Monday, June 25, 2018

Thinking about Software Architecture & Design : Part 13


Harold Abelson, co-author of the seminal tome Structure and Interpretation Of Computer Programs (SICP) said that “programs must be written for people to read, and only incidentally for machines to execute.”

The importance of human-to-human communication over human-to-machine is even more true in Software Architectures, where there is typically another layer or two of resolution before machines can interpret the required architecture.

Human-to-human communications is always fraught with potential for miscommunications and the reasons for this run very deep indeed. Dig into this subject and it is easy to be amazed that anything can be communicated perfectly at all. It is a heady mix of linguistics, semiotics, epistemology and psychology. I have written before (for example, in the “What is Law Series - http://seanmcgrath.blogspot.com/2017/06/what-is-law-part-14.html) about the first three of these, but here I want to talk about the fourth – psychology.

I had the good fortune many years ago to stumble upon the book Inevitable Illusions by Massimo Piattelli-Palmarini and it opened my mind to the idea that there are mental concepts we are all prone to develop, that are objectively incorrect – yet unavoidable. Think of your favorite optical illusion. At first you were amazed and incredulous. Then you read/discovered how it works. You proved to your own satisfaction that your eyes were deceiving you. And yet, every time you look at the optical illusion, your brain has another go at selling you on the illusion. You cannot switch it off. No amount of knowing how you are being deceived by your eyes will get your eyes to change their minds, so to speak.

I have learned over the years that some illusions about computing are so strong that it is often best to incorporate them into architectures rather than try to remove them. For example, there is the “send illusion”. Most of the time when there is an arrow between A and B in a software architecture, there is a send illusion lurking. The reason being, it is not possible to send digital bits. They don't move through space. Instead they are replicated. Thus every implied “send” in an architecture can never be a truly simple send operation and it involves at the very least, a copy followed by a delete. 

Another example is the idea of a finite limit to the complexity of business rules. This is the very (very!) appealing idea that with enough refinement, it is possible to arrive at a full expression of the business rules that express some desirable computation. This is sometimes true (especially in text books) which adds to the power of the inevitable illusion. However, in many cases this is only true if you can freeze requirements – a tough proposition – and often is impossible even then. For example in systems where there is a feedback loop between the business rules and the data creating a sort of “fractal boundary” that the corpus of business rules can never fully cover.

I do not let these concerns stop me from using concepts like “send” and “business rule repository” in my architectures because I know how powerfully these concepts are locked into all our minds. However, I do try to conceptualize them as analogies and remain conscious of the tricks my mind plays with them. I then seek to ensure that the implementation addresses the unavoidable delta between the inevitable illusion in my head and the reality in the machine.


Thursday, June 14, 2018

Thinking about Software Architecture & Design : Part 12


The word “flexible” gets used a lot in software architecture & design. It tends to get used in a positive sense. That is, "flexibility" is mostly seen as a good thing to have in your architecture.

And yet, flexibility is very much a two edged sword. Not enough of it, and your architecture can have difficulty dealing with the complexities that typify real world situations. Too much of it and your architecture can be too difficult to understand and maintain. The holy grail of flexibility, in my opinion, is captured in the adage that “simple things should be simple, and hard things should be possible.”.

Simple to say, hard to do. Take SQL for example, or XSLT or RPG...they all excel at making simple things simple in their domains and yet, can also be straitjackets when more complicated things come along. By “complicated” here I mean things that do not neatly fit into their conceptual models of algorithmics and data.

A classic approach to handling this is to allow such systems to be embedded in Turing Complete Programming language. i.e. SQL inside C Sharp. XSLT inside Java etc. The Turing Completeness of the programming language host ensures that the “hard things are possible” while the core – and now “embedded system” - ensures that the simple things are simple.

Unfortunately what tends to happen is that the complexity of the real world chips away at the split between simple and complex and, often times, such hybrid systems evolve into Turing Complete hosts. i.e. over time, the embedded system for handling the simple cases, is gradually eroded and then one day, you wake up to find that it is all written in C# or Java or whatever and the originally embedded system is withering on the vine.

A similar phenomenon happens on the data side where an architecture might initially by 98% “structured” fields but over time, the “unstructured” parts of its data model grow and grow to the point where the structured fields atrophy and all the mission critical data migrates over to the unstructured side. This is why so many database-centric systems organically grow memo fields, blob fields or even complete distinct document storage sub-systems over time, to handle all the data that does not fit neatly into the “boxes” of the structured fields.

Attempting to add flexibility to the structured data architecture tends to result in layers of abstraction that people have difficult following. Layers of pointer indirection. Layers of subject/verb/object decomposition. Layers of relationship reification and so on....

This entropy growth does not happen overnight. The complexity of modelling the real world chips away at designs until at some point there is an inflection. Typically this point of inflection manifests in a desire to “simplify” or “clean up” a working system. This often results in a new architecture that incorporates the learnings from the existing system and then the whole process repeats again. I have seen this iteration work at the level of decades but in more recent years the trend appears to be towards shorter and short cycle times.

This cyclic revisiting of architectures begs the obvious teleological question about the end point of this cycle. Does it have an end? I suspect not because, in a Platonic sense, the ideal architecture can be contemplated but cannot be achieved in the real world.

Besides, even if it could be achieved, the ever-changing and monotonically increasing complexity of the real world ensures that a perfect model for time T can only be achieved at some future time-point T+N, by which time, it is outdated and has been overtaken by the every shifting sands of reality.


So what is an architect to do if this is the case? I have come to the conclusion that it is very very important to be careful to label anything as an immutable truth in an architecture. All nouns, verbs, adjectives etc. that sound to you like that are “facts” of the real world, will, at some point bend under the weight of constant change and necessarily incomplete empirical knowledge.

The smaller the set of things you consider immutable facts, the more flexible your architecture will be. By all means, layer abstractions on top of this core layer. By all means add Turing Completeness into the behavioral side of the model. But treat all of these higher layers as fluid. It is not that they might need to change it is that they will need to change. It is just a question of time.

Finally, there are occasions where the set of core facts in your model is the empty set! Better to work with this reality than fight against it because entropy is the one immutable fact you can absolutely rely on. Possibly the only thing you can have an the core of your architecture and not worry about it being invalidated by the arrival of new knowledge or the passage of time.


Friday, June 01, 2018

Thinking about Software Architecture & Design : Part 11


It is said that there are really only seven basic storylines and that all stories can either fit inside them or be decomposed into some combination of the basic seven. There is the rags-to-riches story. The voyage and return story. The overcoming the monster story...and so on.

I suspect that something similar applies to Software Architecture & Design. When I was a much younger practitioner in this field, I remember a very active field with new methodologies/paradigms coming along on a regular basis. Thinkers such as Yourdon, de Marco, Jackson, Booch, Hoare, Dijkstra, Hohpe distilled the essence of most of the core architecture patterns we know of today.

In more recent years, attention appears to have moved away from the discovery/creation of new architecture patterns and architecture methodologies towards concerns closer to the construction aspects of software. There is an increasing emphasis on two way flows in the creation of architectures– or perhaps circular flows would be a better description. i.e. iterating backwards from, for example user stories, to the abstractions required to support the user stories. Then perhaps a forward iteration refactoring the abstractions to get coverage of the required user stories with less “moving parts” as discussed before.

There has also been a marked trend towards embracing the volatility of the IT landscape in the form of proceeding to software build phases with “good enough” architectures and the conscious decision to factor-in the possibility of needing complete architecture re-writes in ever short time spans.

I suspect this is an area where real world physical architecture and software architecture fundamentally differ and the analogy breaks down. In the physical world, once the location of the highway is laid down and construction begins, a cascade of difficult-to-reverse events starts to occur in parallel with the construction of the highway. Housing estates and commercial areas pop up close to the highway. Urban infrastructure plans – perhaps looking decades into the future – are created predicated on the route of the highway and so on.

In software, there are often similar amount of knock-on effects to architecture changes but when these items are themselves primarily software, rearranging everything based on a architecture is more manageable. Still likely a significant challenge, but more doable because software is, well “softer” than real world concrete, bricks and mortar.

My overall sense of where software architecture is today is that it revolves around the question : “how can we make it easier to fundamentally change the architecture in the future?” The fierce competitive landscape for software has combined with cloud computing to fuel this burning question.

Creating software solutions with very short (i.e. weeks) time horizons before they change again is now possible and increasingly commonplace. The concept of version number is becoming obsolete. Today's software solution may or may not be the same as the one you interacted with yesterday and it may, in fact, be based on an utterly different architecture under the hood than it was yesterday. Modern communications infrastructure, OS/device app stores, auto-updating applications, thin clients...all combine to create a very fluid environment for modern day software architectures to work in.

Are there new software patterns still emerging since the days of data flow and ER diagrams and OOAD? Are we just re-combining the seven basic architectures in a new meta-architecture which is concerned with architecture change rather than architecture itself? Sometimes I think so.

I also find myself wondering where we go next if that is the case. I can see one possible end point for this. An end-point which I find tantalizing and surprising in equal measure. My training in software architecture – the formal parts and the decades of informal training since then – have been based on the idea that the fundamental job of the software architect is to create a digital model – a white box – of some part of the real world, such that the model meets a set of expectations in terms of its interaction with its users (which may, be other digital models).

In modern day computing, this idea of the white box has an emerging alternative which I think of as the black box. If a machine could somehow be instructed to create the model that goes inside the box – based purely on an expression of its required interactions with the rest of the world – then you basically have the only architecture you will ever need for creating what goes into these boxes. The architecture that makes all the other architectures unnecessary if you like.

How could such a thing be constructed? A machine learning approach, based on lots and lots of input/output data? A quantum computing approach which tries an infinity of possible Turing machine configurations, all in parallel? Even if this is not possible today, could it be possible in the near future? Would the fact that boxes constructed this way would be necessarily black – beyond human comprehension at the control flow level – be a problem? Would the fact that we can never formally prove the behavior of the box be a problem? Perhaps not as much as might be initially thought, given the known limitations of formal proof methods for traditionally constructed systems. After all, we cannot even tell if a process will halt, regardless of how much access we have to its internal logic. Also, society seems to be in the process of inuring itself to the unexplainability of machine learning – that genie is already out of the bottle. I have written elsewhere (in the "what is law?" series - http://seanmcgrath.blogspot.com/2017/07/what-is-law-part-15.html) that we have the same “black box” problem with human decision making anyway).

To get to such a world, we would need much better mechanism for formal specification. Perhaps the next generation of software architects will be focused on patterns for expressing the desired behavior of the box, not models for how the behavior itself can be achieved. A very knotty problem indeed but, if it can be achieved, radical re-arrangements of systems in the future could start and effective stop with updating the black box specification with no traditional analysis/design/ construct/test/deploy cycle at all.

Monday, May 28, 2018

Thinking about Software Architecture & Design : Part 10


Once the nouns and verbs I need in my architecture start to solidify, I look at organizing them across multiple dimensions. I tend to think of the noun/verb organization exercise in the physical terms of surface area and moving parts. By "surface area" I mean minimizing the sheer size of the model. I freely admin that page count is a crude-sounding measure for a software architecture, but I have found over the years that the total size of the document required to adequately explain the architecture is an excellent proxy for its total cost of ownership.

It is vital, for a good representation of a software architecture, that both the data side and the computation side are covered. I have seen many architectures where the data side is covered well but the computation side has many gaps. This is the infamous “and then magic happens” part of the software architecture world. It is most commonly seen when there is too much use of convenient real world analogies. i.e. thematic modules that just snap together like jigsaw/lego pieces, data layers that sit perfectly on top of each other like layers of a cake, objects that nest perfectly inside other objects like Russian Dolls etc.

When I have a document that I feel adequately reflects both the noun and the verb side of the architecture, I employ a variety of techniques to minimize its overall size. On the noun side, I can create type hierarchies to explore how nouns can be considered special cases of other nouns. I can create relational de-compositions to explore how partial nouns can be shared by other nouns. I will typically “jump levels” when I am doing this. i.e. I will switch between thinking of the nouns in purely abstract terms (“what is a widget really” to thinking about them in physical terms: “how best to create/read/update/delete widgets?”). I think of it as working downwards towards implementation an upwards towards abstraction at the same time. It is head hurting at times, but in my experience produces better practical results that the simpler step-wise refinement approach of moving incrementally downwards from abstraction to concrete implementation.

On the verb side, I tend to focus on the classic engineering concept of "moving parts". Just as in the physical world, it has been my experience that the smaller the number of independent moving parts in an architecture, the better. Giving a lot of thought to opportunities to reduce the total number of verbs required pays handsome dividends. I think of it in terms of combinatorics. What are the fundamental operators I need from which, all the other operators can be created by combinations of the fundamental operators? Getting to this set of fundamental operators is almost like finding the architecture inside the architecture.

I also think of verbs in terms of complexity generators. Here I am using the word “complexity” in the mathematical sense. Complexity is not a fundamentally bad thing! I would argue that all system behavior has a certain amount of complexity. The trick with complexity is to find ways to create the amount required but in a way that allows you to be in control of it. The compounding of verbs is the workhorse for complexity generation. I think of data as a resource that undergoes transformation over time. Most computation – even the simplest assignment of the value Y to be the value Y + 1 has an implicit time dimension. Assuming Y is a value that lives over a long period of time – i.e. is persisted in some storage system – then Y today is just the compounded result of the verbs applied to it from its date of creation.

There are two main things I watch for as I am looking into my verbs and how to compound them and apply them to my nouns. The first is to always include the ability to create an ad-hoc verb “by hand”. By which I mean, always having the ability to edit the data in nouns using purely interactive means. This is especially important in systems where down-time for the creation of new algorithmic verbs is not an option.

The second is watching out for feedback/recursion in verbs. Nothing generates complexity faster than feedback/recursion and when it is it used, it must be used with great care. I have a poster on my wall of a fractal with its simple mathematical formula written underneath it. It is incredible that such bottomless complexity can be derived from such a harmless looking feedback loop. Using it wisely can produce architectures capable of highly complex behaviors but with small surface areas and few moving parts. Used unwisely.....

Monday, May 21, 2018

Thinking about Software Architecture & Design : Part 9

I approach software architecture through the medium of human language. I do make liberal use of diagrams but the diagrams serve as illustrators of what, to me, is always a linguistic conceptualization of a software architecture. In other words, my mental model is nouns and verbs and adjectives and adverbs. I look for nouns and verbs first. This is the dominant decomposition for me. What are that things that exist in the model? Then I look for what actions are performed on/by the things in the model. (Yes, the actions are also “things” after a fashion...)

This first level of decomposition is obviously very high level and yet, I find it very useful to pause at this level of detail and do a gap analysis. Basically what I do is I explain the model to myself in my head and look for missing nouns and verbs. Simple enough.

But then I ask myself how the data that lives in the digital nouns actually gets there in the first place. Most of the time when I do this, I find something missing in my architecture. There are only finite number of ways data can get into a digital noun in the model. A user can enter it, an algorithm can compute it, or an integration point can supply it. If I cannot explain all the data in the model through the computation/input/integration decomposition, I am most likely missing something.

Another useful question I ask at this level of detail relates to where the data in the nouns goes outside the model. In most models, data flows out at some point to be of use i.e. it hits a screen or a printout or an outward bound integration point. Again, most of the time when I do this analysis, I find something missing – or something in the model that does not need to be there at all.

Getting your nouns and verbs straight is a great first step towards what will ultimately take the form of objects/records and methods/functions/procedures. It is also a great first step if you are taking a RESTian approach to architecture as the dividing line between noun-thinking and verb-thinking is the key difference between REST and RPC in my experience.

It is hard to avoid prematurely clustering nouns into types/classes as our brains appear to be wired towards organizing things into hierarchies. I do this because I find that as soon as I start thinking hierarchically, I close off the part of my brain that is open to alternative hierarchical decompositions. I try to avoid that because in my experience, the set of factors that steer an architecture towards one hierarchy instead of another are practical ones, unrelated to “pure” data modelling. i.e. concerns related to organizational boundaries, integration points, cognitive biases etc.

Take the time to explore as many noun/verb decompositions as you can because as soon as you pick one and start to refine the model, it becomes increasingly hard to think “outside the box” of your own architecture.

Friday, May 18, 2018

Thinking about Software Architecture & Design : Part 8


It is common practice to communicate software architectures using diagrams, but most diagrams, in my experience are at best rough analogies of the architecture rather than faithful representations of it.

All analogies break down at some point. That is why we call them “analogies”. It is a good idea to understand where your analogies break down and find ways to compensate.

In my own architecture work, the main breakdown point for diagrams is that architectures in my head are more like movies than static pictures. In my minds eye, I tend to see data flowing. I tend to see behaviors – both human and algorithmic – as animated actors buzzing around a 3D space, doing things, producing and consuming new data. I see data flowing, data flowing out, data staying put but changing shape over time, I see feedback loops where data flows out but then comes back in again. I see the impact of time in a number of different dimensions. I see how it relates to the execution paths of the system. I see how it impacts the evolution of the system as requirements change. I see how it impacts the dependencies of the system that are outside of my control e.g. operating systems etc.

Any static two dimensional picture or set of pictures, that I take of this architecture necessarily leaves a lot of information behind. I liken it to taking a photo of a large city at 40,000 feet and then trying to explain all that is going on in that city, through that static photograph. I can take photos from different angles and that will help but, at the end of the day, what I would really like is a movable camera and the ability to walk/fly around the “city” as a way of communicating what is going on in it, and how it is architected to function. Some day...

A useful rule of thumb is that most boxes, arrows, straight lines and layered constructions in software architecture diagrams are just rough analogies. Boxes separating say, organizations in a diagram, or software modules or business processes are rarely so clean in reality. A one way arrow from X to Y is probably in reality a two way data flow and it probably has a non-zero failure rate. A straight line separating, say “valid” from “invalid” data records probably has a sizable grey area in the middle for data records that fuzzily sit in between validity and invalidity. And so on.

None of this is in any way meant to suggest that we stop using diagrams to communicate and think about architectures. Rather, my goal here is just to suggest that until we have better tools for communicating what architectures really are, we all bear in mind the limited ability of static 2D diagrams to accurately reflect them.

Thursday, May 10, 2018

Thinking about Software Architecture & Design : Part 7


The temptation to focus a lot of energy on the one killer diagram that captures the essence of your architecture is strong. How many hours have I spent in Visio/Powerpoint/draw.io on “the diagram”? More than I would like to admit to.

Typically, I see architectures that have the “main diagram” and then a series of detail diagrams hidden away, for use by implementation and design teams. The “main diagram” is the one likely to go into the stakeholder presentation deck.

This can works fine when there are not many stakeholders and organizational boundaries are not too hard to traverse. But as the number of stakeholders grows, the power of the single architectural view diminishes. Sometimes, in order to be applicable to all stakeholders, the diagram becomes so generic that it really says very little i.e. the classic three-tiered architecture or the classic hub-and-spoke or the peer-to-peer network. Such diagrams run the risk of not being memorable by any of the stakeholders, making it difficult for them to get invested in it.

Other times, the diagram focuses on one particular “view” perhaps by putting one particular stakeholder role in the center of the diagram, with the roles of the other stakeholders surrounding the one in the middle.

This approach can be problematic in my experience. Even if you take great pains to point out that there is no implied hierarchy of importance in the arrangement of the diagram, the role(s) in the middle of the diagram will be seen as more important. It is a sub-conscious assessment. We cannot help it. The only exception I know of is when flow-order is explicit in the diagram but even then whatever is in the middle of the diagram draws our attention.

In most architectures there are “asks” of the stakeholders. The best way to achieve  these “asks” in my experience is to ensure that each stakeholder gets their own architecture picture, that has their role in the center in the diagram, with all other roles surrounding their part in the big picture.

So, for N stakeholders there are N "main views" - not just one. All compatible ways of looking at the same thing. All designed to make it easier for each stakeholder to answer the “what does this mean for me?” question which is always there – even if it is not explicitly stated.

Yes, it is a pain to manage N diagrams but you probably have them anyway – in the appendices most likely, for the attention of the design and implementation phase. My suggestion is to take them out of the appendices and put them into the stakeholder slide deck.

I typically present two diagrams to each stakeholder group. Slide one is the diagram that applies to all stakeholders. Slide two is for the stakeholder group I am presenting to. As I move around the different stakeholder meetings, I swap out slide number two.


Tuesday, May 08, 2018

Thinking about Software Architecture & Design : Part 6


Abstractions are a two-edged sword in software architecture. Their power must be weighed against their propensity to ask too much from stakeholders, who may not have the time or inclination to fully internalize them. Unfortunately most abstractions require internalization for appreciation of their power.

In mathematics, consider the power of Euler's equation and contrast it with the effort involved to understand what its simple looking component symbols represent. In music, consider the power of the Grand Staff to represent musical compositions and contrast that with the effort required to understand what its simple looking symbols represent.

Both of these abstractions are both very demanding and very powerful. It is not uncommon in software architecture for the practitioner community to enjoy and be comfortable with constantly internalizing new abstractions. However, in my experience, a roomful of software architects is not representative of a roomful of stakeholders in general.

Before you release your killer abstractions from the whiteboard into Powerpoint, try them out on a friendly audience of non-specialists first.

Wednesday, May 02, 2018

Thinking about Software Architecture & Design : Part 5

Most architectures will have users at some level or other. Most architectures will also have organisational boundaries that need to be crossed during information flows.

Each interaction with a user and each transition of an organizational boundary is an “ask”. i.e. the system is asking for the cooperation of some external entity. Users are typically being asked to cooperate by entering information into systems. Parties on the far end of integration points are typically being asked to cooperate by turning around information requests or initiating information transfers.

It is a worthwhile exercise while creating an architecture, to tabulate all the “asks” and identify those that do not have associated benefits for those who are performing the asks.

Any entities interacting with the system that are giving more than they are receiving, are likely to be the most problematic to motivate to use the new system. In my experience, looking for ways to address this at architecture time can be very effective and sometimes very easy. The further down the road you get towards implementation, the harder it is to address motivational imbalances.

If you don't have  an answer to the “what is in it for me?” question, for each user interaction and each integration point interaction, your architecture will face avoidable headwinds both in  implementation and in operation.

Friday, April 27, 2018

Thinking about Software Architecture & Design : Part 4


Any new IT system will necessarily sit inside a larger context. If that context includes the new system having its own identity – from “the new system” to “Project Unity” - it will be anthropomorphised by its users. This can be good, bad or neutral for the success of the new IT system.


It does not matter what form the naming takes e.g. “Project Horizon”, “Bob's System”, “The new HR system” or even a visual identity such as "the new button", or even a tactile identity such as “the new panel under the desk at reception”. In all cases the new IT system may be treated by existing team members in much the same way as a new team member would be treated.


New systems get “sized up”, so to speak, by their users. Attributes such as “fast”, “unreliable”, “inflexible” or even “moody” might be applied to the new system. These may be factually based, or biased, depending on the stance the community of users adopts towards the new system arriving into their team.


One particularly troublesome possibility is that the new system may be seen as causal factor in events unrelated to it. “X used to work fine before the new system came along....” The opposite can also happen i.e. the new system get plaudits for events it had no hand or part in. Causality versus correlation can be a tricky distinction to navigate.


Takeaway: sometimes the human tendency towards the anthropomorphic can be used to your advantage. If you suspect the opposite may be true for your new system, it can be useful to purposely avoid elaborate naming and dramatic rollout events which can exacerbate anthropomorphisation.


Sometimes, new systems are best rolled out with little or no fanfare in a “business as usual” mode. Sometimes it is not possible to avoid big bang system switchover events but if it is at all possible to adopt a phased approach to deployment, and transition slowly, I would recommend it for many reasons, one of which is the sort of team dynamics alluded to here.


As AI/Robotics advances, I think this will become even more important in the years ahead.


Tuesday, April 24, 2018

Thinking about Software Architecture & Design : Part 3


In software architecture and design we have some pretty deep theories that guide us on our way. We know how to watch out for exponential run times, undetectable synchronisation deadlocks, lost update avoidance etc. We have Petri nets, state charts, entity/attribute diagrams, polymorphic object models, statistical queuing models, QA/QC confidence intervals....the list goes on....


...and yet, in my experience, the success of a software architecture & design project tends, in my experience, to revolve around an aspect of the problem domain that is not addressed by any of the above. I call it the experiential delta.


Simply put, the experiential delta is a measure of how different the “to be” system appears to be, to those who will interact with it – experience it – day to day.


A system can have a very high architecture delta but a low experiential delta and be many orders of magnitude easier to get into production than a system with low architecture delta but a high experiential delta. 


It pays to know what type of experiential delta your “to be” solution represents. If it has high experiential delta, it pays to put that issue front and center in your planning. Such projects tend to be primarily process change challenges with some IT attached, as opposed to being IT projects with some process change attached.


In my experience, many large IT projects that fail, do not fail for IT reasons per se. They fail for process change reasons, but get labeled as IT failures after the fact. The real source of failure in some cases, is a failure to realize the importance of process change and the need to get process change experts into the room fast. As soon the the size of the experiential delta crystallizes.


Indeed in some situations it is best to lead the project as a process change project, not an IT project at all. Doing so has a wonderful way of focusing attention on the true determinant of success. The petri nets will look after themselves.

Thursday, April 19, 2018

Thinking about Software Architecture & Design : Part 2

Technological volatility is, in my experience, the most commonly overlooked factor in software architecture and design. We have decades worth of methodologies and best practice guides that help us deal with fundamental aspects of architecture and design such as reference data management, mapping data flows, modelling processes, capturing input/output invariants, selecting between synchronous and asynchronous inter-process communication methods...the list goes on. 

And yet, time and again, I have seen software architectures that are only a few years old, that need to be fundamentally revisited. Not because of any significant breakthrough in software architecture & design techniques, but because technological volatility has moved the goal posts, so to speak, on the architecture.

Practical architectures (outside those in pure math such as Turing Machines) cannot exist in a technological vacuum. They necessarily take into account what is going on in the IT world in general. In a world without full text search indexes, document management architectures are necessarily different. In a world without client side processing capability, UI architectures are necessarily different, in a world without always-on connectivity.....and so on.

When I look back at IT volatility over my career – back to the early Eighties – there is a clear pattern in the volatility. Namely, that volatility increases the closer you get to the end-users points of interaction with IT systems. Dumb “green screens”, bit-mapped graphics, personal desktop GUIs, tablets, smart phones,  voice activation, haptic user interfaces..

Many of the generational leaps represented by these innovations have had profound implications on the software architectures that leverage them. It is not possible – in my experience – to abstract away user interface volatility and treat it as a pluggable layer on top of the main architecture. End-user technologies have a way of imposing themselves deeply inside architectures. For example, necessitating an event-oriented/multi-threaded approach to data processing in order to make it possible to create responsive GUIs. Responding sychronously to data queries as opposed to batch processing. 

The main takeaway is this: creating good software architectures pay dividends but they are much more likely to be significant in the parts of the architecture furthest away from the end-user interactions. i.e. inside the data modelling, inside discrete data processing components etc. They are least likely to pay dividends in areas such as GUI frameworks, client side processing models or end user application programming environments.

In fact, volatility is sometimes so intense, that it makes more sense to not spend time abstracting the end-user aspects of the architecture at all. i.e. sometimes it makes more sense to make a conscious decision to re-do the architecture if/when the next big upheaval comes on the client side and trust that large components of the back-end will remain fully applicable post-upheaval.

That way, your applications will not be as likely to be considered “dated” or “old school” in the eyes of the users, even though you are keeping much of the original back-end architecture from generation to generation.

In general, software architecture thinking time is more profitably spent in the back-end than in the front-end. There is rarely a clean line that separates these so a certain amount of volatility on the back-end is inevitable, but manageable, in compared to the volatility the will be visited upon your front-end architectures.

Volatility exists everywhere of course. For example, at the moment serverless computing models are having profound implications on "server side" architectures. Not because of end-user concerns - end-users do not know or care about these things - but because of the volatility in the economics of cloud computing.

If history is anything to go by, it could be another decade or more before something comes along like serverless computing, that profoundly impacts back-end architectures. Yet in the next decade we are likely to see dozens of major changes in client side computing. Talking to our cars, waving at our heating systems, installing apps subcutaneously etc.

Friday, April 13, 2018

Thinking about Software Architecture & Design : Part 1

This series of posts will contain some thoughts on software architecture and design. Things I have learned over the decades spent doing it to date. Things I think about but have not got good answers for. Some will be very specific - "If X happens, best to do Y straight away.", some will be philosophical "what exactly is X anyway?", some will be humorous, some tragic, some cautionary...hopefully some will be useful. Anyway, here goes...

The problem of problems

Some "problems" are not really problems at all. By this I mean that sometimes, it is simply the way a “problem” is phrased that leads you to think that the problem is real and needs to be solved.  Other times, re-phrasing the problem leads to a functionally equivalent but much more easily solved problem.

Another way to think about this is to recognize that human language itself is always biased towards a particular world view (that is why translating one human language into another is so tricky. It is not a simple mapping of one world view to another).

Simply changing the language used to describe a “problem” can sometimes result in changing (but never removing!) the bias. And sometimes, this new biased position leads more readily to a solution.

I think I first came across this idea in the book "How to Solve It" by the mathematician George Poyla.  Later on, I found echoes of it in the work of philosopher Ludwig Wittgenstein. He was fond of saying (at least in his early work) that there are no real philosophical problems – only puzzles – caused by human language.

Clearing away the fog of human language - says Wittgenstein - can show a problem to be not a problem at all. I also found this idea in the books of Edward de Bono whose concepts of “lateral thinking" often leverage the idea of changing the language in which a problem is couched as a way of changing view-point and finding innovative solutions.

One example De Bono gives is a problem related to a factory polluting water in a river. If you focus on the factory as a producer of dirty water, your problem is oriented around the dirty water. It is the dirty water that output that needs to be addressed. However if the factory also consumes fresh water, then the problem can be re-cast in terms of a pre-factory input problem. i.e. make the factory put its intake upstream from its water discharge downstream. Thus incentivizing the factory to not pollute the river. Looked at another way, the factory itself becomes a regulator, obviating or at least significantly reducing the need for extra entities in the regulation process.

In more recent years I have seen the same idea lurking in Buddhist philosophy in the form of our own attitudes towards a situation being a key determinant in our conceptualization of a situation as either good/bad or neutral. I sometimes like to think of software systems as "observers" of the world in this Buddhist philosophy sense. Admittedly these artificial observers are looking at the world through more restricted sense organs that humans, but they are observers none-the-less.

Designing a software architecture is essentially baking in a bias as to how "the world" is observed by a nascent software system. As architects/designers we transfer our necessarily biased  conceptualization of the to-be system into code with a view to giving life to a new observer in the world - a largely autonomous software system.

Thinking long and hard about the conceptualization of the problem can pay big dividends early on in software architecture. As soon as the key abstractions take linguistic form in your head i.e. concepts start to take the form of nouns, verbs, adjectives etc., the problem statement is baked in, so to speak.

For example. Imagine a scenario where two entities, A and B, need to exchange information. Information needs to flow from A to B reliably.  Does it matter if I think of A sending information to B or think of B as querying information from A? After all, the net result is the same, right? The information gets to B from A, right?

Turns out it matters a lot. The bias in the word "send" is that it carries with it the notion of physical movement. If I send you a postcard in the mail. The postcard moves. There is one postcard. It moves from 1) I have it to 2) in transit to 3) you have it (maybe).

If we try to implement this "send" in software, it can get very tricky indeed to fully emulate what happens in a real world "send" - especially if we stipulate guaranteed once and only once delivery. Digital "sends" are never actually sends. They are always replications, or normally replicate followed by delete.

If instead of this send-centric approach, we focus on B as the active party in the information flow  - querying for information from A and simply re-requesting it if, for some reason it does not arrive, then we have a radically different software architecture. An architecture that is much easier to implement in many scenarios. (Compare the retry-centric architecture of many HTTP systems compared to, say, reliable message exchange protocols.)

So what happened here? We simply substituted one way of expressing the business need - a send-oriented conceptualization, with a query-oriented conceptualization, and the "problem" changed utterly before our very eyes.

Takeaway : the language in which a problem is expressed is often, already a software architecture. It may or may not be a good version 1 of the architecture to work from. It contains many assumptions. Many biases. Regardless of whether or not it is linguistic or visual.

It often pays to tease out those assumptions in order to see if a functionally equivalent re-expression of the problem is a better starting point for your software architecture.

Friday, February 23, 2018

What is a document - Part 7



The word “document” is, like the word “database”, simple on the outside and complex on the inside. 

Most of us carry around pragmatically fuzzy definitions of these in our heads. Since the early days of personal computers there have been software suites/bundles available that have included distinct tools to manage “documents” and “databases”, treating them as different types of information object. The first such package I used was called SMART running on an IBM PC XT machine in the late Eighties. It had a 10MB hard disk. Today, that is hardly enough to store a single document, but I digress...

I have used many other Office Suites since then, most of which have withered on the vine in enterprise computing, with the notable exception of Microsoft Office. I find it interesting that of the words typically associated with office suites, namely, “database”, “word processor”, “presentation”, and “spreadsheet” the two that are today most tightly bound to Microsoft office are “spreadsheet” and “presentation” to the point where “Excel” and “Powerpoint” have become generic terms for “spreadsheet” and “presentation” respectively. I also think it is interesting Excel has become the de-facto heart of Microsoft Office in the business community with Word/Access/Powerpoint being of secondary importance as "must haves" in office environments, but again I digress...

In trying to chip away at the problem of defining a “document” I think it is useful to imagine having the full Microsoft office suite at your disposal and asking the question “when should I reach for Word instead of one of the other icons when entering text?” The system I worked in in the Nineties, mentioned previously in this series, required a mix of classic field-type information along with unstructured paragraphs/tables/bulleted lists. If I were entering that text into a computer today with Microsoft Office at my disposal, would I reach for the word processor icon or the database icon?

I would reach for the Word icon. Why? Well, because there are a variety of techniques I can use in Word to enter/tag field-type textual information and many techniques for entering unstructured paragraphs/tables/bulleted lists. The opposite is not true. Databases tend to excel (no pun intended) at field-type information but be limited in their support for unstructured paragraphs/tables/bulleted lists – often relegating the latter to “blob” fields that are second-class citizens in the database schema. 

Moreover, these days, the tools available for post-processing Word's .docx file format make it much easier than ever before to extract classic “structured XML” from Word documents but with the vital familiarity and ease of use for the authors/editors I mentioned previously.

Are there exceptions? Absolutely. There are always exceptions. However, if your data structure necessarily contains a non-trivial amount of unstructured or semi-structured textual content and if your author/edit community wants to think about the content in document/word-processor terms, I believe today version of Word with its docx file format is generally speaking a much better starting point than any database front-end or spreadsheet front-end or web-browser front-end or any structured XML editing tool front-end.

Yes, it can get messy to do the post-processing of the data but given a choice between a solution architecture that guarantees me beautifully clean data at the back-end but an author/edit community who hate it, versus a solution architecture that involves extra content enrichment work at the back end but happy author/edit users, I have learned to favor the latter every time.

Note I did not start there! I was on the opposite side of this for many, many years, thinking that structured author/edit tools, enforcing structure at the front-end was the way to go. I built a few beautiful structured systems that ultimately failed to thrive because the author/edit user community wanted something that did not “beep” as they worked on content. I myself, when writing the books I wrote for Prentice-Hall (books on SGML and XML - of all things!), I myself wanted something that did not beep!

Which brings me (finally!), to my answer to the question “What is a document?”. My answer is that a document is a textual information artifact where the final structure of the artifact itself is only obvious after it has been created/modified and thus requires an author/edit user experience that gets out of the way of the users creative processes until the user decides to impose structure – if they decide to impose a structure at all.

There is no guaranteed schema validity other than that most generic of schemas that splits text into flows, paragraphs, words, glyphs etc and allows users to combine content and presentation as they see fit.

On top of that low level structure, anything goes – at least until the point where the user has decided that the changes to the information artifact are “finished”. At the point where the intellectual work has been done figuring our that the document should say and how it should say it, it is completely fine - and generally very useful - to be able to validate against higher level, semantic structures such as "chapter", "statute", "washing machine data sheet" etc.

The big lesson of my career to date in high volume document management/processing is that if you seek to impose this semantic structure on the author/edit community rather than have them come to you and ask for some structure imposition, you will struggle mightily to have a successful system.

Friday, February 09, 2018

What is a document? - part 6



By the late Nineties, I was knee deep in the world of XML and the world of Python, loving the way that these two amazing tools allowed tremendous amounts of automation to be brought to traditionally labor intensive document processing/publishing tasks. This was boom time in electronic publishing and every new year brought with it a new output format to target: Microsoft Multimedia Viewer, Windows Help, Folio Views, Lotus Notes and a whole host of proprietary formats we worked on for clients. Back then, HTML was just another output for us to target. Little did we know back then that it would eclipse all the others.

Just about twenty years ago now - in the fall of 1998 - , I co-presented a tutorial on XML at the International Python Conference in Houston, Texas. [1]. At that same conference, I presented a paper on high volume XML processing with Python [2]. Back in those days, we had some of the biggest corpora of XML anywhere in the world, here in Ireland. Up to the early/mid oozies, I did a lot of conference presentations and become associated with the concept of XML processing pipelines[3].

Then a very interesting thing happened. We began to find ourselves working more and more in environments where domain experts –not data taggers or software developers – needed to create and update XML documents. Around this time I was also writing books on markup languages for Prentice Hall[4] and had the opportunity to put “the shoe on the other foot” so-to-speak, and see things from an authors perspective.

It was then that I experienced what I now consider to be a profound truth of the vast majority of documents in the world - something that gets to the heart of what a document actually is which distinguishes it from other forms of digital information. Namely, that documents are typically very “structured” when they are finished but are highly unstructured when then are being created or in the midst of update cycles.

I increasingly found myself frustrated with XML authoring tools that would force me to work on my document contents in a certain order and beep at me unless my documents were “structured” at all times. I confess there were many times when I abandoned structured editors for my own author/edit work with XML and worked in the free-flowing world of the Emacs text editor or in word processors with the tags plainly visible as raw text.

 I began to appreciate that the ability to easily create/update content is a requirement that must be met if the value propositions of structured documents are to be realized, in most cases. There is little value in a beautifully structured, immensely powerful back-end system for processing terabytes of documents coming in from domain experts unless said domain experts are happy to work with the author/edit tools.

For a while, I believed it was possible to get to something that authors would like, by customizing the XML editing front-ends. However, I found that over and over again, two things started happening, often in parallel. Firstly, the document schemas became less and less structured so as to accommodate the variations in real-world documents and also to avoid “beeping” at authors where possible. Secondly, no amount of GUI customization seemed to be enough for the authors to feel comfortable with the XML editors.

“Why can't it work like Word?” was a phrase that began to pop up more and more in conversations with authors. For quite some time, while Word's file format was not XML-based, I would look for alternatives that would be Word-like in terms of the end-user experience, but with file formats I could process with custom code on the back end.

For quite a few years, StarOffice/OpenOffice/LibreOffice fitted the bill and we have had a lot of success with it. Moreover, it allowed for levels of customization and degrees of business-rule validation that XML schema-based approaches cannot touch. We learned may techniques and tricks over the years to guide authors in the creation of structured content without being obtrusive and interrupting their authoring flow. In particular, we learned to think about document validation as a function that the authors themselves have control over. They get to decide when their content should be checked for structural and business rules – not the software.

Fast forward to today. Sun Microsystems is no more. OpenOffice/LibreOffice do not appear to be gaining the traction in the enterprise that I suspected they would a decade ago. Googles office suite – ditto. Native, browser based document editing (such as W3C's Amaya [5]) does not appear to be getting traction either....

All the while, the familiar domain expert/author's mantra rings in my ears “Why can't it work like Word?”

As of 2018, this is a more interesting question than it has ever been in my opinion. That is where we will turn next.



Friday, January 26, 2018

What is a document? - Part 5


In the early Nineties, I found myself tasked with the development of a digital guide to third level education in Ireland. The digital product was to be an add-on to a book based product, created in conjunction with the author of the book. The organization of the book was very regular. Each third level course had a set of attributes such as entry level qualifications, duration, accrediting institution, physical location of the campus, fees and so on. All neatly laid out, on page per course, with some free-flowing narrative at the bottom of each page. The goals of the digital product were to allow prospective students to search based on different criteria such as cost ranges, course duration and so on.

Step number one was getting the information from the paper book into a computer and it is in this innocuous sounding step that things got very interesting. The most obvious approach - it seemed to me at the time - was to create a programmable database – in something like Clipper (a database programming language that was very popular with PC developers at the time). Tabular databases were perfect for 90% of the data – the “structured” parts such as dates, numbers, short strings of text. However, the tabular databases had no good way of dealing with the free-flowing narrative text that accompanied each course in the book. It had paragraphs, bulleted lists, bold/italics and underline...

An alternative approach would be to start with a word-processor – as opposed to a database – as it would make handling the free-flowing text (and associated formatting, bold/italic, bulleted lists etc.) easy. However, the word processor approach did not make it at all easy to process the “structured” parts in the way I wanted to (in many cases, the word processors of the day stored information in encrypted formats too).

My target output was a free viewer that came with Windows 3.1 known as Windows Help. If I could make the content programmable, I reasoned, I could automatically generate all sorts of different views of the data as Windows Help files and ship the floppy disk without needing to write my own viewer. (I know this sounds bizarre now but remember this work predated the concept of a generic web browser by a few years!)

I felt I was facing a major fork in the road in the project. By going with a database, some things were going to be very easy but some very awkward. By going with a document instead...same thing. Some things easy, some very awkward. I trawled around in my head for something that might have the attributes of a database AND of a document at the same time.

As luck would have it, I had a Byte Magazine from 1992 on a shelf. It had an article by Jon Udell that talked about  SGML - Standard Generalized Markup Language. It triggered memories of a brief encounter I had had with SGML back in Trinity College when Dr. David Abrahamson had referencing it in his compiler design course, back in 1986. Back then, SGML was not yet an ISO standard (it became one in 1987). I remember in those days hearing about “tagging" and how an SGML parser could enforce structure – any structure you liked – on text – in a similar way to programming language parsers enforced structure on, say, Pascal source code.

I remember thinking “surely if SGML can deal with the hierarchical structures like you typically find in programming languages, it can deal with the simpler, flatter structures you get in tabular databases?”. If it could, I reasoned, then surely I could get the best of both worlds. My own data format that had what I needed from database-approaches but also what I needed from document approaches to data modelling?

I found – somehow (this is all pre-internet remember. No Googling for me in those days.) – an address in Switzerland that I could send some money to in the form of a money order, to get a 3.5 inch floppy back by return post, with an SGML parser on it called ArcSGML. I also found out about an upcoming gathering in Switzerland of SGML enthusiasts. A colleague, Neville Bagnall went over and came back with all sorts of invaluable information about this new thing (to us) called generalized markup.

We set to work in earnest. We created our first ever SGML data model. Used ArcSGML to ensure we were getting the structure and consistency we wanted in our source data. We set about inventing tags for things like “paragraph”, “bold”, “cross-reference” as well as the simpler field-like tags such as “location”, “duration” etc. We sent about looking at ways to process the resultant SGML file. The output from ArcSGML was not very useful for processing, but we soon found out about another SGML parser called SGMLS by Englishman James Clark. We got our hands on it and having taken one look at the ESIS format it produced, we fell in love with it. Now we had a tool that could validate the structure of our document/database and feed us a clean stream of data to process downstream in our own software.

Back then C++ was our weapon of choice. Over time our code turned into a toolkit of SGML processing components called IDM (Intelligent Document Manager) which we applied to numerous projects in what became known as the “electronic publishing era”. Things changed very rapidly in those days. The floppy disks gave way to the CD-ROMs. We transitioned from Windows Help files to another Microsoft product called Microsoft Multimedia Viewer. Soon the number of “viewers” for electronic books exploded and we were working on Windows Help, Multimedia Viewer, Folio Views, Lotus Notes to name but four.

As the number of distinct outputs we needed to generate grew, so too did the value of our investment getting up to speed with SGML. We could maintain a single source of content but generate multiple output formats from it, each leveraging the capabilities of the target viewer in a way that made them look and feel like they had been authored directly in each tool as opposed to programmatically generated for them.

My concept of a “document” changed completely over this period. I began to see how formatting – and content – could be separated from each other. I began to see how in so doing, a single data model could be used to manage content that is tabular (like a classic tabular database) as well as content that is irregular, hierarchical, even recursive. Moreover, I could see how keeping the formatting out of the core content made it possible to generate a variety of different formatting “views” of the same content.

It would be many years later that the limitations of this approach became apparent to me. Back then, I thought it was a completely free lunch. I was a fully paid-up convert to the concept of generalized markup and machine readable, machine validatable documents. As luck would have it, this coincided with the emergence of a significant market for SGML and SGML technologies. Soon I was knee deep in SGML parsers, SGML programming languages, authoring systems, storage systems and was developing more and more of our own tools, first in C++, then Perl, then Python.

The next big transition in my thinking about documents came when I needed to factor non-technical authors into my thinking. This is where I will turn next. What is a document? - Part 6.

Monday, January 15, 2018

What is a document? - part 4


In the late Eighties, I had access to an IBM PC XT machine that had Wordperfect 5.1[1] installed on it. Wordperfect was both intimidating and powerful. Intimidating because when it booted, it completely cleared the PC screen and unless you knew the function keys (or had the sought-after function key overlay [2]) you were left to you own devices to figure out how to use it.

It was also very powerful for its day. It could wrap words automatically (a big deal!). It could  redline/strikeout text which made it very popular with lawyers working with contracts. It could also split its screen in two, giving you a normal view of the document on top and a so-called “reveal codes” view on the bottom. In the “reveal codes” area you could see the tags/markers used for formatting the text. Not only that, but you could choose to modify the text/formatting from either window.

This idea that a document could have two “faces” so to speak and that you could move between them made a lasting impression on me. Every other DOS-based word processor I came across seemed to me to be variations on the themes I had first seen in Wordperfect e.g. Wordstar, Multimate and later Microsoft Word for DOS. I was aware of the existence of IBM Displaywriter but did not have access to it. (The significance of IBM in all this document technology stuff only became apparent to me later.)

The next big "aha moment" for me came with the arrival of a plug-in board for IBM PCs called the Hercules Graphics Card[3]. Using this card in conjunction with the Ventura Publisher[4] on DRI's GEM graphics environment [5] dramatically expanded the extent to which documents could be formatted - both on screen an on the resultant paper. Multiple fonts, multiple columns, complex tables, equations etc. Furthermore, the on-screen representation mirrored the final printed output closely in what is now universally known as WYSIWYG.

Shortly after that, I found myself with access to an Apple Lisa [6] and then an Apple Fat Mac 512 with Aldus (later Adobe) Pagemaker [7] and an Apple Laserwriter[8]. My personal computing world split into two. Databases, spreadsheets etc. revolved around IBM PCs and PC compatibles such as Compaq, Apricot etc. Document processing and Desktop Publishing revolved around Apple Macs and Laser Printers.

I became intoxicated/obsessed with the notion that the formatting of documents could be pushed further and further by adding more and more powerful markup into the text. I got myself a copy of The Postscript Language Tutorial and Cookbook by Adobe[9] and started to write Postscript programs by hand.

I found that the original Apple Laserwriter had a 25 pin RS/232 port. I had access to an Altos multi-terminal machine [10]. It had some text-only applications on it. A spreadsheet from Microsoft called – wait for it – Multiplan (long before Excel) – running on a variant of – again, wait for it – Unix call Microsoft Xenix [11].

Well, I soldered up a serial cable that allowed me to connect the Altos terminal directly to the Apple Laserwriter. I found I could literally type in Postscript command at the terminal window and get pages to print out. I could make the Apple Laserwriter do things that I could not make it do via Aldus Pagemaker by taking directly to its Postscript engine. 

Looking back on it now, this was as far down the rabbit hole of “documents as computer programs” that I ever went. Later I would discover TeX and find it in many ways easier to work with than programming Postscript directly. My career started to take me into computer graphics rather than document publishing. For a few years I was much more concerned with Bezier Curves and Bitblits[12] using a Texas Instruments TMS 34010[13] to generate realtime displays of financial futures time-series analysis (A field known as technical analysis in the world of financial trading [14]).

It would be some years before I came back to the world of documents and when I did, my approach route back, caused me to revisit my “documents as programs” world view from the ground up.

It all started with a database program for the PC called dBase by Ashton Tate[15]. Starting from the perspective of a database made all the difference to my world view. More on that, next time.


Tuesday, January 02, 2018

What is a document? - Part 3


Back in 1983, I interacted with computers in three main ways. First, I had access to a cantankerous digital logic board [1] which allowed me to play around with boolean logic via physical wires and switches.

Second I had access to a Rockwell 6502 machine with 1k of RAM (that's 1 kilobyte) which had a callous-forming keyboard and a single line (not single monitor – single line) LED display called an Aim 65[2]. Third, at home I had a Sinclair ZX80 [3] which I could hook up to a black and white TV set and get a whopping 256 x 192 pixel display.

Back then, I had a fascination with the idea of printing stuff out from a computer. An early indication – that I completely blanked on at the time – that I was genetically predisposed to an interest in typesetting/publishing. The Aim 65 printed to a cash register roll which was not terribly exciting (another early indicator that I blanked on at the time). The ZX80 did not have a printer at all...home printing was not a thing back in 1984. In 1984 however, the Powers That Be in TCD gave us second year computer science newbies rationed access to a Vax 11/870, with glorious Adm3a[4] terminals.

In a small basement terminal room on Pearst St, in Dublin, there was a clutch of these terminals and we would eagerly stumble down the stairs at the appointed times, to get at them. Beating time in the corner of that terminal room, most days, was a huge, noisy dot matrix printer[5], endlessly chewing boxes of green/white striped continuous computer paper. I would stare at it as it worked. In particular, finding it particularly fascinating that it could create bold text by the clever trick of backing up the print head and re-doing text with a fresh layer of ink.

We had access to a basic e-mail system on the Vax. One day, I received an e-mail from a classmate (sender lost in the mists of time) in which one of the words was drawn to the screen twice in quick succession as the text scrolled on the screen (these were 300 baud terminals - the text appeared character by character, line by line, from top to bottom). Fascinated by this, I printed out the e-mail, and found that the twice-drawn word ended up in bold on paper.

"What magic is this?", I thought.  By looking under the hood of the text file, I found that the highlighted word – I believe it was the word “party” – came out in bold because five control characters (Control-H [5] characters[6]) had been placed right after the word. When displayed on screen, the ADM3a terminal drew the word, then backed up 5 spaces because of the Control-H's, then drew the word again. When printed, the printer did the same but because ink is cumulative, the word came out in bold. Ha!

Looking back on it, this was the moment when it occurred to me that text files could be more that simply text. They could also include instructions and these instructions could do all sorts of interesting things to a document when it was printed/displayed...As luck would have it, I also had access to a wide-carriage Epson FX80[7] dot matrix printer through a part-time programming job I had while in college.

Taking the Number 51 bus to college from Clondalkin in the mornings, I read the Epson FX-80 manual from cover to cover. Armed with a photocopy of the “escape codes”[8] page, I was soon a dab hand at getting text to print out in bold, condensed, strike-through, different font sizes...

After a while, my Epson FX-80 explorations ran out of steam. I basically ran out of codes to play with. There was a finite set of them to choose from. Also, it became very apparent to me that littering my text files with these codes was an ugly and error prone way to get nice print outs. I began to search for a better way.  The “better way” for me had two related parts. By day, on the Vax 11/780 I found out about a program called Runoff[9]. And by night I found out about a word-processor called Worstar[10].

Using Runoff, I did not have to embed, say, Epson FX80 codes into my text files, I could embed more abstract commands that the program would then translate to printer-specific commands when needed. I remember using “.br” to create a line break (ring any bells, HTML people?). “.bp” began a new page, “.ad” right-aligned text. etc.

Using Wordstar on an Apple II machine running CP/M (I forgot to mention I had access to one of them also...I wrote my first ever spreadsheet in Visicalc on this machine, but that is another story.) I could so something similar. I could add in control codes for formatting and it would translate for the current printer as required.

So far, everything I was using to mess around with documents was based on visible coding systems. i.e. the coding added to the documents was always visible on the screen interspersed with the text. So far also, the codes added to the documents where all control codes. i.e. imperative instructions about how a document should be formatted.

The significance of this fact only became clear to me later but before we get there, I need to say a few words about my early time with Wordperfect on an IBM PC XT. My first encounter with a pixel-based user interface – it was called GEM [11] and ran on top of DOS on IBM PCs. An early desktop publishing system called Ventura Publisher from Ventura Software which ran on GEM. I also need to say a little about the hernia-generating Apple Lisa[12] that I once had to carry up a spiral stair-case. 

Oh, and the mind blowing moment I first used Aldus Pagemaker[13] on a Fat Mac 512k[14] to produce a two columned sales brochure on an Apple Laserwriter[15] and discovered the joys of Postscript.

Next : What is a document? - Part 4.

[5] Similar to this http://bit.ly/2CFhue9