As our investigations relating to collecting practices in the Icon Age have taken a somewhat different turn since we first undertook this work, we have decided to terminate work on this Knowledge Development journey. As noted in the previous post, a wide variety of issues were revealed by the study. It made it clear that the texts selected for mark-up in the present day were often, but not always, distinct from those that had been selected many years before. This was not, in itself, surprising, but we wanted to understand what trajectories of reasoning had changed the most and this was not so easy to pinpoint. We were interested in whether things like the type of document marked up, when it was originally marked up, the place of work at the time, and the interest in the topic would make a difference to the degree of contiguity between the original markup and the new markup. However, none of these seemed to have a consistent effect. More than anything, we have concluded that understanding the nature and relevance of document markup practices for knowledge development is something that would require significant further investigation in its own right.
The second of our investigations involved me marking up 19 documents on which I had previously placed sidebars next to parts of the text between 22 and 39 years ago. This was completed in August last year, and, since then, we’ve been analysing the results. My collaborator, Peter Tolmie, made a numerical assessment of the match between my mark-up of the original and my new mark-up; and I reviewed my original and new mark-ups and documented my comments on what I had done. We then recorded a one-hour video conference in which we discussed what had occurred with three of the documents – we had no time to discuss any more because such a wide variety of general points emerged from this conversation. The recording was processed through the Otter transcription service from which a written transcript of the session was produced. Peter is now planning to produce an overall interim report on our findings to date.
Our original plan (documented in the post for 10th August 2019) involved undertaking a further similar investigation of mark-ups on documents; but this time conducted in the form of a face-to-face interview in order to be able to examine the reasoning for what is marked up in greater detail, and to discuss the findings of the previous two investigations. However, the pandemic has interrupted these plans. We still feel this final investigation is worthwhile undertaking if and when travel restrictions are lifted. In the meantime, however, the Otter transcript and the interim report Peter plans to produce, will be the primary outputs from which we hope to be able to summarise some hard and fast findings in a post later this year.
The second of our investigations into the memorability and impact of information nuggets focuses on work documents that I read decades ago. I used to draw lines next to text I thought significant, so our experiment has taken a random sample of nineteen such documents, removed the marks I drew back then, and had me re-read and re-mark them.
Identifying and preparing the documents was quite a demanding process in its own right. My collaborator, Peter Tolmie, sent me document reference numbers identified by using a random number generator to select items from the 17,000+ entries in my document index. I then used Windows Explorer to pick up the first file in each reference number folder (making sure that the descriptive part of the file name was not visible) and sent the files to Peter who assessed if they possessed any marks. It took us about 5 iterations of this process, working with around 200 Ref Nos in all, to obtain nineteen suitable documents. Peter then removed all marks from the digital documents by a combination of cropping and overlaying white boxes, and sent the finished files to me.
I marked up the documents over 14-15th August, using the same reading approach I believe I have always used since those days i.e. not a detailed line by line read but more of a rapid scan through to pick up the gist of the contents and to identify key text to which I pay more attention and from which nuggets are drawn. There were about 190 pages to get through across the nineteen documents, and I used my PDF application to mark up the nuggets in a tasteful shade of green.
I had vague recollections of some of the documents, and no recollection at all of others. However, I don’t feel this particularly affected my choice of nuggets. Nor do I think that the context in which I was re-reading the documents (i.e my current retired state as opposed to the work I was doing at the time when I originally encountered the documents) was influencing my nugget selection. I started to think that perhaps my selections would be the same as the selections made by anyone – almost as though each document possesses some elements which are inherently nuggets in their own right regardless of reader. However, I did come across a few exceptions to this: for example, I marked up one short para simply because it mentioned the name of someone I knew. Another document was very specific to the organisation I worked for and I suspect my choice of nuggets was influenced by my own particular perspectives on the topics being addressed. This observation has made me muse about the possibility that each document may possess more or less ‘inherent’ nuggets depending on its place in a spectrum of document types ranging from general purpose article to company specific work text.
The exercise has also got me thinking about the difference between summary text and nuggets. Sometimes a short para summarising some key points is worth highlighting simply because it’s a quick route into the document. However, it begs the question as to whether the points being made within the summarising para, are of any great value.
These are all questions which I’m anticipating we will address at some point downstream. However, the immediate priority is to analyse the results of the exercises we have conducted. For this latter marking up exercise, my new mark- ups will be compared against the original mark-ups to see if there is any similarity. We’ll be posting the results here sometime in the next 12 months.
I’ve just acted as subject in our first investigation into the memorability and impact of information nuggets. The nugget material, in this case, was mindmaps of key points in nineteen esoteric-type books which explore perceived unresolved mysteries from ancient Egyptology to modern secret societies. I discovered that I could remember almost none of the points presented to me and was unable to link any of them to a particular development in my thinking. My immediate reaction to this disappointing – but probably to be expected – finding was that these are not actually nuggets of information but instead are just parts of a summary of each book.
However, on reflection, I’ve reversed that view. After all, when I was picking out the points as I read, I must have thought each of them to be significant – otherwise I wouldn’t have picked them out. So, how is a key point in a book different from a key point in, say, a five page article? Well there are some obvious differences like the book is a lot bigger and has a lot more stuff in it – most of which I’m not familiar with AT ALL. Unless one has a photographic or otherwise superb memory, you wouldn’t expect to remember everything in such a book after one quick casual read. Of course, I have the books on my bookshelf and have the look of each one locked in my memory with some ideas of what it’s about. However, this is the case because there are just a few hundred of them, and they have a rich content and the covers and spine usually have distinctively memorable images. In contrast, the articles and documents in my work collection (which are due to be investigated next), are much more numerous; are hidden away in my computer (with just a few in my physical archive box); and they all look very similar and have very few distinctive markings.
I guess I’ve expanded my thinking this morning about all this. However, I’m only the subject and we’re only half way through the overall exercise. The interesting bit will be what the researcher concludes from it all.
This entry has been jointly authored by Peter Tolmie and Paul Wilson
A key part of the motivation for keeping texts is that they contain information – nuggets – that have some kind of future value. It’s rare for a whole document to be seen as having such value in its entirety. Nuggets are more often certain sentences or paragraphs. The question is, what happens to that value over the lifespan of an archive. Does the value of specific nuggets persist? Does it change? Does it grow or reduce in relevance? Does it become eroded to the point of obsoletion? And, given the same documents at some later point in time, would the same nuggets be identified, or would something else stand out as being more important?
To assess the use and impact of identifying nuggets, three separate investigations will be conducted using documents containing nuggets that were collected over the last 38 years. In all cases, investigations will be undertaken with reference to the individual who originally identified the nuggets. Two of the investigations will focus on the individual reflecting himself upon his use of nuggets. The third investigation will focus upon extracting information about the nuggets and their use through discussion.
The investigations will be performed using two separate sets of material:
Set 1: A set of 19 MindMaps relating to esoteric books.
Set 2: Documents from the PAWDOC collection.
Investigation 1: The first investigation will be a written exercise using the set of Mindmaps and will attempt to assess:
a) Whether individual nuggets can still be recalled;
b) Whether the sources of those individual nuggets can still be recalled;
c) How significant specific nuggets are considered to be by the individual concerned;
d) What other nuggets, if any, are associated with the one in question;
e) What specific concept(s) individual nuggets are believed to have contributed towards.
Investigation 2: The second investigation will also be a written exercise and will explore what nuggets, if any, can be identified by the individual concerned in unmarked versions of randomly selected documents from the PAWDOC collection where nuggets have previously been identified. From this exercise it is hoped to deduce:
i) Whether nuggets lose their vitality over time, with new concepts being derived and becoming established.
ii) Whether looking at the documents anew, within a distinct context, will lead to the identification of different nuggets with different relevance to the individual.
Note that the first 2 investigations are interleaved so as to be able to maximise the amount of discussion possible in the concluding interview.
Investigation 3: This investigation will be conducted as an interview and will explore similar themes to those tackled in the prior investigations but in a more open-ended way, so that the reasoning involved in retaining documents for the sake of specific features can be examined in greater detail. The interview will also seek to examine in detail topics that have spanned all three investigations.
Investigation 1: Nineteen nuggets, each from separate MindMaps, will be randomly selected from the MindMaps of books on esoteric subjects. Randomisation will be achieved by creating a template that divides an A4 page into 24 numbered areas of equal size. A random number generator will then be used to create nineteen separate numbers ranging from 1 to 24. The Researcher, Peter Tolmie, will generate the numbers and use each one in conjunction with one of the MindMaps and the template to identify the nugget(s) present in that location. If more than one nugget occurs in that location, the topmost one will be selected.
The Researcher will assemble the nineteen nuggets and send them to the PAWDOC Owner and Subject, Paul Wilson, who will be asked to provide a written answer to the following questions in relation to each one:
- Do you remember where this nugget came from?
- Do you remember why you might have marked this out as a nugget?
- How important do you consider this nugget to be now?
- Do you remember what other nuggets were associated with this one?
- How did you use this nugget and what other things did you develop on the back of it?
The Subject will then be sent the original MindMaps in which the nuggets appeared, and will be asked the following questions for each one:
- Do you remember more about why you marked this as a nugget now?
- Now you see the MindMap it came from and the other nuggets it was associated with, do you see it as more or less important?
- Do you remember anything more now about how you used it or what other ideas it may have contributed to?
The Subject will return his responses to the Researcher, who will categorise them and place the results in an analysis spreadsheet. The overall analysis and specific instances regarding the Subject’s responses and reactions will be written up as the findings for Investigation 1.
Investigation 2: Nineteen separate documents, each containing nuggets, will be randomly selected from documents included in the PAWDOC collection between 1981 and 2011. Randomisation will be achieved by using a random number generator to create numbers relating to the 16925 entries in the PAWDOC Index sequenced in the order in which they were created. The Researcher will generate the numbers and use his copy of the Excel version of the Index to identify the first nineteen of the entries for which an electronic file exists (some entries just contain the information they relate to and have no associated documents – these are identifiable by the contents of the Movement Status field); and for which the associated electronic files are likely to contain highlighted nuggets (items such as, for example, Health & Safety booklets, are unlikely to contain highlighted nuggets). He will send the list of numbers to the Subject who will open the folder labelled with each particular number, take a copy of the first file that appears in each folder and will send the files back to the Researcher. The Subject will take as little notice of the file titles as possible by setting up the Windows File Explorer window to display only the beginning of the file name so that only the Reference Number, which is always at the start of the file name, is visible. Should the Researcher deem any of these files to be unsuitable, he will send additional numbers to the Subject until a satisfactory set of nineteen documents that contain nuggets has been obtained. Then, using a cropping tool or by obtaining a clean copy of the document from elsewhere, he will send clean unmarked copies back to the Subject. The Subject will read the documents, mark up any text that he considers to be nuggets, and will send the documents back to the Researcher.
The Researcher will then make a comparison between the original nuggets identified and the new nuggets identified and record the results in the analysis spreadsheet. The overall analysis will be written up as the findings for Investigation 2.
Investigation 3: Nineteen nuggets, each from a separate document, will be randomly selected from randomly selected documents placed in the PAWDOC collection between 1981 and 2011. Randomisation will be achieved by using the same procedure employed in Investigation 2. The Researcher will take the first nineteen suitable documents that contain nuggets and randomly select one specific nugget from each document using a random number generator. The nineteen nuggets will be assembled together and presented to the Subject who will be asked to answer a similar set of questions to the first set of questions in Investigation 1. These questions will form the rough frame for the first part of an interview in which the various nuggets will be discussed.
The Subject will then be shown the original documents in which the nuggets appeared, and will be asked the following questions for each one, which will form the second part of the interview mentioned above:
- Do you remember more about why you marked this as a nugget now?
- Now you see the document it came from and the other nuggets it was associated with, do you see it as more or less important?
- Do you remember anything more now about how you used it or what other ideas it may have contributed to?
- Are there things in the original document that you didn’t mark as a nugget at the time that you would mark as a nugget now? If so, why?
A third and final part of the interview will explore the responses from across all of the assessments in a more open-ended fashion to generate deeper insights and discussion.
The Researcher will then transcribe the interview, categorise the responses and place the results in the analysis spreadsheet. The overall analysis and specific instances regarding the Subject’s responses and reactions will be written up as the findings for Investigation 3.
Conclusions: The Researcher will use the findings from all three Investigations to write the overall conclusions of the investigation.
To review what I’ve done on the topic of information nuggets, I’ve been trawling through the PAWDOC Index and files. The earliest example of sidelined text that I can find in my document collection was from October 1981 when I was working at the National Computing Centre. I can’t remember why I started to do it – but it may well have been prompted by the method that NCC’s Chief Editor, Geoff Simons, used to construct his books. He explained to me that he read everything about a subject, identified key points and put them on Post-it notes which he stuck on the wall. When he was ready to write, he rearranged the Post-its into separate sections and in sequence within the sections – and I saw examples of this in his office. At some point I started to employ this technique to construct the best practice books I wrote at NCC – but using the word processor on our new Zynar Office System to assemble and organise the key points. Sidelining text was an obvious way to identify key material to feed into that process.
Around 1994 I started talking with City University academics Clive Holtham and David Bawdon with a view to undertaking a joint project on ‘The Paperless Office Worker’. A key strand of this work would involve me digitising my PAWDOC collection. Extracts from emails between us in the early part of 1994 included the following:
Email from Clive Holtham: ‘….I don’t record each document as Paul does, but file at a quite detailed level. What I am conscious of is how much I forget about what I already have. The equivalent of underlining is important – we need to consider something more than keywords to store with each piece.’
Reply from Paul Wilson ‘…I agree with needing to deal with the underlining problem. I have sidebars on most of my material – they are the information nuggets; but I don’t know how much use they would be out of context.’
Email from Paul Wilson: ‘… with reports , papers etc. I usually mark the nuggets of info within them – presumably these are the bits of information I really want.’
The collaboration with City University proved very productive: Clive Holtham introduced me to a product Manager in Fujitsu who loaned me a scanner; and to the owner of a small company called DDS who loaned me the Paperclip document management software. Soon I was scanning my existing paper documents and new ones as they arrived. In January 1997 I issued my 3rd briefing note on these activities and included the following towards the end of the four page document:
Despite the close relationship between filing and information use, contemporary filing systems provide little other than title and index fields to support the knowledge acquisition, synthesis and use process. Filing systems, it seems, are there just to store items and to aid their retrieval. Unfortunately, the personal knowledge acquisition, synthesis and use process is not supported adequately outside filing systems either. Some standalone packages do exist, but they are not intended to be used in a day to day manner for personal knowledge acquired in documents, electronic files and other artefacts. In fact, even the need for such support is not widely recognised.
It is not yet clear to me what support could be most beneficial. However the clues are littered throughout the practices of knowledge workers like myself. For example, whenever I read articles, papers and reports I always mark the good bits – the nuggets of information. These are key points which I particularly want to augment the knowledge in my brain. Sometimes when I have been researching a topic I collect together all the nuggets I can find, categorise them and reorder them, and synthesise a new view of the topic in question. Unfortunately, like the magnesium nodules on the floors of the deep ocean, huge numbers of nuggets now litter my filing system unseen and inaccessible. I hope they are in my brain and that they have been used to develop my current state of thinking – but I’m not so sure that their huge potential has been fully exploited.
For my filing activities to really start adding value I need tools which can record those nuggets as I consume and index each item, and which can enable me to reorganise those nuggets, add more nuggets, and synthesise new nuggets, in the process of actively developing my ideas. Such tools would, of course, maintain the links to the original source material stored in my files. And my files would become a combination of original source material and the representations of my developing thoughts and ideas.
Now that I am confident that I have the paper scanning and electronic file indexing activities under reasonable control, it seems high time to start addressing the critical area of information use and its role in knowledge management.
These are the earliest mentions I can find of information nuggets in the PAWDOC collection – and they give no indication of where I picked up the concept from. In fact I can only find one published mention of the term and that was in the Lotus Notes-oriented magazine ‘Groupware and Communications Newsletter’ from April 1998 in which one Ted Howard-Jones gave a brief description of a service implemented by a major financial institution to capture competitive information. He wrote:
‘Called Report-It!, this service captures knowledge using a secure voice-mail system and delivers categorised information directly to the desktops of office-bound managers and competitive information professionals. These nuggets of professional information are disseminated via Notes.’
The use of the term Knowledge in this article, and in my briefing paper mentioned above, reflected the fact that, in the late 1980s, the term Knowledge Management started to became fashionable and by the late 1990s had become a holy grail of IT Professionals, Management Consultants, and Academics. The first mention of the term in the PAWDOC Index appears in 1990, and occurs in a further 137 Index entries from then to 2016. The company I worked for (Computer Sciences Corporation – CSC) was a global computer services organisation with tens of thousands of employees worldwide. Its involvement in the Knowledge Management topic came from three angles: first, its clients started to ask about it and how to do it; second, its consultants and salesmen saw it as a potential source of revenue; and third, its employees and management began to think that they needed it internally to improve the effectiveness of the business. Hence, as a consultant in the UK end of the business, I was aware of or got involved in:
- A number of initiatives to develop an offering or information for clients, including:
- the development of a KM service by CSC Netherlands in 1990;
- discussions with CSC UK Management Consultants who were developing KM propositions, in 1996-8;
- the definition and design of a KM service by CSC UK personnel to address opportunities in a number of clients including ICI Paints, LUCAS Engineering, John Menzies and United Distillers, in 1996-7;
- the publication a CSC Research Services Foundation report on KM in 1998;
- news of CSC’s Global Knowledge Management services, in 2001.
- The development of such systems for clients (primarily using web pages on intranets), including:
- a presentation to ICI Paints in 1992;
- the development of systems for KM and for a web-based ‘Gazateer’ of all KM, architecture and other organisational information, for the Nokia SCC – a new organisation being set up by CSC UK for Nokia, in 1997;
- the design and population of a web-based KM system for Dupont Agriculture’s architectural components, in 1998.
- the development and use of internal CSC solutions, including:
- attendance at internal CSC workshops on developing an organisational learning infrastructure in 1996;
- the development of an improvement process for CSC’s new application development organisation that was being designed and built from scratch, in 1998;
- the design of a practical KM programme for CSC UK’s reorganised Consulting & Systems Integration unit, in 1999;
- knowledge management work being done for BAE, in 1999;
- an internal Community of Interest on the subject of Personal Knowledge Management, in 2001.
Of course, my rather lowly personal filing perspective had to be rapidly expanded as I entered the Knowledge Management (KM) arena to accommodate both high level Management Consultancy notions of ‘Intellectual Capital’ and the distinction between Knowledge and Information; and the practical need to derive benefits from an investment in KM by effectively sharing the knowledge that had been acquired. Indeed, one of my contributions to the internal Community of Interest mentioned in the last bullet point above seems to herald a change in my thinking. My opening sentence reads:
‘Since collaborating with everyone in this shared space I’ve had my eyes opened to the concept that KM is all about enabling people to find things and work together, as opposed to the idea that KM is all about nailing down bits of knowledge and providing it to people. I realise that there is significant crossover between the two approaches – but nevertheless giving priority to one or the other will result in significantly different activities.’
During the 1990s I learnt a great deal about what people thought Knowledge Management was, and also about the powerful potential of the new web technology to support KM. However, throughout this period I don’t recall any specific conversations or documents about dealing with underlined, sidelined, or highlighted text. Indeed, by the time I completed the draft of the paper summarising my PAWDOC findings in June 2001, it seems that my ideas on the subject had advanced no further than that reported above. The paper was published in the journal Behaviour & Information Technology (BIT), and addressed the subject as follows in the section on ‘Areas of Investigation and Summary Findings’:
Q27. How can an electronic filing system be used to develop and use knowledge?
- Include substantive information in the index entries, for example phone numbers, book references, and expense claim amounts.
- Identify the nuggets of information (i.e. the valuable bits) when you first read a document
- Capture and structure the nuggets into the overall nugget-base at the same time as indexing the item
Status: ideas formed
Q28. What is the best way to capture and structure information nuggets?
Probably by using a Concept Development tool. Some initial prototyping has been done using the Visual Concepts package and the eMindMaps package.
Status: ideas formed
Q29. Is it feasible and practical to capture and structure information nuggets as well as indexing items?
Status: not started
Q30. Is it worthwhile building and developing an information nugget base?
Status: not started
The Concept Development prototyping mentioned in the answer to Q28 above probably started in April 2001 when I acquired a free copy of the eMindMaps software, and by the end of 2001 I had started making MindMaps of books on esoteric topics such as the Egyptian pyramids and the origin of Atlantis. I have no record of my detailed intentions in doing this, but I guess I wanted to experience the process of recording all the nuggets I found in a book – and then to explore what could be done to integrate and exploit the material from several different MindMaps. In all, I made MindMaps of 19 books over a two year period; but that’s as far as I got. At the end of 2001 I started a new job in bid management and my energies were increasingly taken up with managing very intensive bids, with documenting the bid process, and with operating a Lessons Learned programme. I had no time to pursue these information nugget ideas any further.
This concludes my review of my previous activities in the use of information nuggets. The questions I posed in the 2001 BIT paper still remain largely unanswered, and the operation of the PAWDOC system has not provided any further insights on the subject since then. However, the existence of the sidelined documents, and of the 19 MindMaps, do provide an opportunity to undertake some rudimentary practical work to explore if the information nuggets identified were memorable and of any use. Subsequent entries will outline the methods that will be used to undertake these investigations, and will report on the results.
In 1981 I was working in the newly formed Office Systems team in the UK National Computing Centre, and I was interested in how the new technology could support the management of an individual’s office documents. So, a colleague and I decided to experiment with our own documents. This was the start of a still-running practical exploration of how to manage personal documents using digital technology.
It was my practice to highlight key text with a side line as I read documents, and, as my document collection grew, I began to wonder how I could make explicit use of this very specific information. No doubt the act of highlighting was in itself helping me to assimilate documents; but I wasn’t sure if all the highlighted facts were being retained in my brain and being used to develop new concepts.
During the 1990s, the trendy new topic of Knowledge Management emerged which provided a recognised arena in which I was able to explore these ideas. Sometime during this period, I latched onto the term ‘nugget of information’ (the first published mention of this in my filing index is in an article by one Ted Howard-Jones in the March 1998 issue of the Groupware and Communications Newsletter). My attempts to relate lowly personal filing to the Knowledge Management field eventually fizzled out in the face of much sexier concepts such as an organisation’s ‘intellectual capital’. However, in the early 2000s, I did make a specific attempt to see if I could use Concept Mapping software to capture nuggets, by applying it to 19 new age books on the pyramids and the like; but that is where my knowledge nugget endeavours ended.
Now that I’m trying to find a home for my document collection, and to identify the findings from its long term operation, it seems a timely moment to review this particular aspect, to do some practical work on the nuggets I’ve identified over the years, and to draw some conclusions on the topic.