Main Area and Open Discussion > General Software Discussion

I'm thinking of going primitive, with discursion into zettelkasten

<< < (169/244) > >>

panzer:

 But I'm sure there is similar software for MarkDown.
-Shades (November 14, 2020, 09:34 PM)
--- End quote ---

https://www.gitbook.com/
http://daux.io/
https://www.mkdocs.org/
https://docsify.js.org/
https://docpress.github.io/
https://dotnet.github.io/docfx/

Dormouse:
Sorry to remain harping on about AsciiDoc
-Shades (November 14, 2020, 09:34 PM)
--- End quote ---
I'm happy for you to keep on with it.
For me asciidoc is better than markdown but worse than org-mode .
And none have some features that would be essential for full utility (for me). So, theoretically, I don't mind moving around. But so far I don't see another program that looks as if it will be as good as Obsidian will be.

I do consider Obsidian's block reference method a misstep, but it won't affect me if I don't use it. Apart from that, I think they're doing well.
(I think a better approach would have involved a better file explorer, options for automatically creating new notes from a next command with a folgezettel name that's hidden in editor, and using transclusions for display etc. That would have been elegant and stuck to their first principles, but couldn't have been done with a quick hack.)

Dormouse:
OK. Things have moved on swiftly since my last post and my Plan B has now become Plan A.

I said that I considered Obsidian's block reference method a hack. Not using it meant that it didn't affect me, but it was still something I noted.

I was prompted a few days ago to take a look at exactly what data Obsidian was storing and where it was stored. In the Discord it was stated that data was stored in the vault, but actually most is stored in systems folders predominantly, on Windows, in Appdata\Roaming\ObsidianCache which contains .json files for every vault that has been opened. The .json contains the names of all files in the vault, plus headings, and links - in plain sight. A user might encrypt the vaults themselves but this information would remain easily read.

(This can be dealt with by simply deleting the obsidian file in Roaming every time the program is closed - it will just rebuild it when it is opened again. Presumably there will be a speed penalty which will become noticeable with very large vaults.)

I assume that starting Obsidian calls or creates these files, which are used to provide fast access to items that can be linked. This allows fast responses without having to load the full content of files into memory. JIT method.

When I first used Obsidian all information was kept in the vault. The switch to the system folders came in 0.8.7 and was apparently to address problems with sync programs.

I also noted that one of the recent plugins deleted user data before it was updated to address the issue. Plugins are optional but regarded as a central feature of Obsidian's design.

I drew a number of conclusions.

* Until this is fixed, which I assume it will be, there is a potential security issue for any user who puts private information in file names or headings, unless they are confident of the security of their computer.
* The method was chosen without any thought to possible security implications, despite a thread on the forum where a number of users were finding ways to circumvent employers restrictions on installing software so that they could install and use Obsidian on work computers (this would increase their risk if their computer were audited).
* Together with the block reference methodology, my confidence in the expertise of the developer has reduced.
* The developer seems to be responding to pressure to increase the number of features which would only work well in a database. Given that a large proportion of the posting community seems to be made of students, many of whom were either previous users of Roam or aware of its features but not keen to pay Roam's price, I would conclude that this is the group driving the direction of travel.
* I see Obsidian currently as being a partial database with linked files. afaics it will end up as a full database program with associated files. They are discussing ways of saving folds between sessions, and it's hard to see how that can be done without a database. The argument will go 'we already have quite a large partial database, what would be lost if we move to a full database model?' and I think that's true. I now see Obsidian as a database program rather than a files based program  - I had assumed that this type of stuff was only in a cache which disappeared when the program was closed.
Since I don't want another database program, and since my needs aren't closely concordant to students, it seems likely that it will diverge further and further from something that works for me. So no longer Plan A, just a makedo until I have a better solution. The Plan B from last time is my new Plan A. And I'll maintain an open mind about the possibility of a better Plan B.

40hz:
@ Dormouse, Wraith808, JavaJones:

Thx for taking the time to answer my question in such detail.  :Thmbsup:

Interesting to see how, for you guys at least, this technology definitely is evolving to be an “extra” head. Which is how the authors of the software (especially Obsidian) seem to have envisioned it.  :)

wraith808:
@Dormouse- that's a no-brainer for me too.  It's one of the reasons that I stopped depending on Obsidian pretty quickly, and moved over to the Frankenstein model that I have (Using Foam and Memo extensions).  But the plugins with Visual Studio Code has been working really well for me.  I wish that I had the ability to embed references- it's one thing that would have solved some issues I've had.  But it has worked out fine, especially with previews inline.

And for anyone using VS code, I just found a new model built on top of it - Dendron (https://www.dendron.so/).  I can't say too much about it yet, but I'm definitely going to take a look to see what's different between that and the others.

Navigation

[0] Message Index

[#] Next page

[*] Previous page