More gRSShopper

I have played quite a lot with gRSShopper because I like it.

Result 1

One result is a list of Stephen Downes’s articles, by categories:

I still have not found how I can use the categories table for the optlist in the entry form.

Result 2

I am rather proud of the next result,

that allows you to Display a gRSShopper List or add to a Condensr Map (Condensr is my own think tool). The sample data used is the same as in a previous blog post.

First, this is a gRSShopper page

and the Javascript is https://github.com/x28de/mytooldemo,
the demo version of my free Java program.

It is called from another page

[<a href="<st_cgi>page.cgi?x28cat=[*x28cat_id*]">List</a>]
[<a  href="http://mmelcher.org/x28map.html?http://mmelcher.org/cgi-bin/page.cgi?x28cat=[*x28cat_id*]&format=condensr" target="_map">Map</a>] [*x28cat_cat*]<br />

on the same server (to avoid a CORB problem), but I was not able to
use the <st_cgi> tag again because I don’t have https where the javascript
comes from, due to an old plan.

This page, in turn, points to a view called x28cat_condensr:

<admin x28cat,[*x28cat_id*]>
<keyword db=x28text;cat=[*x28cat_cat*];format=condensr;sort=crdate DESC;>

and a view called x28text_condensr

[*x28text_title*]    [*x28text_excerpt*]

(with a tab in between), and such tab-separated lists (label TAB detail)
are the input that can be dragged and dropped onto my canvas.

Five pictures

To understand gRSShopper better, I made the 5 pictures, according to Christina Wodtke’s 5 models

  • “Mind Maps, to gather your thoughts
  • Concept Maps, to organize your understanding
  • System Maps, to map the system (a tautology, but an accurate one)
  • Mental Models, to understand and communicate your user’s understanding
  • Concept models, to message a way to think about a complex system”

(which I had also done her https://github.com/x28de/mytool/wiki/Design for my own tool).

Concept model

Mental model

System map

Concept map

Mindmap

The mindmap’s content (without the cross connections) is better readable here:

gRSShopper

  • elided benefits of RSS
    • harvested in the background
    • decentralized
      • no hub needed
      • no gatekeeper
      • no patronizing
    • pull, not push
      • less incentive for loud voices
    • no stream enforced
      • less context-switching
      • less built-in confusion
        • no need for recommendations
      • more room for infrequent writers
    • no tracking and privacy abuse
  • full workflow
    • aggregate
    • organize
    • publish
      • even distribute
  • incredibly powerful common interface
    • for common tasks
      • many channels in & out
      • many shipped examples
    • for custom tables
    • even for system tables
      • sometimes scary
  • emerged from practice
    • proven & very effective
    • many subtle accelerators
      • need time to appreciate
    • continuously improved
      • perpetual beta?
    • tailored & biased
      • courageous ‘under the engine hood’
      • specialities for large audiences
  • beyond consumption
    • handy for involvement
      • annotate/ refine/ put in context
      • bookmark/ share/ tag
    • convenience for making
      • comment or cite
      • curate or compile
      • write
    • beyond page impressions
      • pre-process the incoming
      • store resource meta-info

 

Posted in Personal Productivity | Leave a comment

#el30 Technical: Week minus one

I just realized that week ‘-1’ (Getting Ready) has already started yesterday. So here is a ‘Hello World’ test posting with the new category “el30”, according to the hashtag “#el30”.

The feed address (i.e., XML, not HTML) of this category is now
  https://x28newblog.wordpress.com/category/el30/feed/
(note the appended “feed/”, for Blogspot look here), and it can be harvested by gRSShopper as follows.

Install gRSShopper on Reclaim Hosting as described here.

In the PLE, click the database icon, then Make > [New] Feed:

Then click the “Harvest” tab:

Then insert the feed link:

Click anywhere outside the fields, and watch for a wait icon (spin) as shown here:

Finally your insert will be acknowledged with a green “OK”. Then you must approve the new feed, clicking “Approve”:

Now list the feeds (Make > [List] Feed) and note that the new one has now a green light:

Click the Edit icon (a pen) next to it and then the Harvest tab: you should see a green light, too, and new buttons such as “Harvest”:

I would recommend not to use this button but instead, set up the automatic harvesting via “Cron Job” in the Reclaim hosting cPanel, as explained in the above-mentioned installation instructions (and perhaps issue #6). After maximal 10 minutes waiting, the feed should harvested such that you can read the posts via “Read”.

Posted in EL30 | Leave a comment

Depolarization

Mike Caulfield speaks out for depolarization, and I think he is right. It seems like the only viable way to cope with the current cold war of lies.

The noble struggle for ‘truth’ reminds me of demanding some kind of ‘peace’ to end war — which was a big topic when I was young. Just as peace was often meant to be more than merely the end of war, truth is not the logical negation of lying. Not lying is the negation of lying, because truth it is not a simple binary thing, as Jenny Mackness’ thoughtful post impressively shows.

And lies have more in common with war. Accusing each other as liars, is like the brutal machinery of ‘kill or be killed’ that can’t be escaped: Once somebody threatens my credibility with calling me a liar, I have to destroy his or her credibility, because otherwise I will be very ‘dead’ in our information society. In fact, the defenders of a liar often honestly believe that the other side lies even more. And we won’t convince them with an arms-race of ridiculing, if they, for whatever reason, just distrust those who ‘own’ the truth.

Posted in Misc | 2 Comments

Exploring gRSShopper

I knew that gRSShopper is able to harvest RSS feeds, but it is much more powerful than that.

I wondered if it could replace my local WordPress note-taking, where I display my small note snippets on categoy pages. As a first test, however, I used the excerpts of my public blog.

Here is how I pieced the building blocks together. (Don’t be confused because all my tables and fields are prefixed with “x28”, just to tell them apart from the built-in stuff.)

1. Got a “Reclaim Hosting” webspace, so I have now a database with a database password, and a folder “public_html/” where I can transfer files.

2. Installed gRSShopper (following the instructions here).

3. Created two SQL import files of my data, using a quick utility intended for my own think tool — but it can also be used as standalone Java program (wxr2sql.jar, source here WXR2SQL.java) which converts the WordPress export format to SQL.

4. Using the phpMyAdmin tool from the Reclaim cPanel, I imported this SQL into my database.

5. Learn how to create a “view”:

  • navigate to your PLE: https:// yourservername/cgi-bin/page.cgi?page=PLE
  • in the upper left, click the green database icon to reveal the navigation pane
  • in the top left menu, click the “Make” tab;
  • in the row saying “[New][List] View”, click “New”;

6. Create a view with the View title of x28text_html and the following View text:

<h2>[*x28text_title*]</h2> 
<p>[*x28text_excerpt*]</p>

(To save, just click somewhere outside the entry fields. You should see a wait (spin) indicator somewhere below. Also, your new view should appear in Make > [List] View ).

7. Now you can try to retrieve a text, say, text number 3, by typing this URL into your browser: https:// yourservername/cgi-bin/page.cgi?x28text=3

8. Create a view with the View title of x28cat_html and the following View text:

<h3>[*x28cat_cat*]</h3>
<admin x28cat,[*x28cat_id*]>
<keyword db=x28text;cat=[*x28cat_cat*];format=html;>

9. Try to retrieve a category, say, category number 3, by typing this URL into your browser: https:// yourservername/cgi-bin/page.cgi?x28cat=3

Done. The magic is in the “<keyword ” construct.

To add even more subtle accelerators, I created a view called x28cat_list containing

<li><a href="<st_cgi>page.cgi?x28cat=
[*x28cat_id*]">[*x28cat_cat*]</a></li>

which is called from a new page

<keyword db=x28cat;format=list;></keyword>

Here is my page: page

Of course there was more cosmetic work to be done, such as finding and adapting the page_header, page_footer from Make > [List] Templates, and publishing the page to a “Page location” (.htm) via Make > [List] Page and the Page tab.

But the main building blocks are the colored ones that fit to each other as shown in this diagram:

Four code snippets with colored arrows connecting the matching terms.

Which is not very different from this:

Clipping from an IKEA assembly instruction, showing a screw and two holes, and an arrow through the first hole to the second hole.

(Source: IKEA)

Posted in Personal Productivity | Leave a comment

Blog parade “AI for Common Welfare?”

(Deutsche Version unten)

@algoethik is running a blog parade “AI for common welfare?”. I tried to respond to their questions.
Logo of @algoethik

  1. To what extent are ethical and Common Welfare aspects of AI already covered by the key points paper?
  2. Which relevant common welfare aspects are missing in the key points paper?
  3. A look outside: What have other countries done better in their strategies?
  4. What concrete measures should be included in the German AI strategy so that algorithms and AI are used for the benefit of society and for the benefit of all?

#1: The paper relentlessly mentions ethics and an ethics commission; it adds clauses like “taking into consideration…”, and so on. There are many honorable declarations of intent. I particularly like the “traceability” of algorithms. This is fine.

But (#2) I wonder if this intent can be sustained once the economic pressure grows and the jobs become scarce.

I do not believe that the change on the labour market only means that we will have different jobs (such that employees just need more education). We must not kid ourselves. The small number of new jobs for programmers and algorithm supervisors, or for managers with the much heralded new kind of soft-skills, cannot outweigh the jobs lost to the cognitive automation, IMHO. Understanding what AI can do now, we will no longer underestimate their competitive power on the labour market. (And there are even voices who say that today, humans themselves strive to become more machine-like — i.e., even easier to replace.)

The paper speaks of a “basis of justified trust and acceptability for users” (sectiion 3.9), and this needs, IMHO, a palpable guarantee (#4) and commitment for jobs. Not job security for existing jobs, but for sufficient jobs. And not just welfare (in a reservation where AI keep us benevolently as well-fed pets?), but sufficient gainful employmnet.

The paper sets the ambitious goal that Germany shall become a worldwide leading location for AI. If workers have a worldwide leading, justified, trust in their security, this may well happen, because then they will embrace the development rather than procrastinating and resisting it. Perhaps the time has come that an enforceable ‘right of work’ finally enters into the laws. For example, for every Euro of revenue, x Cents must be paid as wages — with these obligations being tradeable, of course. Otherwise, I think, Luddism will be inevitable, when workers feel like the Silesian weavers when the mechanical loom was introduced.

The horror scenario would be that algorithms are employed to administer the shortage of jobs. Any administration of shortages done by AI will be intolerable. Even if there was traceability of the algorithms. I also doubt that this traceability can really be achieved, except in simpler cases. After all, trial-and-error plays a big role in machine learning, doesn’t it? And if we hope that we can still supervise the algorithms, we must not forget the fatigue effect which inevitably occurs when 98 % of the AI decisions are reasonable. Perhaps a criterion for the admission of an algorithm would be that if it cannot be traced it must not administer shortages.

(Re question #3, I don’t know enough about others, but Finland seems to have thought about labour market aspects: see here.)

German:

@algoethik veranstaltet eine Blogparade. Ich versuche, die Fragen zu beantworten:

  • Inwieweit werden ethische Gesichtspunkte und Gemeinwohl-Aspekte von KI durch das Eckpunktepapier bereits abgedeckt?
  • Welche relevanten Gemeinwohl-Aspekte fehlen im Eckpunktepapier?
  • Der Blick nach draußen: Was haben andere Länder in ihren Strategien besser gemacht?
  • Welche konkreten Maßnahmen sollten in der deutschen KI-Strategie auftauchen, damit Algorithmen und KI in den Dienst der Gesellschaft und zum Wohle aller eingesetzt werden?

 

(1.) Das Papier erwähnt unablässig Ethik und eine Ethikkommission; es enthält Klauseln wie “unter Berücksichtigung von…”, usw. Es gibt viele ehrenwerte Absichtsbekundungen. Besonders gefällt mir die “Nachvollziehbarkeit” von Algorithmen. Das ist gut.

Aber (2.) ich frage mich: können diese Absichten aufrecht erhalten werden, wenn der ökonomische Druck steigt und die Arbeitsplätze knapp werden?

Ich glaube nicht dass der Wandel auf dem Arbeitsmarkt nur bedeutet, dass wir andere Jobs haben werden (so dass Arbeitnehmer nur mehr Weiterbildung brauchen). Wir dürfen uns nichts vormachen. Die wenigen neuen Arbeitsplätze für Programmierer und Algorithmen-Supervisoren, oder für Manager mit der vielbeschworenen neuen Art von Soft-Skills, können nicht die Verluste ausgleichen, die durch die Kognitive Automation entstehen, denke ich. Wenn wir verstehen, zu was KI jetzt fähig ist, können wir ihre Wettbewerbsposition auf dem Arbeitsmarkt nicht mehr unterschätzen. (Und manche Stimmen sagen, dass der Mensch selbst ehrgeizig immer maschinenähnlicher wird, also noch leichter zu ersetzen.)

Das Papier spricht von einer “Grundlage für berechtigtes Vertrauen und Akzeptanz für Nutzerinnen und Nutzer” (Abschnitt 3.9), und diese braucht m.E. eine greifbare, konkrete Garantie (4.) für Arbeitsplätze. Keine Garantie für die derzeitigen Stellen, aber für ausreichende Stellen. Und nicht nur Gemeinwohl im Sinne von Wohlfahrt (in einem Reservat, wo uns die gutwilligen KIs halten wie wohlgenährte Haustiere?), sondern ausreichend Erwerbsarbeit.

Das Papier setzt sich das ehrgeizige Ziel, dass Deutschland der weltweit führende Standort für KI werden soll. Wenn die Arbeitnehmer ein weltweit führendes, berechtigtes Vertrauen in ihre Sicherheit haben, kann das Ziel durchaus realistisch werden, denn dann können sie die Entwicklung bereitwillig annehmen, statt Verschleppug und Widerstand zu zeigen. Vielleicht ist nun die Zeit gekommen, dass letztendlich ein einklagbares “Recht auf Arbeit” in die Gesetzgebung einziehen muss. Etwa, von jedem Euro Umsatz müssen x Cent als Löhne aufgewendet werden — wobei diese Obligationen natürlich gehandelt werden können. Andernfalls, glaube ich, wird zwangsläufig eine Maschinenfeindlichkeit entstehen, wenn sich die Arbeitnehmer fühlen wie die schlesischen Weber, als der mechanische Webstuhl eingeführt wurde.

Das Horrorszenario wäre, dass Algorithmen die Mangelverwaltung für Arbeitsplätze übernehmen. Jede Mangelverwaltung durch KI wäre unerträglich. Sogar mit Nachvollziehbarkeit der Algorithmen. Ich bezweifle auch dass diese Nachvollziehbarkeit wirklich erreicht werden kann, außer in ganz einfachen Fällen. Schließlich spielt Trial-and-Error doch eine große Rolle baim maschinellen Lernen? Und wenn wir hoffen, dass wir die Algorithmen ja noch überwachen könnten, dürfen wir nicht den Ermüdungseffekt vergessen, der sich zwangsläufig einstellt wenn 98 % der KI-Entscheidungen vernünftig sind. Vielleicht wäre es ein Kriterium für die Zulassung eines Algorithmus, dass er keine Mangelverwaltung betreiben darf wenn er nicht nachvollziehbar ist.

(Zu 3. weiß ich zu wenig über andere, aber Finnland scheint schon über Aspekte des Arbeitsmarktes nachzudenken, siehe hier.)

 

Posted in Knowledge | 2 Comments

Cmaps and the “Split Attention Effect”

A big benefit of Concept Maps is that similar concepts are spatially close to each other — even those that are not explicitly connected by lines, become coherent by the gestalt principles. But what if you need more text to explain your ideas, such that the closeness gets lost? As Stephen Downes recently observed about concept maps:

“they’re the sort of thing that I think has never really made the transition to digital media. I think they just need bigger screens.”

It is hard to swallow that despite all the urge for digitalization, exactly this tool is left behind which might so optimally support networked thinking.

It seems that instructional designers traditionally shy at the “Split Attention Effect” that was described in 1992 by Chandler and Sweller. Which says that the cognitive load is increased if an annotation is not immediately next to the item it refers to.

I wonder if this is still as true as it was 1992 in a paper world, where the annotations had to be found somewhere in the linear text. Today, it is possible to catch annotations with an eye saccade movement if it appears always at the same location — such as the preview pane in the “coordinated views” technique used by ordinary email clients.

An eye saccade works extremely fast and effortless because the jump is ballistic (like a missile), i.e. the target fixation is determined at the start and it cannot be changed later (I learned this from S. P. Ballstaedt, on p. 20 of his German lecture notes).

In my experience, this works much better than a popup annotation which is commonly used as the digital equivalent of the callouts bubble in a diagram of the paper era.

Now this is the effect that is the basis of my own think tool Condensr.de — so I would love to hear your feedback!

Posted in Visualization | Leave a comment

Guesses about wisdom

Recently, the term ‘wisdom’ caught my eye multiple times, so I tried to clarify for myself what it means to me (and so I’m extending my old DIK post of 2005 to DIKW).

The term occurred in diverse contexts and senses:

  • as opposed to knowledge here on Mastodon,
  • as a gift which, like wealth and other powers, calls for moral accountability, here on Halfanhour,
  • as insight into the deepest structures and shapes of the universe, here by McGilchrist,
  • and as ‘wisdom of crowds’ here in a nice new ‘explorable explanation’.

1. In the Mastodon discussion, some definitions of wisdom seemed to me like just a higher level of knowledge, or like ‘meta’ knowledge, which I would rather call ‘critical literacies’: they depicted knowledge as simply knowing a googleable fact, and wisdom as something that is beyond that. But if this were true, everybody would have to be wise in the future — or else they would be displaced by artifical intelligences. Emergent knowledge based on long experience and intuition, will become more necessary but IMHO, it is not already wisdom.

Also, if knowledge cannot be transferred, or ‘told’, it is not already wisdom. Of course, wisdom cannot be told, but can it at least be taught? If teaching is Downes’s “to model and demonstrate”, this is certainly a useful prerequisite for developing wisdom, since we can sometimes recognize wisdom in other people even if we cannot describe it. But an important connotation of the term is, IMHO, that it takes a long time to develop wisdom. So, it is not for impatient teachers who expect an instant impact of their interventions.

In my understanding, wisdom grows very slowly, and it is often about what is really important, or actually, what is not important. For example, all the grumbling by the ‘wise’ wisenheimers who criticize everything but don’t offer an alternate solution — gradually becomes unimpressive.

2. By contrast, people who are wise enough to perceive their wisdom as a gift (and not as result of hard work that needs compensation), see their responsiblity.

3. McGilchrist, in turn, talks about the ‘path to wisdom’. And he likens it to how the hunter behaves (who “is careful, quiet, listening, attentive”). While this does not yet sound ‘left-brained’ to me, ‘Be a hunter of wisdom‘ (here) seems a bit misleading, because wisdom is not a target that can be shot, no more than happiness should be hunted for.

Furthermore, “insight into the deepest structures and shapes of the universe” is certainly a great thing, but it is so much entangled with fighting for truths about unknowable things, that I am not sure how important it should be to me. (Notwithstanding McGilchrist’s deep insight into the two modes of the brain, which are as important and fundamental to me as the flexors and extensors of the mind, i.e. relevant throughout everyday life.)

Another project of deep insight is currently undertaken by Dave Gray here (although he does not speak of ‘wisdom’ but of ‘level two’). His emphasis on ‘purpose’ can be seen as complemetary to the above ‘structures and shapes’, and while the latter might be seen as rather ‘right-brained’, the former is very ‘left-brained’. For me, wisdom seems more balanced.

A diagram showing 5 players, a button for launching the next match.

Screenshot of ncase.me/trust

4. In wisdom of crowds, finally, it is the aggregated knowledge of the many that creates the higher level, that cannot quickly be attained by a single human. And in a twin simulation here, it is the aggregation of many iterations of a trust game that leads a wise behavior — which, again, points to the long time that wisdom needs to be built.

Posted in Knowledge | 2 Comments

Myths, semi-myths etc.

I bought Clark Quinn’s new book about training myths, semi-myths and misconceptions, and I can whole-heartedly recommend this exciting, in-depth, clinical and precise work.

Continue reading

Posted in Learning | Leave a comment

Zooming is overrated

For the last release of my thought condensr tool, I had to think a lot about zooming. Besides text zooming — which was certainly important — the question was: Should I try to zoom the map, i.e., offer to edit the visual overview on different zoom levels?

Many people are enthusiastic about the ‘Big Picture’ attitude. Some even quote Dyson’s distinction between birds (who “fly high”) and frogs (who “live in the mud”).

A boss flying above a landscape, and some frogs sitting near the bottom.Photo credit: Flickr user youdid2, cc-by-nc-sa.

But I am sceptical, and I doubt that such a detached view of a ‘Big’ picture can be the necessary rich picture. It must leave out a lot, and usually there is a predetermined hierarchy that controls what to show and what to hide.

A geographical map, for example, needs some cartographic generalisation for zooming out: the labels of the big towns must be bigger than the others, and the smaller towns are gradually omitted, just as the smaller rivers and streets vanish. Otherwise there would be no orientation possible on the larger area. (This is why Google maps are becoming unusable: because the locations of the advertisers cause ever more clutter.) And this hierarchy may have distorted the rich picture already.

In my tool, there is no pre-existing hierarchy. On the contrary: it tries to facilitate an overview without premature hierarchy-building and pigeon-holing. And it has a different approach to handling the relationship between overview and detail: all the overview on the (visual) left pane, all details on the right (text) pane.

So, my new ‘bird’s eye view’ is just for making long-distance moves easier — not for withdrawn attitudes.

Posted in Visualization | Leave a comment

Time: Back or forward?

Many people were confused about whether the clock would be switched forward or back today. This is a nice opportunity to visualize how unintuitive our abstract concepts can be, even if they seem so natural all year long. Why do we hesitate to accept that we have to forward “the time” ? Because it seems that in reality, it is being switched back: sunset is delayed.

A landscape with the sun being pushed back, and an overlaying clock-face showing the spring time switching.

It is a similar effect as the confusion about how a document on the screen should be moved when we want to proceed to the later sections: Scroll down, or pan or swipe upwards? We’ve got accustomed to ignore the reality of the paper, and use the scroll bar as a tool of abstraction. But now ever more user interfaces want it the other way around.

In McGilchrist’s terms, the abstraction tools such as the scroll bar or the clock, belong to the “emissary” mode of thinking, while the ‘real world’ such as the sun or the paper, belong to the “master”. (Actually, he talks about two hemispheres seeing the world differently. But for me, thinking of my brain as two separate “seeing” entities, is rather unintuitive, too. I prefer to think of their difference in much more mundane and elementary ways, as the flexors and extensors of the mind.)

Posted in Intuition | Leave a comment