I want a space ...

I want a space ...

Something I wrote in a bit of a haze a couple of weeks back. It seems to hold together better than I'd feared, and addresses my goals in online media and discussion.

Originally shared by Edward Morbius

A long rambling post with strong, but well-founded, opinions

And no, it comes to no definite answers or conclusions. The path may still prove of interest to some.



I want a space where I can focus on the things I want to say, how I say and present them, and in which the people I most want to discuss them with either are present or can be readily included in the conversation. I think this is the nutshell attractive-pull of online publishing, social networking, whatever.

I want a space where interesting, relevant, truthful (where that matters), quality content can surface.

I want a space where others who share these ideals feel welcomed, rewarded, and safe, in participating, as writers (or musicians, or artists, ...), or as readers and viewers of the works presented.

I want a space that's fun. Not bog serious all the time.

I want a space that is safe. Where information that's meant for a limited number of eyes, or ears, reaches only those. Where ideas that may be disruptive or subversive can be explored, safely.

I want a space that does not reward stupidity, ignorance, malice, crassness, hate, oppression, prejudice, blind conformity, pretentiousness. These are the major aversion-push away from online publishing and social spaces.

I want a space that allows for some rough-and-tumble, but also recognises when it is presented with a fragile flower, and provides the space and protection to allow that to bloom.

I want a space that allows connections to be drawn between ideas, where conceptual relationships can be exposed and explored.

I want a space that allows for discovery and resurfacing of gems whose brilliance was not apparent when they were first created.

I want a space that challenges my beliefs, prejudices, statements, self-centeredness. That checks me when I am mistaken (though not to bully me into conformance with some ideology). That allows me to move from being wrong to being write.

I want a space that can present, and inspire beauty. But doesn't blind itself to ugliness where and as it exists either.



How the hell do I get there? Because I've been looking for 40 years. In person and online. Via Usenet, mailing lists, BBSes, Slashdot, HN, Reddit, Google+, Ello, Imzy, Mastodon, Diaspora, and others.

I've also been obsessing over the question of what media is, how it interacts with society and social institutions, and how each of these feeds back on the other. A topic I'm starting to feel I could write a book on, though there are some excellent ones out there. (Elizabeth Eisenstein's The Printing Press as an Agent of Change among the best, it was strongly inspired by McLuhan's The Gutenberg Galaxy.)

Which means that if I do succeed in my quest, I will almost certainly change myself and the realm I exist in.

Medium matters. Media is intermediate agency, as I precociously titled a G+ collection of mine beyond all conscious awareness of significance at the time. One side project is compling a list of communications methods and listing out characteristics of these: persistence, range, immediacy, bandwidth, arity (a new favourite word), latency, cognisability. Concepts on which Communications majors might have a considerable leg up on me.


Descending within potential sight of solid ground, in an online context, communications tools vary along another set of axes. Length. Formatting. Multimedia potential. Push vs. pull technologies. (We laughed at Push in the 1990s. It won.) Centralisation versus decentralisation. Standard vs. proprietary formats. Presence of encryption or authentication, for session or content, in flight or at rest. Transactional vs. agglomerative media -- streams vs. Wikis (or ledgers vs. balance sheets). Rewriting, editorial process & flow, verification, validation, fact-checking, peer review, wisdom of crowds, independent verification. Advertising, ideology, propaganda, disinformation, distraction, intimidation, dehumanisation. Overload. Encoding, formatting, presentation, transmission, request, metadata, table-of-authority, subject classification, and document identification, standards. The question of identity of author, editor, proofreader, fact-checker, publisher, payment processor, distributor, syndicator, and ... consumer.


Clearing through the clouds a bit: present communications systems tend to distinguish themselves by sensory channel, distribution medium and network, presentation, directory (a huge and central control point), notifications system (push) or publishing catalog (pull), length, and complexity of the delivered work.

A buzz or blink. Alarms. A tickler item. Onscreen notifications. A message delivered in an application context: SMS, MMS, ASCII mail, ISO-9660, UTF-8, HTML, RTF, PDF, Markdown, LaTeX, DOCX, JPG, PNG, GIF, SVG, MP3, MP4, WEBP, WEBM. Email, Web page, document, article, book, series, brochure, map, spreadsheet, Jupyter notebook.



In the beginning was the word.

Early Unix systems offered text. It was delivered via 'write', 'wall', 'talk', or 'mail', or could be placed in a .project or .plan. Usenet grouped message delivery by forum rather than recipient, growing out of email. IRC generalised talk into multi-user live chat. In a computing environment based on ubiquitous multiuser networked textual capabilities, these evolved naturally.

The graphical world -- Macintosh, later Windows, both MS and X11 -- introduced apps, and cut the direct line between receiving text and processing it. Unix pipes, bumps on a wire, allowed for user-supplied pre- and post-processing of messages as the were sent or received. Graphic applications tended (though, true, don't fully require) to remove the option of discretionary arbitrary message manipulation in ways not specifically supported within the application.

And the applications often support little. Some won't wrap text lines, or support find, let alone replace. Those that do offer advanced manipulation frequently require application- or vendor-specific tools or languages to do so.

They did, however, add the capacity for visuals, audio, and video to be supplied, and these were. Useful capabilities -- it is vastly more efficient to look at an object than to have someone attempt to describe it to you. But in much the same way, it's vastly more efficient to describe many complex ideas than to try to picture them. This post as a cartoon or set of ideographs would not be as detailed. (It might be shorter, and quite possibly more entertaining.)

There's a strong case for documents comprising more than simple text. Looking at books, document preparation systems, Web pages, and more, I've come to see a general hierachy of increasing structure.

Letters, words, sentences, paragraphs.

Sections, headings, metadata.

Emphasis: italic, bold, underline, strikeout.

Lists, unordered, ordered, hierarchical.

Blockquotes.

Formatted lines -- poetry, with fixed line endings.

Technical text: unformatted, monspace, code.

Superscript and subscript.

Extended characters: fractions, symbols, formulae, non-ASCII glyphs, non-Roman glyphs, typographic and other marks.

Unicode. The tenth circle of Hell.

Tables.

Notes and references.

Boxes, frames, and other page layout.

Line drawings and figures.

Shaded art.

Photographs, detailed images.

Animations. Video.

Audio.

Data.

Programs.



Apologies for spelling this out in obvious detail, but as you, more or less, descend through this stack, you're finding increasingly complex, but also increasingly capable, document types. The power can be (and is) abused, but it's also useful. The key is in discretion, objectives, rewards, and dissuasions.


Key aspects of online publishing are high speed, all-but-zero cost, interactivity, an often large canvas, interactivity of the work itself, access to mixed media, and the capacity to gather response from readers.

Or non-readers.

A typical interaction will consist of a text, direct responses (or reactions), and further discussion.

The text can be as simple as a few words, an image, a video, or a longer-form work. The responses may be like/dislike, tags, or commentary. Longer discussions may involve multi-reference responses to the original and/or one or more comments. We've seen this.

The discussions are often not particularly high quality. XKCD was complaining about the quality of YouTube comments in comic 202, published before August 14, 2007:
https://www.xkcd.com/202/

The situation's not improved markedly.

Many online sites have disabled or done away with comments entirely. NPR did so in 2016:
https://www.npr.org/sections/ombudsman/2016/08/17/489516952/npr-website-to-get-rid-of-comments

They cite an article listing another seven major publications doing similarly:
http://www.niemanlab.org/2015/09/what-happened-after-7-news-sites-got-rid-of-reader-comments/

The New York Times was employing a team of 14 to review 11,000 comments per day, or 785 per person. This appears to be close to the peak of sustained content-response I've found documented.

https://www.nytimes.com/interactive/2016/09/20/insider/approve-or-reject-moderation-quiz.html

Consummate digital consumers Walt Mossberg (WSJ, ReCode), and Stephen Wolfram each have a daily volume of roughly 100 - 300 messages per day. Wolfram actually tracks this, and his outbound mail, whilst peaking near 250, has a mean generally less than 50/day on a monthly average (peaking at 100/day over several months), and a mode of roughly 25.

http://blog.stephenwolfram.com/2012/03/the-personal-analytics-of-my-life/

People can't take much more than this.

I've looked at the number of news stories produced in a day, many accounts of top or most significant news stories, items in broadcast news programmes, and more:

A contemporary news site such as the NY Times, WashPo, or WSJ, publishes about 150 - 500 original items per day (most on weekends).

The news wires, AP, Reuters, and AFP, publish from 1,000 - 5,000 items per day.

The US Presidential Daily Brief typically includes 10 items, over 1-5 pages. This is the most important news, for the most powerful person in the world.

A top-of-the-hour NPR news summary typically includes about 5-8 items, and an hour of Morning Edition or All Things Considered, about ten segments of 4-5 minutes each.

Time, attention, is the ultimate rivalrous resource, and is critical to all media. (And yes, I've been taking up far too much of yours.)



Text Types

The original text was the word, spoken. Messages might be interactive, if among a small number of people, say, 2-5, rarely more. Or monologues or speeches, to larger groups. Interactivity is a function of audience size and fungibility.

The first written texts were small in content, though often durable: cave walls, stones, clay tablets. Flexible media -- papyrus, bamboo, or vellum scrolls, repeated the sequential access of a spoken speech in print.

The codex offered random access to the work, once completed, but production itself was slow and expensive. Pre-Gutenberg books were highly valuable treasures, often secured by chains in libraries, occupying a scribe for a third to half a year, and claiming the hides of a large herd of cattle.

Printing created the uniform work, identical in all reproductions. It reduced costs to the point that it was cheaper to bring the text to the language of the reader, via translation, than the reader to the language of the text, through instruction. The vernacular literature was born.

Word separation, punctuation, paragraphs, page numbering, notes, title pages, indices, and bibliographies, all had to be invented. Several of these were conveniences for production (not printing or binding books in the wrong sequence) which proved useful for readers. They represent over 1,000 years of technological development in textual representation conventions.

Carl Linnaeus faced a problem as he attempted to categorise the world's life-forms: capturing information in a format which allowed for flexible re-presentation and re-organisation. The result was the index card, which rapidly became the favoured tool of encyclopaedists, lexicographers, librarians, census-takers, spy chiefs, and conspiracy theorists. John McPhee and Vladimir Nabokov are inveterate users of same.

Changing information created the requirement for changable texts. The loose-leaf booklet appeared in the late 1890s, serving a business need for updating their burgeoning, and changing, procedures and manuals books.

https://books.google.com/books?id=LGQbAQAAMAAJ&pg=PA1604&dq=%22loose-leaf%22&hl=en&sa=X&ved=0ahUKEwiIw4264PjdAhUDYKwKHa0VDOgQ6AEIKDAA#v=onepage&q=%22loose-leaf%22&f=false

Sequentially published "yearbooks" and quarterly or monthly supplicants appeared at about the same time, addressing the need for standard, but current documentation in a changing world.

The concept of structured data, a database, dates to the 1970s. This combines earlier notions of tabular data with the ability to dynamically insert, delete, or update records.

Hypertext created the notion of interlinked documents, dating to Ted Nelson in 1963.

The versioned document, first as whole incremented version as with DEC's versioned files on VMS, later as patch-based diffs, as with Larry Wall, was another milestone. It was no longer necessary to supplement a document, it could be updated in place. We are at 1985.

The Wiki combines elements of hypertext, versioned documents, databases (for metadata and change management), as well as structured, formatted, and multimedia interactive documents to create yet another novel document type. Traced to Ward Cunningham, 1995.

One might include the thread, or interactive discussion (named after Usenet or Unix email's threaded presentation of discussion trees' child-parent relationships) as another intrinsic document form, though I prefer to think of that as a document relationship. (In a future work I hope to come up with a comprehensive classification of ontological category boundaries.)

The history of texts: Speech, tablet, scroll, codex, mass-published book, card database, loose-leaf, supplement, database, hypertext, versioned file, wiki, stream, thread.




A rich, productive, useful media system should support multiple media types

There is a very small, and perhaps quite dull, point buried in all of this.

The media systems we've been presented with in most major social media systems appear rich but are in fact poor. They offer a very thin selection of possible offerings. They lean to the engrossing categories, those which capture and monopolise attention: the stream, images, video, audio. Games.

Again, not that these are bad, but they are a subset of the possible.

Very few social media systems offer anything remotely resembling a wiki functionality. Reddit is a notable exception, though its wiki is quite weak (no search, no dynamic renaming, limited versioning). There are Wiki sites, and they are quite popular, even beyond Wikipedia -- Wikia and numerous fan- and other user-oriented wikis exist.

A surprising number of social media systems have limited, crippled, internal-only, or no search.

Google Plus, produced by a company synonymous with search, had no search when it was released. I believe that was added in early 2012, roughly 6-9 months later, but even then, the search was limited. Until the past year, searching by metadata -- poster, commenter, date, media types -- was not supported. Reddit has full post search, but no comment search. Ello has an almost perfectly useless search. Facebook doesn't allow external indexing of much of its data. Mastodon has no native search (though content has been well indexed externally).

I live and die by search, as I'm constantly re-referencing my own writing, and searching for related discussions or information. The lack of search is crippling. The difference between a search-crippled and a search-enabled site is that of passive consumer and of active participant.

Funny, but when you put it like that, the strategy sounds deliberate.

(In fairness: search is computationally expensive, finicky, surprisingly complex in edge cases, and an all around pain in the ass, much like myself. But it remains exceptionally useful.)

A reasonably complete media system, or set of interconnecting systems, should support multiple formats.

I would like long, rich, complext text capabilities. Sufficient for publishing scientific articles, housewares catalogues, photo albums, novels, and short but very sweet love letters.

It should allow (and provide user controls for) images, audio, video, data presentation (including data sorts, summaries, filters, pivots, and graphs), and in limited cases, programmatic content. (Security and other concerns emerge here, I'm well aware.)

The simple should be simple, the difficult should be possible. Novice users are supported, expert users are empowered.

Publishing scope, reading interest, response prerequisites (look up "Scholastic Instruction" within Wikipedia's "Scholasticism" article for an item I've spared you here), should be within the control of the author or audience. Elements which might be distracting (or worse) to readers can be preemptively disabled (or worse) by readers.

Short advertisements of longer material are possible. Something of a cross between Twitter (or Mastodon) and RSS.

(Twitter is an interactive RSS feed. Hrm....)

Discussion can occur, in large groups or small. And information that might be captured from a discussion can be, and restructured, refactored, rewritten, in a wiki-like environment. Knowledge can grow.

(The empirical testing of the presumed knowledge is out of scope of the system, though not of the quest for truth itself.)



Conversation scales poorly

There are 7 billions and 400 millions of souls on this planet. Each of them has a story. If you were to listen to all of them, over an 85 year lifespan, you could afford each just over one third of a second.

(And yes, I know, I've claimed far more than my share. I apologise.)

The residents of what is today a modest-sized town, or a mid-sized university, of 10,000 could occupy you for 3 days each, over the same span.

A social network with "only" 10,000 members is considered tiny by today's standards, though if it were the right 10,000 members -- a sampling of selective-admissions universities, or global politicians or business or artistic leaders -- would be exceedingly compelling.

Early social networks -- the telephone system, as an example -- provided far more value in who they excluded than who was included. If the phone rang, you knew it was important, for this reason alone, let alone the cost of the connection dissuading frivolous contacts.

When was the last time you were excited to answer the phone?

When it was The Harvard Facebook, and even as it grew through other Ivy, selective, and general collegiate populations, there was a cachet, something to be gained, from that association. I've watched a new network, Mastodon, grow over the past year and a half, and the inevitable complaints that it Ain't What it Used to Be (usually meaning: the normies have invaded).

Facebook was once Literally Harvard. It is anything but that now.

Early-phase social networks are exceedingly sensitive to founding cohorts. Facebook succeeded like gangbusters. It stole a script for Usenet, which lacked a monetisation strategy (well, other than MAKE MONEY FAST!!!!). Slashdot, Reddit, and Google+ largely leveraged off of a founding tech cohort (I've argued that Google's co-population of advertising and SEO types created an early disutility for the network). Some networks start bad and get worse: Voat and Gab. Some turn toxic: Imzy. Many simply fail to sustain traction: Ello.

Network effects matter, but not as much as the cohort.

Site dynamics also matter, hugely. Mastodon and Tildes seem preternaturally aware of this, and have been making some very good decisions. This isn't a stage of growth at which they'll guarantee success, but intelligent design (the real variety) can avoid snatching defeat from the jaws of victory.

The key is to have enough people present that a productive discussion on topics of interest to you, based on a relatively small number of participants -- as few as 3-4, rarely more than 20-30 -- can occur. With lurkers, occasional participants, and the like the total size of the group may extend beyond this -- by the 90/9/1 rule, individual groups of perhaps 300-500. But much above this with active participation the quality trend seems to run markedly downward, and quite possibly at much smaller levels.

Several social networks have attempted to specifically engineer for Dunbar's Number, notably. This ... doesn't seem to work either, though then again, most social networks simply don't. I suspect that optimising for small groups too early may itself be a failure path (no reference intended).



Utopias do exist

If the notion of a utopia is an intentional community, then it turns out that these do exist, and are amongst the most durable of all human institutions.

Academic institutions.

Many of these include a considerable surrounding community as well. The reasons for this institutional longevity (and various factors surrounding schism) might prove interesting. There are 40 European universities dating to before 1500, three before 1200. The very oldest remain among the highest ranked in the world.

Maybe there's something in the water?



Stop and smell the primroses along the way

As advertised (and as I feared) this has been a long and meandering sojourn through a number of topics for which I hope my barrister can make a credible case for relatedness. I don't have a surefire formula to obtain my menu of desires listed above. Some of the landmarks I've pointed out may be useful. Thank you for your indulgence.




https://www.xkcd.com/202/