passes I am expecting the perceived risk to fade in any minds that can persevere and piece together all my argument image and confirmed prediction That picture s very high degree of coherence and growing cogency is I guess not so obvious at this stage After MACK s launch if necessary without the teammates hereby sought the whole ballgame changes anyway as the completed picture s clarity will indeed shine out as the targeted complexity simplified Hopefully more convinced now that MACK is on the right kind of track despite the horrendous complexity and the associated Scylla and Charybdis Syndrome of which the Divine Programmer Syndrome is part and despite my own expository fog let us at last get on with the actual design of MACK and Metaset The point of MACK and a high level lesson from the epistemological images At least the impetus and shaft to the point are very much in line with the familiar mainstream Consider the delightfully apt very first paragraph of Chapter 1 Complexity of Grady Booch s Object Oriented Analysis and Design with Applications 2 nd ed 1994 Benjamin Cummings A physician a civil engineer and a computer scientist were arguing about what was the oldest profession in the world The physician remarked Well in the Bible it says that God created Eve from a rib taken out of Adam This clearly required surgery and so I can rightly claim that mine is the oldest profession in the world The civil engineer interrupted and said But even earlier in the book of Genesis it states that God created the order of the heavens and the earth from out of the chaos This was the first and certainly the most spectacular application of civil engineering Therefore fair doctor you are wrong mine is the oldest profession in the world The computer scientist leaned back in her chair smiled and then said confidently Ah but who do you think created the chaos Booch then reviews The Inherent Complexity of Software with its features and causes so by page 15 he can stand back and reflect that there is a factor that dominates the fundamental limitations of the human capacity for dealing with complexity How true But the final point to the thrust of The Mainstream is both ancient and refreshingly simple we deal with complexity better by means of simple systems not complex systems Mere naïve innocence and persistent folly Not really Thanks to the present state of IT we may surprisingly easily follow The Mainstream of history We may project and take it into the future in a very consciously fine tuned fashion as the natural way of dealing with complexity by means of simplicity We need merely better pursue The Mainstream of the evolution of species and artifacts We need demand no great adaptation by our users We shall cater to them as they are The right seeds already exist and the good ground is in place The Mainstream is not some Utopia It is where we are already and always have been We need merely see through the fog complete a bit of non divine programming then everybody will soon ensure that the mess is cleared up and we shall really Ride The Mainstream further There will be no appeal to any grand idealism or extraordinary capabilities or superhuman efforts That is just as well as history has also shown how such appeals if they are heard at all tend towards the slippery slope into dictatorship or totalitarianism We shall see below how more modern conceptions of leadership protect us from that Procrustean danger But is such a grand project feasible On page 8 op cit Booch laments Our failure to master the complexity of software results in projects that are late over budget and deficient in their stated requirements We often call this condition the software crisis but frankly a malady that has carried on this long must be called normal Yes there is that obvious crisis the mess and its underlying Syndrome but pursuing the point I have to insist that normal may certainly be redefined somewhat and yet remain natural We can clarify The Mainstream and ride it better Booch then still on p 8 provides a handy point of departure for us the underlying problem springs from the inherent complexity of software From there we now dig deeper than he set out to do in that book The Mainstream epistemology giving the MACK reality model It seems some mere wordplay by MACK as it shifts the perspective from that of Booch s cited book It is not software but reality that is inherently complex I suspect that Booch would agree with that as the philosopher which in that book he did not claim to be However I must add that his approach in that book as quoted above and though it is doubtless not his true self does make him appear a most typical instance of the classic inhabitant of the Charybdian figtree the professional who appears to have the edge in supposedly mastering complexity when in reality it is mainly artificial complication that is being addressed and no breakaway from that static position seems imminent In the MACK perspective software can at most be intricate or elaborate or complicated as in artificially complicated or convoluted as it is mere humans that make it so It is indeed a complex task to sort out the mess but that is solely because of the industry s humanly real component People are involved But the complexity is elsewhere and certainly not hidden or even overt in any Divine Programmer s attempted orchestrations Within that perspective we may certainly still talk of complex adaptive systems a useful concept which Jeff Sutherland has most pertinently placed at centre stage for this Workshop and which will recur in this paper However its complexity aspect derives from how the real world via its various and generally unpredictable inputs into the adaptive system adds a reality component The thus complexified whole system may then be considered a truly interesting part of reality But the model component not only remains artificial it should even be made and kept as simple as possible but no simpler being the added tail in what I called Einstein s Imperative in my 1996 paper Thus the shift in perspective is not mere wordplay The difference is absolutely fundamental far reaching and squarely part of The Mainstream of the evolution of species and artifacts as it has been represented in this paper It is the axiom already posited Restated for this context we have It is not the IT system that is complex but the reality it models The model itself should be simply structured simply accessible and simply adaptable Here I merely present the end result of the whole long story where as usual a picture is worth a thousand words It will only be a mental picture One thousand graphic words is worth one million purely abstract words as any one actual picture brings its own distortions and here I can surely rely on architects good visual imaginations It is also worth recalling from the metastory that the picture is of both kinds of model knowledge models in our minds or on our computers We shall also even see how the model itself even portrays what is possibly the major difference between the two We now look at the detail of the agate as already introduced Despite and yet also because of its rock solid apparent stability like the flint that sparks it is a form of chalcedony a microcrystalline quartz the agate in many significant ways portrays the phenomenon of knowledge as we observe it with its structure and evolutionary dynamics the way we humans live it within the ultimately unknowable complexity it models All images having limited applicability I might preface the incongruity of the very idea with the reassurance that we shall also see aspects where the image breaks down It is not followed slavishly The picture is of agates as they grow in their geological situations They typically form in cavities in a rock matrix as left by condensing gases in igneous rock that has solidified under low pressures near the surface of the earth In South Africa the rock is basalt from deep in the Earth s crust but now forming the Drakensberg mountains the highest in the sub continent My own agate specimens then on my desk in Cambridge I had found on Mont Aux Sources where our dry country s three largest rivers arise But that specific background is a merely poetic touch Such once molten and quickly cooled rock is characterized by its formlessness having no layered or visible crystal structure The rock matrix represents our given vast unstructured complex reality from which all knowledge derives Think of us cognitive beings as those cavities individual or group minds or databases to be filled with true and usable knowledge The image is quite the opposite of that of the classical tabula rasa a smooth surface on which our knowledge will be neatly inscribed The insides of the basalt cavities on the contrary are formless and especially rough even jagged and inhospitable yet ultimately fertile like that of lava flows destined for much smoothing shaping transformation and growth from appropriate seeds Now the rock matrix the womb over the ages secretes aqueous solutions rich in silica and other mineral components leached from the surrounding reality They percolate through networks of channels in the matrix and in due course start crystallizing out in the cavities The first deposits are from colloidal solutions and produce the microcrystalline chalcedony They line the rough boundary of the cavity so have no apparent form but that which is given by that unstructured reality Thus the outer surfaces of the agates represent our formless sensory and intuitive world

    Original URL path: http://jeffsutherland.org/oopsla98/SpottComplexity.html (2016-04-27)
    Open archived version from archive

  • Organization in a Chaotic World
    Slide 1 of 22 Notes Chris Marshall is a director of the Object Technology Group which is a subsidiary of SES Software

    Original URL path: http://jeffsutherland.org/oopsla98/marshall/sld001.htm (2016-04-27)
    Open archived version from archive

  • Business Object Stereotypes
    unit to optimize service levels within inventory constraints the production unit to improve efficiency and quality and the procurement agent to reduce costs and lead times The static purpose of such a unit is defined by its policies which detail the work it can do the conditions under which it is done and the courses of action that fulfill cancel or violate them Dynamic purpose is defined by its schedule

    Original URL path: http://jeffsutherland.org/oopsla98/marshall/sld002.htm (2016-04-27)
    Open archived version from archive



  •