Comments on: Universal academic directory http://www.hyperorg.com/blogger/2008/09/30/universal-academic-directory/ Let's just see what happens Sun, 16 Jun 2013 09:24:11 +0000 hourly 1 http://wordpress.org/?v=3.5.1 By: Charles Eisenberg http://www.hyperorg.com/blogger/2008/09/30/universal-academic-directory/comment-page-1/#comment-38815 Charles Eisenberg Thu, 16 Oct 2008 00:06:22 +0000 http://www.hyperorg.com/blogger/2008/09/30/universal-academic-directory/#comment-38815 I like to file papers and reports in manila folders to stay organized at work. I want to be able to find things quickly. Sometimes it is difficult to know how to label a new folder; at other times, it is difficult to know in which of my folders a particular file best belongs.
In the first case, I am often puzzled defining the most useful category—one that will best jog my memory, and remind me of the folder’s contents. I want to avoid having to look through each folder, or even a large subset of my folders, to find the particular target (paper, report, memo, policy, etc). For example, a folder labeled “memos” won’t be very helpful. I have to search through the entire folder to find the target memo. Also, I have five areas that I supervise, within my department of Behavioral Health Services. I often want to separate my folders along those lines, keeping certain folders for Inpatient, others for Outpatient, still others for the Substance Abuse clinic, etc.
In the second case, it is sometimes difficult to know where to file something, as it has important relevance to several concerns of mine, each with their own folder. So then I ask myself the second best question, “Where am I most likely to find this target, when I go to look for it?” I say this is second best, as ideally I would like my folders to have some kind of coherent, and consistent organizational unity, independent of my practical needs. As if the unity of my mind and my work had found some physical expression and independent existence in my desk.
Even when my filing principle ‘devolves’ to “Where am I most likely to find this target?” the choice is not always easy. Reflection sometimes reveals that I can easily first think of the target under one rubric, then under another, and vice versa. For example, Things relating to the Outpatient license could be put in the folder label “Outpatient license.” But what about my notes on supervision of the Outpatient director, where one of the topics was about the Outpatient license? Should I place supervision notes in the folder labeled “Outpatient license?” That does seem right both as regards organization per se and as regarding recall. On the other hand, if I place it in the folder labeled “Janet Supervision notes,” then I really lose the mental tags altogether. There’s no way I will be able to recall where I put the important information, discussion, work assignment regarding the Outpatient license. So I might, as a ‘safety’ measure to insure recall of the target, make a copy and place the target in both folders. Some such organizational decision plagues me from time to time.
With the computer and my Microsoft Word documents, the same as above applies. There I have the additional tool/ headache of availing myself to subfolders, for additional specification. The problem with subfolders is two-fold. First, too many subfolders leads to a different kind of bulky plurality, within which one must search. The real index to the folders and subfolders lies within my head. To write it out would be tantamount to listing all of my files (targets), and then “What’s the point of having folders?” Nothing is saved in efficiency and organization. The folders and subfolders work for me (I wonder how it works for others?) as mnemonics, memory pointers, supplementing my memory recall with memory recognition triggers. The folder (or sub, sub, sub-folder) jars my memory, in a variety of ways, which I generally don’t even notice, because they occur so rapidly, and I’m not in the habit of attending to them. “What was I thinking when I filed the contents of that folder? Roughly when did I create that folder? Was the target created before or after I created this folder? Which of these folders is the most likely way I would have thought of the contents of the search target?” And so on. Of course, sometimes none of that works and I have to resort to opening up large ‘regions’ of folders and subfolders to take a look down the list of file titles. Occasional I cannot find something that I know has been saved. Months later, when I stumble upon the target, often I have been shocked to see where I put it, and how bizarrely—though understandably—I conceptually pigeon-holed the target. It made perfect sense to me at the time! Now I have gotten into the habit of “multiple location filing” for something that I think may be a bit of a stretch for my future self to recall. But that’s a technique I discussed in my use of physical papers, that’s to say, material organization of my targets.
The second aspect of problem with my Microsoft Word subfolders is that they, like the targets (files) themselves, often should sit in two or more distinct folders. I have yet to duplicate a subfolder and allow it multiple locations (copies) in my Microsoft Word “OPEN” dialogue box. And yet, I do allow myself the liberty to store targets in multiple folder and sub-folder locations. That seems inconsistent to me now.
What I have been pointing to is what might at first blush be considered under the rubric of “The Organization Features of Material Things.” I say, “At first blush.” But is the organization of folders, sub-folders and files in my Microsoft Word really material? Do the problems that attend its organizational limits have the same roots, grounds or causes as those that attend my physical papers and manila folders sitting in my physical desk? The answer is Yes and No!
Microsoft Word’s logical structure, on the user level, meaning at the “OPEN” dialogue box level does mimic the “spacialization” of organization that we see in play with real material organization systems like my papers and manila folders (By the way, there are cardboard accordion folders which one could buy and use as containing sub-folders, akin to what I do with my computer files.) But we know that the little file ICONS that I see in the “OPEN” dialogue box are not tiny real, physical manila folders, which take up space, and which are subject to the limits and nature of space, e.g. two things cannot be in the same place at the same time, the change of place requires transversing space, etc. But in another very, very real and important way, Microsoft Word’s organization of files, folders and sub, sub, sub-folders is spacial and is subject to the same limitations.
The file I am typing now is ultimately stored on my hard drive, once I save it and close it. But what does that really, really mean? It means, in general terms, that somewhere on magnetic surface, something is done by the writing head of my hard drive, to alter the material, in its magnetic aspect. That is, the material is subject to magnetic properties, and the computer alters the “local landscape” by use of the writing head of the hard drive, so that an enduring “design” is laid down on material. If we magnified the material 100,000x and made ourselves equally small, we would be able to walk around the molecules, so to say, and see arrangements that weren’t there yesterday, and which are perfectly describable. These arrangements were created by the writing head of my hard drive, and roughly speaking are in a very special code or language that the computer can manipulate.
I took the time for that excursion into my imagination in order to drive home the point that computers are physical and their files are physically stored somewhere—even if that somewhere is really teeny-tiny!
(What I am about to say is probably wrong as to the details, but is definitely right as to the general principles. And it is the general principles that are interesting, for the purposes of this paper. )
My computer takes this file I am writing, and stores it as magnetic “code” in a variety of places on the hard drive. The hard drive is organized in something like sectors. These sectors have addresses just like my house and your house have addresses. The computer has defined these sectors. I don’t need to know these definitions—thank god! I just hit SAVE and the computer places aspects of this file in one address, and other aspects in another address. When it does so, it also writes down something like reminders, where it put everything, so that later if I ask to see this file I am typing now, it can know where to go—what addresses in which sectors—to put the file together for the user level, which I am now attending to on the surface of my monitor. The computer may store information about the fonts in one place, information about the page layout in another, information about the font color in another, one word over here, another on a completely different sector, the end of this paragraph ahead of the beginning of this paragraph, and so on. So long as it has a “recipe,” or what I would call an algorithm—that is a description of a procedure to follow to re-assemble what it stored—then the file is endlessly recoverable and storable. The algorithm might go something like—look up the table of sectors, look up the addresses within the sectors, pull the list of items with the list of their addresses in the sectors, grab this item, then that on and so on. The basic point is that the file in a computer is nothing like the manila folder with files sitting in my desk.
Computers assemble stored files when we need them. And even when we are using them, reading and typing right now, the information on my monitor is being re-freshed at a bewildering rate so that I don’t see any flickering images, as you might with early silent films running at slow frame rates. Furthermore, the file is broken up into distinct and non-contiguous places in the Random Access Memory (RAM) registers. The font styles I have used in this document may be located—physically located—next to the information on page formatting, or next to the code for this letter “a.” The computer doesn’t need contiguity. In a real sense, it is assembling and re-assembling this file that seems pretty constant on my monitor. In comparison, the physical Manila folder (and its files) seems to have a greater degree of permanence, both when I am working with a sheet of paper, and when the sheet of paper is placed in a specific manila folder and closed in its draw (Alphabetical order, of course!)
(On a fine-grained level, the manila folder is also being assembled and re-assembled, again and again but that is for another paper on Quantum Physics!)
I like to think of the computer as something like a factory, where the files are being called up and assembled, as needed. This is clearly different than the physical files, with their ink, and cellulose pulp, i.e. paper, lying within physical manila folders, which lie side by side, in alphabetical order, within my desk. This difference has led to “tagging.” Tags point to web pages but are not web pages. The web pages are assembled, as needed—as called up. The web pages are not stored in our computers. But first a word on tagging .
Tagging means putting a label to something, in this case, a file. Tagging has become a popular notion in our culture through the use of the internet, although the fact of its popularity, and its birth on the internet is not terribly important to what it is. It is a means by which you can locate an item, in a diversity of contexts. For example, you can stumble upon a web page on neuro-anatomy and tag it as “Neurocognitive studies” as well as “Computer Science” “web pages with green backgrounds” “stuff I found while drinking hot cocoa” and “brain.” There is no end to the list of tags you can create. And that one page can simultaneously have all those tags associated with it.
This is a bit different than anything we’ve seen before, regarding problems and solutions to keeping organized. With manila folders, you can’t put the same file in two different folders at the same time—unless you make redundant copies. But you can’t put the same paper in two folders at the same time, nor can you have two folders inside of each other at the same time, and so on. With tagging, one web page is located in multiple contexts simultaneously. So if I forget where I put that web page on Neuroanatomy, I can guess it is in “brain,” “neuro-cognitive science,” “computer science,” etc. and I will find it there. And when I store it, if I want to be sure that I see it when I’m on a page on the latest anti-psychotic medications, I can add that tag “Psychopharmacology.” But as we saw with ordinary filing, there is a point where proliferating tags is no longer helpful. If I have 10,000 tags on all 5 web pages in my total collection, then it doesn’t matter which tag I hit—I will see the list of my total collection of 5 web pages in all 10,000 tags. However, if 1 of those 5 tags is only tagged in 1 of the 10,000 tags, then I stand a good chance of never finding that web page. These are some of the ridiculous outside limits of tagging. But I want to slowly come in on the practical limits (inability to find web pages) to tagging web pages. Maybe even develop a formula. To take another outside limit, suppose I have 10,000 web pages and 1 tag. Well, that amounts to absolutely no efficiency and no increased organization. How’s 2 tags? Better but still a bit unwieldy. How’s 10 tags? How’s 100 tags, with 100 web pages in each tag, organized along topics. Better but still unwieldy. Maybe we are being unfair in starting with 10,000 web pages (But there is an important point in this observation—one which the ‘pack-rat’ knows.) How are 100 web pages, as our base? We all love 100. And we’ll have a few tags (2, 3, 4…..n) up until 100 tags. Different problems beset the owner of this collection of web pages, as the number of tags increases. As we said, 1 tag for all 100 web pages adds nothing. 2 is a bit better. 10 is even better—just from a quantitative point of view. However, we should also examine the content of the tag. For example, “web pages I found while drinking hot cocoa” may or may not be useful. If I am referring to a particular day of drinking hot cocoa, then it might quite clearly jog my memory of what web pages I’ll find through that tag. But it is probably a useless tag. So “what” a tag is, and how many tags, to how many web pages—that’s what we have to consider.

1 tag—dumb. 100 tags—dumb. 10,000 web pages—dumb. 1 web page and 100 tags—dumb. Let’s see if we can generalize. With a reasonable amount of web pages, say 100, then at some point—which varies person to person—the tags become too numerous to efficiently and accurately identify a given target web page. Let’s use the convention of tags/web pages.
In a uniform distribution of tags:
100/100 = 0 efficiency.
10/100 = 10
1/100 = 0
A graph of tags as against efficiency would produce a bell curve of some sort. But that shape would change with the number of web pages being considered, and the number of tags that tagged a given page.
Again assume uniform distribution of tags:
5/10 = 2
And yet, I am fairly certain that I could find things in a set of 10 web pages with bad tags much quicker than in a set of 100 web pages with 10 excellent tags.
Also, assuming 100 web pages, without uniform distribution of tags and the efficiency can go way down quickly. Imagine a set of 10/100, but 8 of the tags cover 20 items and the 2 other tags cover 80 different items. That really amounts to a reduction in the effective power of the tags. You really have something like 2.xx tags/100, hence the weakness.
I think what is decisive is one’s ability to frame pointers (tags) that will elicit a rough and fairly accurate guess of the pages that fall into that tag. If you label tags numbers 1, 2, 3, 4….10, then you really have to memorize what attaches to each tag. So we have the dimension of specific versus general, and what falls somewhere in between (“Stuff I found while drinking hot cocoa,” “ tag #4,” “neuroscience,” etc.)
The pointers don’t even have to elicit a clear and distinct memory of what attaches to a tag. The pointer “Neuroscience” really borrows on your memory of what that tag is and isn’t, and you set up a working hypothesis of whether the target page is attached to that tag or not. Even narrowing your search to 4 tags is a big time-saver. You may still glance through 50 web page titles but that’s better than going through 100. But what about just listing the web pages alphabetically. The titles don’t often usefully lend themselves to that.
Tagging is also used to stumble upon, so that looking at a web page on Neuroanatomy, I see it is tagged with “Doug Hofstadter.” And that may lead on to connections I would have otherwise missed out upon.
SUMMARY
1. How can you measure a tag’s power to elicit good guesses?
2. Plug that into the number of tags/web pages
3. Describe the distribution, and assign a “distribution power quotient” to particular tags in that set of tags/web pages. (For example, in a group of 10 tags, each of 2 tags covering 100 of the pages would be weaker than each of the other 8 tags covering 100 of the same pages)
4. Factor in the number of web pages in the collection being considered
These are the dimensions that I have identified in describing the cognitive power of tags. On another occasion I will work on a formula for quantifying the strength of tags to serve as pointers to target pages.
I want to return to the main thread of this paper—determinate versus fluid organizational formats. As I said earlier, the computer is something like a factory, where the files are being called up and assembled, as needed. This is clearly different than the physical files, with their ink, and cellulose pulp, i.e. paper, lying within physical manila folders, which lie side by side, in alphabetical order, within my desk. Tagging allows me to “file” (in quotes) pages. But what is really occurring is that now we are collecting algorithms for assembling pages.
In the old days of actual cellulose papers and manila files, etc. space became a limiting factor, in several senses. First, it quickly accumulates, as any hospital medical records department can tell you. Secondly, your filing must obey those nasty rules of 3 dimensional life—no simultaneously diverse location, and two things can’t occupy the same place at the same time, be in each other at the same time, be organized distinctly by two different people at the same time, etc.
Much of the fever regarding the internet has to do with the content of what you find there. But some of the fever regarding the internet has to do with the maddening desire to be everywhere at all times, with all possible content, real and imaginable. And to have none of that determine what your next moment can be. No limits! Wow! File sharing, social networking, and tagging have made possible the access to and accumulation of unimaginable amounts of content, in a variety of media. Organization and effective access to one’s collection of content is becoming a bigger and bigger topic. Books have been written about this topic, which either avoid the question of the limits to this explosion of access, or happily suggest that there is no end in sight. But avid users of the internet—users of file sharing, bookmarking, social networking, and tagging—are quite aware of the limitations and paradoxes that occur when virtually endless content is there for the taking.
The computer appears to free us from the limits of all taxonomies framed under the constraints (tacit or explicit) of space—I will call them “spacial taxonomies.” The files in my desk are one example. So are the Dewey Decimal System, and even the idea of alphabetizing. All this spacializes logical relations, and has led to a general confusion of the limits of space with the constraints of taxonomical systems. The computer age has brought these unexamined assumptions to the foreground. And the pundits declare the war on taxonomical constraints is over! And we have won!
However, I have this nagging feeling that the limits of space will swiftly be upon us. And war will break out anew. This time it is a war between the actual and the possible.
Computers have limits to storage, limits to working memory (RAM) and limits to the speed of the hard drive—which whizzes and zips around just like any other mechanical object. They also have limits to their synchronous internal clock, which ticks away at a set speed. I am sure there are other limits as well. But what does this all amount to.
Our collections no longer are constrained by spacial taxonomies because computers do not store physical papers, steel bins, aisles, wooden shelves, corrugated boxes, microfiche, etc. of content. Computers are little factories, where we order up what we need to meet the demands of the present project, e.g. writing this essay. We don’t store papers, files and folders in this new world, as we once did in the old world. We store algorithms for building content (files, web pages, compilers, programs). And there is the catch:
1. The storage is still physical.
2. And the immensely expanded capacities for “weaving” together what the user calls up has led to a proliferation of content that is starting to exceed what one user, or even a group of users can usefully organize, track, attend to, and integrate—on a human scale.
My lap top—the one upon which I am word processing (we should always remember that we are using a word processing program and not a typewriter)—has 220 GB of hard drive memory with 3 GB of RAM. When it gets full, that’s that. Either I delete something or I am done collecting. 220 GB is pretty damn big. But not limitless. Actually, you can fill it pretty quickly with some full length DVD movies, and a few thousand songs, plus a couple of virtual reality programs like “Second Life.” This is a horizonal limitation, in that, like the horizon, which constantly moves as you move, these limits can never be ultimately overcome. (Never say never.) Maybe I can buy a free standing hard drive of say 1 Terabyte. And that will help me for some time. Maybe forever, given my purposes. But the real problem for us humans—us human users—is not the limits of computer memory and speed. Computers can easily contain content and “pointers to content” which far exceed what a human being can make use of. This is what I want to dwell upon, because the issue is upon us. What is meant by “useful?”
Let me start by making an analogy. When I go food shopping—any shopping really—I bring back a limited selection from the supermarket. I don’t bring back the supermarket. I don’t bring back the even larger supermarket distribution warehouse. And I don’t bring back the world, in the entirety of its goods. Even if I could, where would I put the world? It already takes up the world! Now let’s say, the inhabitants of this good Earth agree to give me the world. I sign the deed and it is conceptually established that the Earth is mine. If I want to know everything that’s in it, that’s gonna be a big task. Where is Joanna who lives in a small town outside of Amsterdam? What is she wearing today? What is the weather in Figueres, Spain? What store in Figueres has the best price on umbrellas? Abstractly I have the world—it’s mine. I have the deed! But what is likely to occur is that I will explore, act and utilize locally and limited, even though I have the world as possibilities. So if I re-design the streets of Minsk, I turn my back for a second, and someone is changing what I did! I need an army of overseers, set up in a hierarchy, so I can organize my world, and keep it the way I like, know its particular local features, and the possibilities that exist there. Maybe I need some guide books to each of the regions. Do I organize it by Continent? Too big a book, with a very big table of contents and index. How about by country? I’ll want a workable book, not too detailed, so that a country guide book might be manageable. But then I miss out on knowing the particular local flavor and possibilities for action, thought, etc. Maybe I should create a library of town guide books, and an army of town guides. (But then they have the knowledge in their heads. I want it all in mine. And how do I even know what I should be asking about, when I talk to the local town guide). A library of 120,000 town guide books, each with their detailed accounts. And 120,000 town guides. That is still inadequate. You get the point.
One more analogy. If you have a map of the United States, which is the same size as the actual United States, what have you gained in comprehension?
So I reassert that the immensely expanded capacities for “weaving” together what the computer user calls up has led to a proliferation of content that is starting to exceed what one user, or even a group of users can usefully organize, track, attend to, and integrate—on a human scale. It is for this reason that we will find ourselves—human users—de facto limiting our collections, and thereby being in the dark about what it is we don’t know, don’t have, didn’t utilize. We won’t know that we don’t know. You find this with the specialization of knowledge, in every domain of human thought. Who knows all of the knowledge base contained in Neuro-cognitive science—especially in light of the exponential advances in this particular field in the last 5 years? What you gain in sweeping survey you lose in detailed know-how on the fine granular level, and vice versa. And how many of the neuroscience experts are proficient in computational mathematics, which clearly is being used to develop accurate and detailed models of the brain and its functioning. We turn to the experts. I can foresee a day when people specialize in the organization of knowledge—not library science—but like a project manager, someone who knows how to effectively integrate the specialists and their specific funds of knowledge. This specialty would be akin to a composer’s role, but the elements would not be notes, tempos, instrumental voices, and scales. It would be the various disciplines, with their specialists, and sub-specialists, woven together to create something that can no longer be resident, in all its details and mutual relations, in the head of one human being. It is not hard to imagine a day when the extension of human knowledge and cognitive function will be supplemented by interface—direct bio-technical interface—with aspects of computers and their architecture . Perhaps additionally, it will be common place for each of us to have a ‘pedagogical robot’ to organize, access, call up, and manipulate what has already exceeded the human scale . This is already taking place on a small and uncoordinated scale in algorithms that websites like “Google” and “Stumble Upon” utilize to guide their users’ internet searches.
In the interim, I find that the means of expanding our reach into the wide womb of possibility has led to a renewed dichotomy between what is possible and what is possible for me.

]]>