4 ms·

These systems always strike me as a failure of the file system and desktop metaphor. _The computer_ was intended to be the system you stored documents in, with

by jrmg 4y ago

These systems always strike me as a failure of the file system and desktop metaphor. _The computer_ was intended to be the system you stored documents in, with paths to link between them, and any number of applications to do any number of things with them. Extended attributes are available for metadata. There’s not really any reason why things like Dropbox and iCloud couldn’t extend this to the Internet. But it somehow became too unwieldy to really work this way (for me, too - I’m not arguing that it’s a good way to do things with the systems we really have). It’s a shame.

swalling 4y ago
It’s true that the organizational model of the file system and desktop metaphor should be usable for this. There’s another huge gap though: documents in your local machine don’t have easy to use hypertext. Apple Notes app is very nice, but neither it nor TextEdit nor Pages can do what a Notion document can do. The OS needs a HyperCard to make it adopt Web-native tools for thinking.
- DerCed 4y ago
  Hence, Obsidian.md
  - whats_a_quasar 4y ago
    Yeah, to expand Obsidian is almost exactly what you're describing. It's a wiki-like software that works entirely on markdown and data on the file system. Has similar functionality to Notion but feels more like an IDE
    - swalling 4y ago
      Gotcha. I think I never dug into it because normal people generally don’t want to write in Markdown. The effort in doing that and manually managing cloud sync isn’t worth it. Apple needs to build it with WYSIWYG and have it work out of the box with iCloud.
lmm 4y ago
People have been trying to turn the file system into a database for decades at this point. It was supposed to be the killer feature of Windows Vista. IME the filesystem is both too complex and too simple; it's completely underspecified, has no memory model or semantics, and all the advanced features like extended attributes are unportable to the point of uselessness. There's literally one filesystem metadata feature that actually works reasonably reliably and it's the one most of HN hates: Windows three-letter extensions to indicate file types. It would be great to have a portable, standardised, database-like layer - SQL isn't it either - but I think it would have to be something ran as a userspace layer rather than built into the filesystem.
npteljes 4y ago
>a failure of the file system and desktop metaphor. I agree. I think this is because the file system is very rigid in terms of what you can do with it, especially regarding extending it, and managing it from a higher perspective. It would be very nice to use labels/tags for files and folders. To have a nice index that can be easily viewed, sorted, filtered. The file systems are also very rigid. To create filesystems while computer is running, mounting them around, resizing etc on the fly, this is unthinkable to me with existing file systems, and I have used some. The lacking of file systems are demonstrated I believe in file formats that are basically file systems themselves. Archive formats come first to mind, and that basically every game I have played grouped its assets into large data files.

by jrmg 4y ago

swalling 4y ago

It’s true that the organizational model of the file system and desktop metaphor should be usable for this. There’s another huge gap though: documents in your local machine don’t have easy to use hypertext. Apple Notes app is very nice, but neither it nor TextEdit nor Pages can do what a Notion document can do. The OS needs a HyperCard to make it adopt Web-native tools for thinking.

DerCed 4y ago
Hence, Obsidian.md
- whats_a_quasar 4y ago
  Yeah, to expand Obsidian is almost exactly what you're describing. It's a wiki-like software that works entirely on markdown and data on the file system. Has similar functionality to Notion but feels more like an IDE
  - swalling 4y ago
    Gotcha. I think I never dug into it because normal people generally don’t want to write in Markdown. The effort in doing that and manually managing cloud sync isn’t worth it. Apple needs to build it with WYSIWYG and have it work out of the box with iCloud.

lmm 4y ago

People have been trying to turn the file system into a database for decades at this point. It was supposed to be the killer feature of Windows Vista. IME the filesystem is both too complex and too simple; it's completely underspecified, has no memory model or semantics, and all the advanced features like extended attributes are unportable to the point of uselessness. There's literally one filesystem metadata feature that actually works reasonably reliably and it's the one most of HN hates: Windows three-letter extensions to indicate file types. It would be great to have a portable, standardised, database-like layer - SQL isn't it either - but I think it would have to be something ran as a userspace layer rather than built into the filesystem.

npteljes 4y ago

>a failure of the file system and desktop metaphor. I agree. I think this is because the file system is very rigid in terms of what you can do with it, especially regarding extending it, and managing it from a higher perspective. It would be very nice to use labels/tags for files and folders. To have a nice index that can be easily viewed, sorted, filtered. The file systems are also very rigid. To create filesystems while computer is running, mounting them around, resizing etc on the fly, this is unthinkable to me with existing file systems, and I have used some. The lacking of file systems are demonstrated I believe in file formats that are basically file systems themselves. Archive formats come first to mind, and that basically every game I have played grouped its assets into large data files.