Dave Winer, who probably coded the first PC outliners in the 80s, is trying to wrap his head around the new outliners in the market, and has some interesting thoughts in his blog.

Most recently he's discussed the use of double-square-brackets, and sharing, and outline-blogging:

"There are also big differences in the way our outliners work. I could explain that in a podcast, but don't have the time to write it up now. The takeaway is that without open modular connections between the components via APIs, this new burst of outlineism will end up with a sad story of silos, not a wonderful explosion like the earlier open platforms. One silo vs a thriving ecosystem -- either approach could win. It's happened both ways in the past. We have to be careful to learn from past mistakes in evolving open platforms."

http://scripting.com/2021/05/13/124723.html

(Dave is also a bit of a grouch, and often has to revise his opinions [a good thing!] and tends to dismiss siloed apps that export and import to multiple formats. Still, an interesting read, and blog.)

I know I am not the only one who remembers that Dave was the original host of this forum, the one who got this ball rolling. I'm not sure of what he'd make of CRIMPing, and I don't think I want to find out.

Dave Winer wrote

this new burst of outlineism will end up with a sad
story of silos, not a wonderful explosion like the earlier open
platforms.

Well, there are more and more tools that support markdown. And there are Obsidian, Notebooks app and others that read and write MD files in the file system, Obsidian even supports links between files.
That seems a good way to avoid being trapped in a silo.

There is a necessary tension between experimentation and compatibility. If we had to agree to standards before creating each new app there would be no room for experimentation, but endless experimentation makes it hard to ever stop and develop standards. Look at how many different versions of Markdown there are and how many apps adapt Markdown in ways that make it slightly incompatible with other versions. At the same time, most of these apps all support some kind of plain text output, and it is possible to write scripts to "translate" the unique codes used in other apps. For instance, Obsidian does a decent enough job of translating Roam's markdown output into its own, though some stuff gets lost in the process. (Right now Roam supports nested titles like [[[[apple]]pie]] while Obsidian does not.) My feeling is that this is "good enough." Even though some workflows might be broken, it is all plain text and you can search and find stuff. In some ways, the power of modern search has made it less necessary to rely on standards. I think of the difference between the old Yahoo! search engine where everything was cataloged in a nested hierarchy and the modern Google search engine (or Spotlight on the Mac) where you can just find stuff even if it was filed away in the wrong place...

Sometimes ideas just get translated into another medium.

And the fact that it was possible to translate them perhaps means that there are no silos.

I haven't tried Obsidian, so I can't comment on it, but Roam to me is essentially an online version of ConnectedText, a personal wiki, which itself was a translation of online wiki standards.

ConnectedText only worked on one PC at a time (it was possible to work across Dropbox but it could get messy, so I never went down that route).

A Roam account on the other hand allows you to have multiple instances and you can easily switch from machine to machine and one platform to another.

To me that is progress.

And as Franz said, the openness is there in terms of using conventions such as Markdown and even wiki links.

Maybe I'm misunderstanding Winer's point. I liked OPML, but ultimately anything can become a standard, even something of very poor quality, if enough people adopt it (such as QWERTY).

So if Roam becomes a runaway success and everyone in the world will be using it, or a clone of it, well, it will set the standard, it won't be a silo.

Indeed, though what may not be findable are the relationships between that stuff--and this is often times as important as the stuff itself.

Luhmann wrote:

My feeling is that this is
"good enough." Even though some workflows might be broken, it is all
plain text and you can search and find stuff.

This is an important point. But leaving the relationships between stuff to a single app is, I increasingly think, a self-limiting strategy.

Increasingly, I’m looking at ways of creating my main data repositories in the file system, then using powerful and highly flexible search engines to tie it all together. I’ve been experimenting with FoxTrot Pro (on Mac) for a while now, and I find that research using the latter is much faster and more efficient than using any number of clever, attractive, but ultimately limited information management apps. There are, of course, any number of powerful desktop search engines available for Windows – more than for Mac, in fact.

Having said that: to manage the results of what you ferret out in FoxTrot Pro, you do need some kind of information management app… ;-)

And this is what’s interesting about info management – how many “layers of correlation” you need will depend on the ultimate purpose of your research. For single projects, a couple of layers will be enough (a search tool will usually allow you to save “smart” searches, which are sufficient for one-off projects). For a complex project, you will need many more layers of correlation. This will usually exceed the capabilities of a search engine, no matter how sophisticated it is, because information management = information structuring, which is almost always matricial (regardless of the human mind’s fondness for hierarchical structures).

Alexander Deliyannis wrote:

Indeed, though what may not be findable are the relationships between
that stuff--and this is often times as important as the stuff itself.

Luhmann wrote:
>My feeling is that this is
>"good enough." Even though some workflows might be broken, it is all
>plain text and you can search and find stuff.

This conversation may be going over my head, so if I'm stating something completely off-base apologies up front.

But when Dave Winer was talking about silos, I took him to mean that most of the applications he's refering to encourage links to notes internal to the app and not to the world outside the app. I don't use Roam or Obsidian much (though, of course, I've tried them), so I don't even know if that is true or not. But if it is, he may have a point (again, only if I interpreted him correctly). Dave has always been about creating wider networks of knowledge.

Steve

@Stephen:
Roam is a silo, Obsidian isn't.

Thanks Steve, re-reading Dave's post it seems clear that you interpretation is right, though the point is made towards the end of his text.

I can't think of any information management software --whether discussed here or not-- which does not reference the outside world, even in the form of simple URLs. Of course Dave is talking about open APIs and the like, which would allow one to leverage information structures already developed by others.

For me such a information management world would be ideal, and would take learning and knowledge management many times forward. That said, I believe that it's as much an issue of technology as it is of culture.


Stephen Zeoli wrote:

But when Dave Winer was talking about silos, I took him to mean that
most of the applications he's refering to encourage links to notes
internal to the app and not to the world outside the app. [..]
Dave has always been about creating wider networks of knowledge.

Franz Grieser wrote:

@Stephen:
Roam is a silo, Obsidian isn't.

You can export all your Roam notes as markdown files though, so there is no lock-in in that sense.There was some talk of Roam API coming at some point as well.

Of course it does depend on what you mean by "silo". Text can be extracted from any app. I don't use Roam or Obsidian, but from what I've read, everyone loves the linkage between the notes and the tagging. You may be able to export all your md files, but if all those connections no longer work you have a problem. I had this when I used to use Ulysses and they had their bespoke version of markdown. Yes I could export all my text, but lot's of functionality was lost.

If an app is classed as not being siloed because it can simply export the text, then most proprietary apps aren't silos. I would question how useful this is with thousands of interlinked notes?

Simon wrote:

I had this when I used to use Ulysses
and they had their bespoke version of markdown. Yes I could export all
my text, but lot's of functionality was lost.

The default choice in Ulysses is their own Markdown XL, because features Ulysses provides like Comments, Annotations, local image and other (notes, goals) embedding, Delete are not part of the original Markdown syntax.

But if you don't want to use those you can change it easily in Edit > Convert Markup, letting you choose between Markdown XL, regular Markdown, Minimark, Textile'd... or even your own custom variant, like grafting citations to documents.

https://ulysses.app/answers/reference-managers

Personally the only thing I dislike about Markdown XL is the lack of an option to always show URLs when writing. But I got used to the more WYSIWYG view.

Simon wrote:

Of course it does depend on what you mean by "silo". Text can be
extracted from any app. I don't use Roam or Obsidian, but from what I've
read, everyone loves the linkage between the notes and the tagging. You
may be able to export all your md files, but if all those connections no
longer work you have a problem. I had this when I used to use Ulysses
and they had their bespoke version of markdown. Yes I could export all
my text, but lot's of functionality was lost.

You have a point.

But your post made me look into my Roam export file and I was quite happy with what I saw.

As each link is denoted by the [[double brackets]] and the content within the double brackets is the name of the md file, it is not too hard to follow manually the links, as one just needs to look through the alphabetical list to find the target of the given link.

It would be usable even as a paper-based Zettelkasten, if it was printed out. It's essentially a simple library reference system.

I was also happy to see that the outline hierarchy of the individual 'blocks' as they're called (within a page) was also preserved.

Granted, I'm a very simple Roam user, I don't use aliases often or build any fancy query pages. Those automatically rendered pages might get lost.