Skip to content

Author: Frode Hegland

Re: Sam’s Dictapedia

In response to Sam’s post.

First of all, as discussed on Skype today with Sam and Stan, Liquid Server can be expanded to do some of this. I have emailed the following to the original developer, Tobias, to see what it would cost to do it this way:



First would be the general cleanup of the issue that the mouse over the menu items (nothing visual happens for MO) and creating a new account needs to work, not solid now.


User Interaction:

• User selects text on blog.
• Liquid list comes up. Top level includes: Glossary. Sub menu is names of any person on this wordpress site which has written a glossary definition for this word. (If no-one, then this menu item will not appear at all.)
• User choses a persons name and is presented with the text of the definition in the normal Liquid window.
• Top left in the window, where the name of the site would be, appears the name of the author of this definition. Click to open that persons glossary definition post.


• The possible list of definitions are from those authorised to be a writer on the wordpress blog.
• This means the system needs to have an index of all the definitions so that the menu can be grown dynamically (on page load I suggest).
• Results are shown as in current prototype.

How to create a Definition:

Data comes from the wordpress site itself: User posts the subject in the format: “Definition: Word” (non-case sensitive).

The body then becomes the glossary entry.





Other Issues


The above system is quite old and I never really manage to get it off the ground. It was co-developed by Fleur and Rob, back in the day.

Concerns about developing it now is that it’s for one platform only, I’d like to be part of a Liquid system, which is faster and works in all apps, but I appreciate that now all KF members are OS X users :-)



We could also look at how Wordnik scrapes for definitions, learning them automatically.

oh, this is an area ripe for innovation….


1 Comment

Response 1 to Sam’s Response to Liquid Space

This is in response to Sam’s Post:

Here are my comments to Sam’s response to my post on Liquid | Space:



Multiple Layouts – created by author

Multiple layouts – selected / controlled by user

Thinking about (all multiples)

relationships (peer-level associations), as well as

drilling into detail (drilling down), as well as

understanding context (drilling up)1


I agree that we need to have and key multiple layouts, what I refer to as ‘Views’. Sam, do you agree it’s the same thing? I’m not sure about the mechanism for this yet though, something to think about hard.

As far as drilling down and up is concerned, my current thinking is to keep it simple and allow the user to link to other .space documents which are rendered in the linking document as a live preview, for all intents and purposes acting like a transclusion, like a portal. The opposite issue, showing how the current document/canvas fits in a ‘higher’ level, is a hard one, since multiple other documents can refer to it and therefore be ‘above’. We therefore need some mechanism to show ‘backlinks’ in this world.

For these issues I hope we can be more inventive than simply adding buttons.

Changes to original document, in new version

Add these as issues to be worked on.


Visual Design

For diagrams, it is useful to use visual indicators (colors, line style, line width, icons, size, etc) to bring out intended message or meme to be conveyed in that presentation instance (message)

Diagram operations can be a separate discussion, as it can be quite an extensive set of actions / use cases


in AUTHOR mode, PRESENTATION is for the AUTHOR, and assists in the creation / authoring / editing / reorganizing / etc. of the object

in READING / VIEWING / NAVIGATING modes, PRESENTATION is for the READER / / and provides ways that ENHANCE the reader’s comprehension of the material2


The question of visual indicators is difficult and important. This can really help communicate, but it can also quickly become a language only the author understands, so needs real work to design.

As for presentation issue, yes, having different modes for author thinking and presentation and reader access is important.

Changes to original document, in new version

Add issue of visual indicators as a point to work on.

Look into modes for Read and Edit.



on keyboards, keyboard shortcuts are faster than mouse operations, and ought be available, at least until keyboards are obsoleted

Rather than having multiple actions create different TYPES of an object class, a SINGLE action ought create an object instance, with easy EDIT-PROPERTIES that allow tweaking it or changing its type, attributes, etc.

HIGHLIGHT NODE – the idea of FOCUS is introduced here – very important. FOCUS can be invoked by click (mouse enabled), keyboard shortcut, glare (eye-sensitive devices)

FIELDS – ought be:

STANDARD / DEFAULT – system-supplied, always available

CUSTOM – user-defined

VISUAL – user / reader CAN choose this, as well as experiencing the AUTHOR’s specified visual behavior


key discussion is CATEGORY TAXONOMY vs TAGS

The LAZY way is tags

most likely to be used / adopted

The intellectually cleaner / elegant way is CATEGORY TAXONOMY

less likely to be used / adopted


adding by drag/drop -> leaves a TODO item to review HOW to integrated it into the FOCUS DOCUMENT3


Keyboard shortcuts are important but this should also take good use of gestures and direct manipulation as via a tablets and laptop with multi-touch trackpad.

The importance of instances and inheritance is crucial.

We could have a very cool way where user types a tag and immediately, visually, a web grows above it, linking it to what is in the taxonomy, letting the user tap on which connections are correct for this tag.

Changes to original document, in new version

Consider keyboard issues as part of the interaction method design, nothing to add right now.

Put down section for inheritance and instances.

We should have a meeting to define the fields as well.


Invention based on this post

Instead of adding tags being plain text or auto-complete, show a visual web/hierarchy connected to the tag based on the document’s world.



1. Sam’s response to Frode’s “Liquid | Space (thoughts for spec)” | Knowledge Federation – The Journal. (2015)

2. Sam’s response to Frode’s “Liquid | Space (thoughts for spec)” | Knowledge Federation – The Journal. (2015)

3. Sam’s response to Frode’s “Liquid | Space (thoughts for spec)” | Knowledge Federation – The Journal. (2015)

1 Comment

How to respond to documents/posts?…

This was orignally written in reposnse to Sam’s post in response to my post on Liquid | Space.

This was written in Author, with citations and headings and pasted into a WordPress posting app (the official one, on iOS) where most of this went missing. The original .author document is available from

Diary Entry of Sorts

I’m sitting at Starbucks in Putney, London, drinking my old favourite, a con-panna, which is two shots of espresso, whipped cream, some sugar and coffee beans on top, mixed, then iced. No healthy for the body but sometimes necessary for the mind.

Of course I am writing this using Author, on my iPad Air 2, propped up on a 12 South Stand (it’s nice, but too heavy and not tall enough really), typing away on my normal Apple keyboard. The promo video for Author for reading is done, with a quote from Vint to start it off. It’s on Facebook and I hope to be able to promote it there when version 1.6 of Author is approved, which should be any day now. It fixes the crash on start bug which only journalist early users have experienced. Yup. Murphy was involved.

Note after writing this document: It would seem that I came to a useful concision of the last issue discussed, ‘threading’ and I would very much welcome comments, while I did not really get to dive into the document space, something I look forward to discussing with my esteemed Amigos.
Document Dialogue

David Price suggested that the third video should be about how documents relate to each other (the first was Reading and the second was Editing). Now that I have put together a proposal for a basic concept map application (tentatively called Liquid | Space) and Sam has blogged (and emailed) a response, I have to figure out a model for how best to respond to him.

The prime objective is to write my reply in such a way that Sam and anyone else can easily follow the thread of the conversation.

Should I publish on WordPress as a comment or a new post? Should I publish an Author document by emailing it or posting it on the web? Should I maybe design a new discussion forum system?…
As Author Document

Ideally I want to write the reply here in Author but I cannot expect the whole world to read documents using Author and a free-floating document (not published onto the web using a URL to find it or academically using a reference number to find it) will be an issue. If I publish it to a server it would have a URL, but would not be readable by a web browser.

As WordPress Post

Publishing my reply via WordPress would keep things nicely in one place to make it easy to follow the thread, if posted as a new post but is linked from the Sam’s post with a link in the comments. I am however tempted to add a link to the original Author document, since I’ll be writing it in Author.
In New Discussion Forum

Too much work. Stop it now.
Results of Reply

I’ll be replying with basically three types of replies for the issues he highlights: Yes, great idea, I’ll add it to the main spec, No, I dont’ agree and here are reasons.

The point is that my responses will be twofold; comments on Sam’s post and a new Spec document, which should also refer back to the original spec, Sam’s comments and my comments + more of the same in the future.
Free-form vs Meta

There is a temptation to add meta information to the Author document saying what it is in response to. An alternative is to simply write it in the document, as I have here, first sentence all the way up top there.

The Document Space

This leads to questions of the document space. How should documents connect?

Links point to a document but they are ‘dumb’ and not stable.

Citations point with more knowledge of what the point to and point to data rather than location (author name, title and publisher etc.).

Doug had powerful arguments for a back-link database so that all documents knew what documents were pointing to them. Today we have Google searches.

I propose that we develop routines and best practices to give documents in points and out points that are easily searched and letting the reader applications find the thread. A design problem for me is that I don’t like having visual interaction elements in documents and users would not necessarily know how to add this met-to the document.

Let’s play with one idea: On creating a document there is a field, right under the ‘Name of Document’ field, which is ‘Document is in response to’ field for a URL or other document information. This is nice and will be part of the new documents meta-information. Then we get onto the issue of threading:

How to refer to each of Sam’s items is also an interesting challenge. Should they all appear in a spreadsheet format? Or should they appear as citations, each with Sam’s name etc. under them? Or something else?
The thing is, it should be possible to comment on Sam’s issues one by one, or in blocks, while making it clear what he wrote and what I wrote. Furthermore, this should be done in a way that the document ‘understands’ which is which, so that we can build further interactions in the future to help people read through document dialogues. I’ll start here and see how it looks, using Author’s Paste As Citation command on the first section I copied from his post:
The text that follows is from Sam:


Multiple Layouts – created by author

Multiple layouts – selected / controlled by user

Thinking about (all multiples)

relationships (peer-level associations), as well as

drilling into detail (drilling down), as well as

understanding context (drilling up)*
You can see that it is a quote, from the * at the end, but since I used the ‘paraphrased’ category rather than ‘quote’ that’s all you can see. I proposed we add a category called ‘Minimalist Quote’, ‘Quick Quote’, or something like that, and show the text as slightly greyed out to indented. Or both.

1 Comment