textorama VS cloudcmd

I decided to go a little more deeply into them, and tonight, I uploaded a repository (https://github.com/joaomelo/textorama) with an open-source project I've been working on nights and weekends for the last few weeks.

I've been interested in two interlinked technologies for some time. The first is Progressive Web Apps (https://web.dev/progressive-web-apps) which lets the websites behave more like native apps, and the second is the File System API (https://developer.mozilla.org/en-US/docs/Web/API/File_System_Access_API) which enables browsers to access the local file system.

I decided to go a little more deeply into them, and tonight, I uploaded a repository (https://github.com/joaomelo/textorama) with an open-source project I've been working on nights and weekends for the last few weeks.

Textorama is a web editor that works with plain text files in your local system. The content is in your disk in a universal open format, and the browser ensures the editor only accesses the files you permit.

There is a working instance available here: https://textorama.online.

I assume it is this one: https://cloudcmd.io/

Personally, and I can't name a tool for you, I consider that:

- files&folder taxonomies are limited. Sometimes a file should be in more than one place, there are links/symlinks but no "backlinks" so it's easy top break things and filenames are not much good for search. Essentially a path in a file&folder classic taxonomy is a kind of limited and limited query to reach some content;

- notes are another interesting things: ALL documents are kind of notes. The fact we have many file formats and apps just to craft document is more a limit and an issue of modern systems that a reasonable thing.

Given the above two consideration I decide for myself to org-attach almost anything. The complete setup is:

- org-roam, org-ql (with a semi-curated catalog to make queries and yasnippets to ensure consistency) and ripgrep as access layer, witch practically means hitting a single key on my keyboard and start typing something. In 99% of the case I get "the good answer" (something already done or new content to add), sometimes I need rg/recoll because just heading/tags search do not work and in that case I adjust/add some roam_aliases to easy mach the content in the future. Sometimes I need queries to work on things, like "check all active contracts" or "current issue" or "last three days notes" etc;

- org-attach and links and dired to craft small "secondary-level file hierarchies" as a storage management layers, something that hide my real home taxonomy (essentially just notes on one root, other files managed by org-attach under another in a cache-like tree) I access via links;

- various org-mode extras to link different kind of stuff I can't org-attach properly, like mails (individual messages, threads, search queries on my mails etc), transactions (hledger via org-babel), mere elisp:(sexp) code to be executed live on click.

Doing so allow me to IGNORE a limited and limited hierarchy, allow crafting dynamic hierarchies as results from SQL-alike (albeit limited and slow) queries, accessing most of the content in search&narrow style something proven to be effective in most kind of UI from search engines to "dashes" instead of "menus" etc and allow to blend a bit most kind of docs in a single "document"/page/live environment witch is VERY useful since we have a single mind, not really compartmentalized and we need different kind of "docs" together often.

This is IMVHO how we should manage files in 2022 BUT since Emacs and classic desktop model for commercial and ignorance reasons is essentially dead it's not something ready out-of-the-box and not something designed for collaboration. It's just a personal HYPER-effective solution that might wrap&hide far less effective one used by collaborators still allowing interaction.

The modern equivalent, far more limited, complex and heavyweight is a DMS (see Nuxeo, Alfresco, ...) mostly crappy WebUIs that wrap Apache Jackrabbit behind the scene and add some forms/tags/ways to classify documents in various "dynamic" and "less constrained" ways. With a bit of hesitation for a small team https://www.tagspaces.org is less crazy to setup and use. Othe simpler but probably too limited options are https://github.com/filebrowser/filebrowser or https://cloudcmd.io/ or https://filerun.com/ or https://www.seafile.com/ or https://tabbles.net/ some are proprietary and all are not much more than classic file browsers served via webapp on a file-server backend storage instead of a local one.

Cloud Commander - Web based remote file manager, while there are a handful of them it's the one I keep coming back to.

Fire up a Docker container of cloudcmd, map your volumes, and go nuts.

check out Cloudcommander (like MC but in a browser)

You should have a look at cloudcmd. It's a browser-based file manager with drag&drop which also offers an SSH shell/terminal emulation.

Cloudcmd - browser-based ssh terminal and file manager (read: byobu, screen, and all the other terminal apps like taskbook, now count as being 'self-hosted') - - there are a few browser-based RDP programs like Apache Guacamole Server, but I haven't tried them (yet).

Cloudcmd?