Hi CRIMPers,
I came across this little corner of the internet when @jimspoon pointed the group to BrainTool[1] (in this thread[2]). I'm glad I did!

I started following @davewiner on twitter recently and understand from his material[3] that an outliner can be understood to be a tool that provides a way to manage hierarchies of information. Standalone outliners died out as hierarchy editors got built into applications. I have been building BrainTool as an outliner that lives in your browser and allows you to create and manage a hierarchy of topics with associated notes and pointers to informational resources.

There's a recent batch of knowledge management and productivity tools like Obsidian, Notion, Roam etc which espouse a less hierarchical, more networked, view of managed information. They seem to be gaining lots of traction.

BrainTool uses org-mode syntax to store the topic hierarchy and so can readily model a network of links between nodes in the hierarchy, but for now it only supports a 'forest' model of multiple top level topics each with a branching set of sub topics.

Does this group have any thoughts or wisdom on hierarchical vs networked models? How do you think about the information you store? Does it really matter if the tree editing and search are good enough?

Thanks for any thoughts.
Tony

[1] https://braintool.org
[2] https://www.outlinersoftware.com/topics/viewt/9390
[3] http://outliners.scripting.com/

Looks like a powerful tool for power browser users. I also appreciate that you opened up outlinersoftware.com in your demo video, as I'm sure my fellow CRIMPers will too.

Steve

I, for one, can make much more sense of things through hierarchies than 'flat' networks, but I understand that reality very often can't be classified in hierarchical boxes. See D. Weinberger's "Everything is miscellaneous".

One approach that I have seen taken by several outliners is the concept of clones, whereby an item can belong to more than one 'box'. This is often referred to as multiple parents. It is equivalent to symbolic links in the file system, which allow a file or folder to be accessed from several folders.

Another approach is to tag items, and to organise the tags in a hierarchy (rather than the items themselves). As an item can have more than one tag, this solves the issue of complementary categorisations.


BrainTool wrote:

Does this group have any thoughts or wisdom on hierarchical vs networked
models? How do you think about the information you store?

There are ways of creating hierarchies within Roam, Logseq, Obsidian as well. I personally have settled on Logseq, which has its roots in Org Mode, and have found their implementation of "page tags" to be a useful way to organize certain highly used top-level categories (such as contacts and books). And since both Roam and Logseq are outliners, hierarchy is central to how you enter information in pages, and how that information shows up in your backlinks.

Cloning can be immensely useful. I spend a lot of my working life gathering, organizing, and searching large sets of data, and I've found that creating a set of mutually exclusive topics or categories is usually difficult, if not impossible. There are almost inevitably items of information that belong in more than one place in your hierarchy.

I agree with @Alexander – hierarchies are more intuitively useful, for some reason.

Having said which, I love the idea of combining hierarchies with a mindmap-like approach: some of the best mindmappers (e.g. XMind) can support mini-outlines or mini-tables in nodes.

I think mindmapping is evolving in much the same way as the recent evolution in task managers: developers have realised that support for graphics, tags and fonts is not enough, and that what users really want is to manage large quantities of information using a variety of methods. The explosion of support for extensive notes in regularly upgraded modern task managers (notably TickTick, Pagico, Todoist and NotePlan) has been interesting to watch – rich-text notes and attached documents have become standard fare, allowing people to put together complete project management “hubs” in a single app. Adding different views is the next step: three of those apps offer various calendar views, and three of them also offer kanban views.

Mindmapping is the next step, perhaps.

Thanks folks, this is good food for thought.

I agree that some combination of hierarchy and cross linking is whats needed. You can certainly create hierarchies with bidirectional links in Roam et al as @Luhmann says, and any note taker tool will have nested bullets somewhere. However like @Alexander and @MadaboutDana, I personally tend to think of the hierarchy as primary. That said, like @James, sometimes I like to change the orientation of the hierarchy (eg instead of Projects:BrainTool:Resources:Chrome-Extensions:page.html I might want to see Resources:BrainTool:Chrome-Extensions:page.html).

My thought for BrainTool in the long run is to combine the notions of Topics (aka outliner headings), Tags *and* bidirectional Links. Each Topic is also a Tag and tags are inherited along the topic tree. Making a link between two topics basically adds the other topics tags to its linked counterpart. I can model this nicely in org-mode with the existing tags and links syntax.

So if I save a web page under Projects:BrainTool:Resources:Chrome-Extensions and then link it to the topic Resources:Healthcare:Standards:FHIR then the saved web page and associated notes will be tagged with all combinations of those Topics and (optionally) 'cloned' to show up in both places in the tree. BTW Roam and Obsidian support 'transclusion' which is, I think, just a fancy name for cloning.

I think the above can make sense in the context of categorizing data and taking notes as you work in a browser in a hierarchical fashion but also support the kind of UIs and visualizations you see in those networked tools. (BTW @Luhmann I like the look of Logseq, I had not realized it was based on org. Maybe I can do an easy integration.)

PS @Stephen I'd forgotten outlinersoftware was in the video! It has become one of my regular reads as BrainTool has started to take shape, so it deserves its moment in the spotlight!

Interesting; you’re describing something like Excel’s PivotTable function, but for hierarchical setups. Challenging!

Many of us are still mourning the untimely death of the hierarchical outliner par excellence, Tree2, which – while it didn’t do PivotTables as such – could extend outlines horizontally as well as vertically. The resulting flexibility was unprecedented. Dashword (also macOS) is a more recent attempt to do more or less the same thing, and is worth a look, not least because you can organise your horizontal outlines in named columns (they disappear in “vertical” view, but reappear when you switch back to “horizontal” view).

Cheers,
Bill

I think this is an interesting subject. Many of us require both, hierarchy and networking and many tools are heading in that direction. At the moment I find them way to clunky. Obsidian is certainly no thing of beauty! I also find search results and links not easily navigated when there is a lot of content. I'm sure they will improve. There's also the question of how these apps deal with the file-fallacy where people think it's in my system and that's enough. Somehow I need a place to throw every snippet and piece of information I collect and another place to work the ideas. The collection chamber however, needs search on steroids!

Neither is sufficient, nor even the combination.

it enrages me so much when these guys claim to emulate the brain when they design their applications:

- there was a talk of imitating the way the brain maps ideas when mind mapping tools came out first.
- and now, those hyperlinking tools again start to talk about building the second brain.

That is nonsense.

First, we really know little on how the brain contains ideas. It is very complicated issue: and the most attractive of the theories is "associationism" which is almost impossible to model in a digital system.

Here is the fundamental issue with hierarchies: most of knowledge is hierarchal. If you are studying the biological classification of organisms, there is always well build hierarchy. Mind maps might work in that case. But, in most other cases, ideas have no hierarchal relationships. Their relation is different sorts.

Networks are better, but they cannot replace the usual extended, long text.

The fundamental issue with networks: networks connect the linguistic terms (words and phrases). But, the real link is not between the terms; it is between the concepts. A concept map is different from a lexical map. (And, on the practical side, hyperlinks created within apps are usually dysfunctional once they moved out of that app. You need a lot of complicated programming if you want to transfer the hyperlinks created in one map into another one. )

If you want to drive a complicated concept home, you need a traditional long text writing; not a number of hyperlinked small texts.

Assume you are reading a book; say Chomsky

Dellu wrote:

it enrages me so much when these guys claim to emulate the brain when
they design their applications:
That is nonsense.
First, we really know little on how the brain contains ideas. It is very
complicated issue:

Thank you for pointing this out. It is also one of my bugbears and makes me groan every time I hear those claims (although it doesn't deter me from using those applications).

As for the topic of this thread, I think hierarchies and networks as methods of organising information are both useful but for different purposes, and so they are not mutually exclusive, in fact it is very useful to have them both in the same app.

When I'm organising a small amount of information (such as a daily todo list or an outline for a relatively small piece of writing), then a hierarchical outline is more helpful.

When I need to navigate a large database, then a network (i.e. wiki links) is a more helpful way of doing it.

This is one reason I like RoamResearch (and previously ConnectedText): I can use outlines for managing daily todos and for writing smaller pieces, but then they are all interlinked to constitute a navigable and searchable network.

We've had personal computers now for 40 years and the Internet for 25 years. There should be no reason to choose between hierarchies and networks. However, if I had to choose one over the other, I'd take hierarchies... just thinking back to the days of Grandview, still the best app ever produced in my opinion.

I agree that saying networks are the way our brains work is false. In fact, I might even venture that hierarchies more closely match what our brains are always trying to do, and that is to make sense of what we see. We want to take the input of our senses and put them into a logical "story." An outline or hierarchy does that better than a network. Networks are very useful, but they are not how are brains work.

Steve

To the hierarchy/network duo, I’d add comparison. Personally, I find the ability to view multiple pieces of information side by side invaluable; it’s why I can’t bear Agenda, which, for all its power and clever features, still doesn’t support multiple windows.

So maybe hierarchy + network + multiview + pivot table?

Cheers!
Bill

MadaboutDana wrote:

Personally, I find the ability to view multiple pieces of information side by side invaluable
it’s why I can’t bear Agenda, which, for all its power and clever features, still doesn’t support multiple windows.

So maybe hierarchy + network + multiview + pivot table?

Precisely what InfoQube is all about !
(you can add Columns, Tags, Google and Evernote sync too :-)

Yes, I know, no native macOS app, but runs well on VMWare / Parallels
If someone wants to help have it run under Wine / Crossover, you're welcome to contact me

Pierre Paul Landry
IQ Designer
https://infoqubeim.com/drupal5/node/48

TheBrain's jump thoughts remain intriguing.

I'm not sure if I correctly understand the difference between network relationships and hierarchical.

I think I use tags to form what you're referring to as network relationships.

I'm reminded of a Dr. Who episode in which time (or in this case information management) is all wibbly-wobbly (not just one thing).

I suppose I'd add 'context' to the list. I don't think there'll be software to match our brains for that (though I suppose software could provide a whack to the head to nudge us).

An interesting example of outliner + comparison (apart from InfoQube – yes, okay, @pierre, you’ve convinced me: I’m finally going to try running this in Parallels!) is offered by Speare (www.speare.com), a web-based writing tool with multiple window panes (between which you can drag and drop paragraphs) as its primary selling feature.

It does look pretty neat, actually.

Cheers!
Bill

IQ takes time to learn. It is well worth the investment though.

Jon

MadaboutDana wrote:

An interesting example of outliner + comparison (apart from InfoQube
– yes, okay, @pierre, you’ve convinced me: I’m finally
going to try running this in Parallels!) is offered by Speare
(www.speare.com), a web-based writing tool with multiple window panes
(between which you can drag and drop paragraphs) as its primary selling
feature.

It does look pretty neat, actually.

Cheers!
Bill

I'm continuing to love the ideas and references this thread is generating. Thanks All!

I'm hearing general agreement on the primary need for hierarchy - which I define as organizing 'topics' (and their notes and pointers) according to a containment relationship (for some notion of containment).

In addition, the ability to link things 'horizontally' across hierarchies into a network is generally desired. The new set of bidirectional linking tools (Roam et al) add unnamed 'related' relationships to the containment ones. IMO if you look at the information you are organizing as a set of nested topics (as in BrainTool) then a bidirectional link between two topics is the same as tagging each with the others topic hierarchy. (BTW BrainTool is inspired by 'Topic Maps'. I wrote about them here[1].)

Then @Mdlynam's 'context' and @MadaboutDana's 'comparison' hint at the need to view the whole topic map at different levels of abstraction (eg zoom in and out, or split a view to navigate multiple paths independently).

All this is a long way from where BrainTool is right now, but I do think it is a good candidate for @Simons "collection chamber" (I'm working on the "search on steroids" part).

Thanks for the food for thought!

PS Totally agree that none of this is actually modeling our brains. But maybe it's moving us along a continuum of increasingly sophisticated ways of getting thoughts and ideas out of our heads. From language, to writing, to the printing press, to telecommunications, to PCs, to the internet, to having all connected information in our pockets.

Once we get those neural implants [2] perfected we'll be able to kick back in our chairs and annotate our memory palaces by smell if we want to! For now though, most of us are just happy to be able to close out some browser tabs without stressing that we'll never be able to find them again.

Happy CRIMPing.

[1] https://braintool.org/2021/05/15/Organizing-your-life-with-a-Topic-Manager.html
[2] https://www.techtimes.com/articles/262957/20210716/neuroscience-researchers-use-neural-implant-help-paralyzed-man-talk-being.htm

Ah, I did love your reference to memory palaces! Good ole’ Conan Doyle…

Just been rereading (in what little leisure time I currently have) Neal Stephenson’s “Snow Crash”.

Now, what about the hypercard stacks that are such a vital part of the infowarrior’s arsenal in his Metaverse? Obviously they require a 3D environment – or do they? Efficient/effective or not? It would, of course, be delightful to have a Librarian like the one that accompanies the p(P)rotagonist during his research, but even there, what do we think about the Librarian’s inability to follow analogies? Or summarise (although reading it again, I find that he does – en effet – summarise quite a lot)?

Any thoughts?