On the Design of a Poetry.Store Territory : Turbocharging Administration

This is the first post in a series.

Problems to Solve

With careful design, I believe that Territories with curated posts, together with T2's capability of tracking how much of an entry is actually read, could resolve various problems in the effectiveness of equivalent entities existing on other social media platforms (i.e. Groups in Facebook, Channels in Discord, Topics in Reddit, and so on). These problems include the following :

• The "This Topic/Group/Channel Is Unmanageable" Problem
• The "This Post Doesn't Belong Here" Problem
• The "It Involves Extra Work for me to Support the Poster" Problem
• The "I Don't Trust this Review" Problem
• The "How Do I Get Reviews Chicken and Egg" Problem

I'll define the first two of these problems in this post, in the context of a potential Territory with the declared purpose of making it easy to both find poetry and support poets (i.e., "Poetry.Store"), and then propose a solution.

This Topic is Unmanageable

Let's assume that you define yourself as one or more of the following :

• a Reader of Poetry (ROP) with a goal of Reading Poetry (RP).
• a Writer of Poetry (WOP) with a goal of Writing Poetry (WP)
• a Buyer of Poetry (BOP) with a goal of Buying Poetry (BP)
• a Seller of Poetry (SOP) with a goal of Selling Poetry (SP)

There are a large number of Facebook groups which are repositories of poetry that you can explore for the purposes of fulfilling your goal. These groups can be divided into two camps : those without effective admins, and those with effective admins.

Let's consider groups without effective administrators. I will pick as an example the group "Poems by Poets; Poets Buy Poems." The title of the group implies that the main purpose of the group is to facilitate the reading and purchase of poetry in some way; and if I read one of the pinned posts I see that this is indeed the case. It states that the criteria for posting to the group is to include both a poem and a link to a book where the poem appears, so that readers looking to buy their next poetry book have a place where they can easily browse and make their choices, without being limited to a particular bookstore. Thus, instead of going to Amazon, finding the Poetry section, "peeking" inside a book in the hopes that a complete poem happens to fit within the confines of the amount allotted by the bookstore for peeking, then making their choice without checking any smaller bookstores in a similar fashion, they can instead browse in the Facebook group and make their choice without depending on either that particular bookstore's inventory or on which section of the book was algorithmically selected for 'peeking' by that bookstore, since the section provided as a preview is made by the author themselves when they created their post in the group and needn't be whatever happens to be contained in the first X percentage of the book.

If I join and peruse that group as a reader (ROP), I rather quickly begin to find posts which are not poems, which is the "This Post Doesn't Belong Here" problem listed above. This is an indication to me that the admins are not "effective," given the non-trivial size of the portion of the posts in the group which do not support the stated goal of the group; and this ineffectiveness makes it harder for me to accomplish my own goal of reading poems (RP). The same would be true if my goal had been Writing Poetry (WP) or Selling Poetry (SP) or Buying Poetry (BP) rather than just reading it. In particular, the pinned post I reference above asks the posters to "please include a link to your book," and the majority of posts which are indeed poems do not include any such link, which is a further indication of the ineffectiveness of the admins, especially for those wanting to Buy Poetry (BP).

Why might these admins be ineffective? I can only hazard guesses without contacting the admins directly; but by viewing the members of the group I can see that there are "apparently" only two admins among over 1,500 members. (I say "apparently" because the two admins are actually one and the same person, whom I happen to know; so I ask readers of this post not to engage in a burst of collectively flaming him for being 'ineffective'). This disparity in numbers leads me to believe that the admins simply don't have the time to properly administer the group - i.e., they have encountered the "This Topic is Unmanageable" problem listed above.

For groups with effective administrators, finding a group to satisfy your goal is less of an issue - take for example the group "Facebook Poetry Society." If I join and peruse that group as a reader, I note that pretty much all entries are poems, allowing me to reach my goal (RP), despite the fact that at the time of writing there were over 85,000 members of the group.

How are the vast amount of submissions from such a group handled? Examining the group administrators reveals that there are over 200 of them. Does Facebook have a way for these admins to easily coordinate with one another? Not that I can find, though it does provide a helpful article on ad-hoc strategies to use.

The trick, then, to accomplishing your goal when joining a group, is to find a group which purports to help you fulfill that goal and which then happens to have both enough administrators to handle the amount of traffic and the ad-hoc structure created by those administrators to coordinate with one another in an effective way such that the criteria for posting are followed by posters and verified by administrators in a timely manner. On Facebook, finding such a group can be challenging - I find that I have to pick groups based on their title, join them, then examine the entries looking for effectiveness, according to what my goal is.

One could thus rephrase the problem of "This Post Doesn't Belong Here" - as encountered by anyone in the set {ROP, WOP, BOP, SOP} - to "Finding a Topic which Satisfies my Goals Is Too Time-Consuming," with its root cause being the separate problem - encountered by the set of group administrators - of "This Topic is Unmanageable."

One Solution

For a Topic to be manageable, one could strive to automate as much as possible the assignment of administrators such that the number of available admins at any given time is enough to process (in this case, to read and filter out according to some criteria) the number of posted articles within a certain time frame. The ability to measure the amount of time spent by each administrator and then either warn that administrator that they are falling below a desired threshold, or to take remedial action by automatically 'activating' additional admins, thus picking up the slack, could be a part of the solution.

For example, imagine X number of people wanting to be administrators of a territory to which Y posts per hour are submitted. Each of the X possible administrators has a different amount of time that they estimate they can spend administering. In a similar way, each of the Y posts has a different amount of time needed for processing, on average.

Each X could 'apply' to be an admin. If X were a user who had already read enough articles on T2 such that an 'average reading speed' is already established for X, then that user would not need to perform any additional action to apply; if instead that user had not yet read enough articles to establish a trusted baseline then the user would be informed that they need to read at least Z more articles before doing so. (this assumes that the admin needs to read an entire post to determine that it meets all acceptance criteria).

Meanwhile, every user - regardless of whether they want to be an admin - who reads any article helps to establish that article's average amount of time needed for reading.

So let's say that the purpose of the proposed "Poetry.Store" territory is the same as that of the Facebook Group "Poems by Poets; Poets Buy Poems" - to enable poems to be easily read, written, bought, and/or sold while avoiding constraints inherently imposed by relying on browsing a particular bookstore's website - with the addition of the ability to get rewarded for the act of browsing itself - which is inherently granted by hosting the "Group" (i.e. Territory) in T2. Annie wants to be an admin of this group. She says she can spend two hours a day reading submissions (presumably, poetry) to determine if the submission meets the criteria for acceptance. What would she need to do as an admin to perform her job ?

• Read the submission and determine if it is a poem
• Determine if a link to the book which contains the poem is included
• Determine if the link actually goes to a "real" bookstore (say, Amazon) or NFT Marketplace (say, Cryptoversal Books)
• Determine if the link is active

What might go wrong in this situation?

1. She ends up spending less than two hours a day, on average, reading submissions, AND the number of posts Y awaiting acceptance ends up increasing over time as a result.
2. She ends up spending a full two hours a day, on average, reading submissions, but the number of posts Y is accelerating over time such that she cannot keep up

If the amount of time spent by each administrator X can be tracked over time, where X is a member of the set of all admins, and the rate Y at which posts are processed can also be tracked, then the detection of an emerging problem for any member of X as described above ought to be detectable given some algorithm A. A could then control the assignment of posts to various X members such that the load is 'rebalanced' to fit the reality of what they can handle given the current load. Algorithm A would thus essentially be equivalent to a load balancing algorithm employed by a network of computers to ensure timely responses to queries.

The important point here is that the potential admins are not burdened with coming up with their own ad-hoc processes for handling who does what such that things don't become unmanageable. This is a problem that, to my knowledge, platforms such as Facebook, Reddit, and so on have not solved.

What could algorithm A consist of? I will leave that up to the developers of T2 to ponder.

Required Infrastructure

Note that the ability to implement algorithm A implies that the admins of the Territory should be able to specify their criteria in an enforceable manner in a way which is independent from other Territories. In other words, the criteria for posting into various Territories will likely differ depending on the type of posts expected in those Territories - so, for example, the Poetry.Store Territory would require that a poster supply a poem and a set of locations where some item containing that poem can be purchased, while the supplier of say a News.Balanced Territory might require that a poster supply a short factual summary of a current event along with links to three archived articles (via the Wayback Machine for example, or a censorship-resistant blockchain) which cover the same current event in greater depth from news organizations known for having three different viewpoints (say, from a "Liberal" news organization, a "Centrist" one, and a "Conservative" one), with the goal of the Territory being to provide a truly 'balanced' perspective and to provide a sort of proof that particular readers have 'open minds' with respect to different opinions - as evidenced, presumably, by their willingness to actually read different viewpoints.

This means that an administrator (or perhaps an administrator of administrators) should be able to create a template made up of components which enforce criteria specific to a Territory. So for Poetry.Store, the admin could create a template with a textbox for the content of the poem, followed by a drop-down which contains N links, each of which is restricted to links to known actual bookstores - which would be vetted by the admins via something like pre-supplied regular expressions for amazon URLs, apple URLs, etc. Posters would need to successfully fill in these components in order to submit a post; the regular expressions in the post would reject attempts from malicious posters to direct readers to look-alike or fake stores, while the admins would be left to verify that the book pointed to actually corresponds to a book authored by the poster and which actually contains the poem, after verifying that the text supplied is, in fact, a poem.

But how can an admin verify that a particular book actually contains a particular poem? One answer to this is supplied by an analysis of a potential solution to the "I Don't Trust this Review" problem mentioned above, which will be the topic of a future article in this series.