There are too many words to read, to listen to, to watch. It's increasingly hard to find: for readers, what would really most help; and for writers, the audience who would most benefit. Prompted by friends, a new, complex idea came to me, based on my CHOICE ideas, which could drastically reduce the amount and improve the relevance of what one has to read. I only have an outline, not completely detailed, so there is plenty of opportunity to join in to help me bring it alive.

I'm writing this primarily for people who already have an interest in CHOICE / ReGenMatch, as this represents a significant step forward in the scope of the approach. I'm also writing it for anyone with a very intelligent long-term interest in solutions to the challenge of information glut, and how writers and readers can connect more effectively. Please get in touch if you're interested in developing a service along these lines.

Too much information, yes; and so many ways to try to find it. (Will Google always give you what you want? Does Wikipedia always give you the right leads at your level? Hardly!) You could call the question of where and how to find the information, a meta-information challenge. Metadata people: librarians; Dublin Core; Schema.org and so many other good people, have made great efforts in their own way.

You can relate this to what you might call “information literacy”, and that is still a very useful skill for people to have, enabling you to find the information you want in a library or collection. But is it all down to you and your skills? In libraries, there are librarians, and in some good bookshops there are knowledgeable staff. These people can, of course, help you find what you want, if you are clear what you want. But there is another possibility: if they get to know a little about you, they might also be able to help you know what there is, in the library or bookshop, that would suit you. Would a large language model be able to serve as this kind of helpful librarian for the Internet? For sure, it's easy to imagine even better, AI-enhanced, search services to help you find what you say you want. But a LLM is about what is out there, not about you, so I doubt that, any time soon, a LLM will be able to use knowledge about you to help you know what you yourself want, and what is looking for you. There are questions which no amount of experience and skill in crafting LLM or search engine queries will enable you to answer. Unless, that is, you are happy with sharing a whole lot of information about yourself with the AI service. Does that feel secure and comfortable?

Let's turn the tables, and consider you as a writer, rather than a reader. From that perspective, it becomes very clear that AI is not much use in finding the people who would really appreciate what you write. I've heard questions recently along the lines of: “How can I, as a writer, reach the audience I want?”. Search engines find material for viewers, not viewers for material. Reflecting on this, things dropped into place for me. That's what SEO is trying to do: search engine optimisation means trying to get the search engines to allow readers to find your pages, rather than the pages that other people have written. That's rather like how advertising is trying to get people to notice a particular product or service, rather than some other competing product or service. Advertising and search-engine optimisation are so clearly zero-sum games. What is more, search engines are commercial services, and their priority will inevitably be to serve their paying customers, the advertisers, first.

What if we could, somehow, escape from those zero-sum games? What if readers and writers could collaborate in finding each other? If readers can find what they really most want or need to read, and spend less time on distractions and items of lesser value, then writers will also benefit from the best quality of readers and reading they can have.

Recently (2024-03-11) I complained about us – non-celebrity, common people – using commercial platforms like Substack or Medium to reach other common people like us. Well, yeah, that's one attempt at solving the question of writers finding audiences: write on a commercial service, and offload the challenge of finding your audience to that service. The downside: then, hold your nose when reflecting on where the money comes from and who sucks it up. For advertising, search engines are not really different in principle; it's just that they lead to recommendations for products and services, while the recommendations given to you by a publication platform will be for reading other writings.

So let's broaden out again to the more general question of how writers and readers find each other. With old-fashioned letter writing, the writer is writing for a very specific reader. Fine on paper. But then, once we get the easy replicability of e-mail, there comes the problem with unwanted Cc:ing; with spam; with phishing.

Another approach is to focus around topics. You write “about something”, and I want to read “about something”, and our two somethings somehow coincide. We see this in newsgroups; and in threads on platforms like Discourse, Slack or Discord. Related, but slightly different, are individual subscriptions to authors or channels — this is more like “about me” than “about something”, but anyway, we have RSS feeds for blogs; subscriptions on YouTube, and many more.

These are human variations on the theme of what is called the publish–subscribe pattern, or “pub-sub” for short. (If you search for that you'll get a lot about the internal computer system communication, and only a little about human-to-human communication.) In my experience, pub-sub services have some advantages over plain search, but still as a reader, if I subscribe to everything of potential interest, I am still wading through so much … content. Still much too much. And if I don't subscribe, I am liable to FOMO. So what do I want?

Back to writing, how many people take trouble in selecting their audience and writing for them? I've noticed that professional, successful writers often make this explicit. One book on my bookshelf has a page called “Is this book for you?”, helping readers know what is promised, and whether to take the time to read it. Many other books do it less explicitly, with introductions, forewords, prefaces, etc. Some rely on the reputation of the authors, or what other people say about them.

But for the median, commoner writer like us who doesn't make a living from writing, we sometimes completely forget about who we are writing for. Maybe we are actually writing for ourselves? Even if we know who we are writing for, we often don't make that explicit. Luckily enough, I did make it explicit in the only book I have had commercially published, in 2009. My Introduction starts with a “guide for readers”, beginning:

Who are you, O reader? Depending on who you are, or really, at this point, what you are interested in, I’d like to point each of you to the parts of this book that are most relevant to you as individuals, but first I’ll start by mentioning what the parts contain.

Also in the introduction I have a short statement about What this book is not, meant to head off potential disappointment. The Preface is talking about how e-portfolios need to focus on who they are being written for — clearly closely related.

But do these attempts answer the real question? Let's ask that question more directly.

Let's start with what I want from reading and writing (do you share my desires here?). If I had a magic wand, what would I conjure into existence? Here's my imaginary answer …

As someone looking for information, my interests vary from day to day, sometimes from hour to hour. Naturally, I want to read or see what interests me at present. I don't want to be distracted by what pulls at my instinctive base senses, but I want to focus either on what brings me joy, or on what is really mine to do; to contribute; to give to the world; both are to do with what makes me more fully human. I want a service that I can ask, “tell me what I need to know to move forward most effectively with my life goals” or something like that, adding “and don't give me anything that will distract me!”

Conversely, as a creator or curator of knowledge, I want to know, who would really benefit from what I am writing about? To whom would it bring joy? Perhaps, yes, who will find it of interest, of entertainment; but more pointedly, who will use it for the good of the world?

And here things start to interest me even more. How does what I am writing fit in with the bigger picture? What is the need out there that I could contribute towards? Working back from that, I would love a service that helps me discover what is uniquely mine to do; what my purpose in life is. Now, my belief is that most of us, at least, cannot discover that through introspection alone, or solitary meditation, prayer, or anything like that. And to follow that thought through, what I would find absolutely awesome would be a service that would use information about what other people are looking for, to guide me in what I can best be writing about, which would naturally as good as guarantee me a readership or audience.

And to complete the circle, I would like it if what I expressed that I wanted could guide other writers to writing things that are an even better match to my interest than what I find through search engines.

So now I turn back for a moment to my old ideas on CHOICE currently branded as ReGenMatch. I was thinking, for all these years, that CHOICE had to do just with people meeting people, not simply through what they were writing or reading. For sure, shared interests will feature in any case, but the CHOICE architecture wasn't originally intended for focusing on finding anything other than people.

The really smart thing about CHOICE is that the whole process is inherently two-way, or both ways if you prefer. You only find me if at the same time I am looking for someone like you. An unlimited number of questions, of which you only need to answer a few, is made possible through the technique of asking you to answer just those questions that matter to the people you are interested in. A part of this is the process of discovery. The process of finding people who really fit together is complex, not just complicated. And the way it seems natural to do this is by exploring. I see the process in practice being like improvised dance, changing requirements and adding answers until matches emerge. This will also be a process of self-discovery, discovering where what we want overlaps with what there is that wants us.

So, yes, we could have questions about what you're interested in, but in this matter of reading and writing words there is a lot more than just that. Indeed, connecting over a single topic of interest is what has been the dominant paradigm all these years, like library classification systems like the Dewey Decimal Classification or the Library of Congress Classification. Faceted classification systems like the Universal Decimal Classification are more expressive, and also widely used by libraries. But who, apart from librarians / information science professionals are going to use these? Given that there isn't even one universal standard, I can't see any common writers taking the (great) trouble to index what they write in this way. And even if they did, would that solve the challenge?

Moving forward, I had a promising development of viewpoint when it was pointed out to me that, unlike the application of classic dating, for recruitment and employment there it is probably easier to see the challenge as involving three things: the job-seeker, the employer, and the job itself. Maybe we could see a job offer as a little like a piece of writing, seeking readers? So

This is what has brought me now to looking deeper into how CHOICE/ReGenMatch might handle writing, and see if that approach has any merit for recruitment. If it does, we have a greatly promising advance!

I'm a writer, writing about various topics, and you can see many of those topics in my list of terms that have appeared in the last year or two in my writings. I want people to read what I write. But I recognise that reading time is limited (sometimes severely!). I want only the right people to read what I write, not just anyone.

So now I'll lay out my current ideas on how we might do this within the CHOICE methodology. The conceptual model for CHOICE needs some readjustment; we'll look at how this could be done; then I'll say more about how I see this working out in practice.

As suggested above, the CHOICE conceptual model now needs to include three things, and the attributes of three things:

1. the authors or publishers of the work;
2. the people viewing, reading or using the work;
3. the work itself.

The first two are no different from ReGenMatch and previous conceptions of CHOICE. What then of the work itself? Based on what I'm familiar with from the world of metadata standards, a reasonable starting point for representing a creative work is some of the schema.org CreativeWork, optionally enhanced by some of LRMI. A quick look at the CreativeWork in schema.org reveals a long list of properties, many of which are represented simply by text. But text by itself this won't work for the CHOICE approach — we need a question structure that doesn't rely on words entered by users, so that they are equally meaningful and accessible to both parties, and can be chosen as desired. Let's look at this in more detail.

There is a lot of work to do here, answering these questions.

• What kinds of work we are going to have in the system? Writing? Audio? Video? Anything else?
• For each kind of work, what properties / attributes / characteristics will be represented our system?
• For each characteristic, how is it going to be represented in a way that is readily understandable by both creators and readers / viewers?

Doing this all in detail is beyond the scope of what I'm writing now. But let me try to draft some ideas on the simplest case, writing, such as blogs, I would expect something like this, as an initial list of attributes / characteristics / properties of a piece of writing, for the purposes of CHOICE.

• creative authorship: who wrote it, not necessarily who is curating it
  • this will need to be an internal ID, to maintain privacy if wanted
• a person who is effectively curating the writing matching process
  • these people have to be on the system and responding to questions
  • this will also need to be an internal ID, for the same reasons
  • this may or may not be the same as the author
• Creative Commons licence, or similar
• language (human, natural, or otherwise)
• number of words
• reading difficulty level
• conceptual level
  • in normal teaching materials for school or student use, this is often specified in terms of the typical age of the reader
  • for adults, and topics not typically covered in educational institutions, this might best be done in terms of a list of the different levels that make sense to common readers and writers: created and maintained as a commons by a partner organisation
• list of main and subsidiary topics
  • this may be the hardest part of the whole exercise
  • not any old user-defined tags (i.e. not a “folksonomy”)
  • AI may be of some initial help here
  • we would again be wanting to create a community maintained knowledge commons around this
• platform, or technical format, as this may affect how it can be accessed
  • including information about any paywall or other restrictions
• accessibility features
  • these will need to be carefully agreed and defined
• relationship to people, things, concepts, or other works
  • the relationship terms to use here need quite some thought, ideally built and maintained as a knowledge commons
  • one particularly useful feature is, what concepts does this piece assume the reader is familiar with — these could be called pre-requisite concepts

Some of these can naturally be generated automatically. For the ones that cannot be, we will need to do a lot of ontological commoning to evolve and maintain a suitable ways of representing them. This is serious ontology work. We can draw inspiration from schema.org, LRMI and other standards, but the key here is that we are ensuring, as far as practically possible, that everything is understandable to common people, and that the process of choosing is also as straightforward as can reasonably be achieved.

It may be worth emphasising here that one of the core values in all CHOICE work will always be doing things in as simple a way as possible, inclusively and accessibly, probably through something like ontological commoning into knowledge commons.

But to come back to the vital point here, remember that none of this information is mandatory. The CHOICE paradigm is that people only need to answer questions about themselves, or in this case the works they are curating, when it is relevant to the people that the work is aimed at, or designed for. To me, this puts a quite different spin on all this metadata. It is informative, if people genuinely want to know, but is not an obstacle, if not needed.

The key principle of CHOICE questions is that if particular answers to a question are able to be required on one side, those questions have to be answerable on the other; and vice versa. But exactly how to handle this for creative works is not completely clear to me. My guess is that it will become clearer as we look into real examples of what people would actually like to read, what writers would like to write, and what they imagine their readers to be like.

One obvious challenge is that many writers simply want to be read, by anybody. How do we stop this flooding the inboxes of many readers? One promising way to counter that is by allowing a reader to require that an author specify a particular feature either of themselves or of their work, before a “match” is made. For example, as a reader I could ask to see only writings that have been explicitly targeted at people over 60. In fact I wouldn't do that, as I would risk being flooded by material about cruises…

It is hard to know how much complexity to allow in readers', or indeed in authors' requirements. My sense is that we won't know this until people try out the system, and then actual experience will shape the answers quite quickly. This, to me, is typical of a complex human task — we won't know how it's best done until we've tried it.

So … when a reader who matches a writer's imagined audience profile finds a piece of writing that matches what they are looking for, what happens? I can imagine various scenarios.

1. Possibly the simplest is that links to the matching piece or pieces of work are delivered to the reader. Writers would have the choice either to reveal their identity and some contact method; or to try to remain anonymous, remembering that it is very hard to protect anonymity when you have written anything substantial. The reader could choose to reveal their identity as well, but may well prefer to remain anonymous.
2. This could be a one-off search, or there could be the equivalent of a subscription service, where the reader opted to continue to receive any other pieces of work that matched both ways.
3. The reader could rate the piece of work either as a general usefulness rating, or in terms of whether it lives up to any of the particular features or characteristics that formed the reader's requirements. This could potentially work in with a collaborative filtering approach, as has been in use for at least all this century!
4. The reader or the writer could opt to propose personal contact. This would be in the same way as for ReGenMatch. Each side could examine the matched details of the other before agreeing to mutual revealing of personal details. In essence, this would be very similar to using ReGenMatch, but with the reader using the writer's writing as an extra criterion to know whether to initiate contact.

I'll keep this short. This the outline of what I'd like to merge into the general approach of CHOICE / ReGenMatch.

• Recognise the distinction (albeit not black and white) between two-way matching of people and both-ways approaches that include details about the relationship or its focus as a separate thing.
• This will mean considering, first, employment (I'll change that page accordingly) and then the other enquiry types. In each case, what is the “third party” “thing” type?
• For each type of thing identified, get working on establishing a commons ontology for that thing. Yes, and we remain open to inspiration from schema.org
• Find partner organisations that can help develop CHOICE-like question sets for each type of thing.
• If any two kinds of thing have substantially the same question sets, treat them as the same kind of thing. This will avoid proliferating kinds of thing unnecessarily.

That will be plenty for the time being!

Contact me to get an account on this wiki which will allow you to comment. Or be e-mailed with updates…