Why would I want to store my files on someone else’s computer rather than in a git repo on mine?
The killer feature I need is a share target so I could select a snippet of text, hit share, and paste it into a particular note.
That said, I would definitely prefer pointing Glow to any standard git repo. This gives it all the security Charm is claiming to offer, versioning, etc, essentially for free. (Obviously that defeats the purpose of Glow being a marketing tool for Charm)
- You might believe the cloud folks do better backups than you.
While this looks like a really neat utility, I'll pass until there's a self-hosted option under a free license.
To each their own.
- ‘Cloud folks’ might have more stable hardware and software, but have built themselves a deadly reputation for unreliable policies.
Edit: It looks like they generate their own SSH key instead of using an already present one[2]. So you'd presumably need to copy that to each machine that you'd want to use so that it can unwrap the real (cloud-stored) decryption key.
[1]: https://github.com/charmbracelet/charm#how-it-works
[2]: https://github.com/charmbracelet/charm/blob/master/keygen/ke...
pandoc -i file.md | w3m -T text/htmlhttps://github.com/Orange-OpenSource/pandoc-terminal-writer
I contributed some changes back in 2019. The writer is written in Lua and the lead maintainer seems happy to welcome PRs.
It's functional but crude so nothing on github just yet ;)
That said, can someone please post a summary of what charm.sh is about? I don't use javascript and nothing comes up on the page.
Then I saw Goyo Vim plugin
Also, sadly GitHub and GitLab only have all/none access to private repos. OAuth access to specific repos isn't a thing. So Desktop support is more important for me.
Also, here is another similar project - https://github.com/taniarascia/takenote
The two main blockers are moving away from Firebase analytics (not open source) and setting up an alternative payment method.
It's something I want to get done by the end of the month [0]
https://github.com/nvim-treesitter/nvim-treesitter
Specifically, for Markdown: https://github.com/ikatyang/tree-sitter-markdown
We actually issue a new Charm specific SSH key for you. We then allow you to link machines together with our `charm` account utility. The symmetric keys used to encrypt data are encrypted for each public key linked to your account so you can access your data from multiple machines.
Just out of curiosity: have you seen or considered age[1] for the symmetric pair? I've used it on a few projects where cryptographic flexibility wasn't necessary.
Your concerns are well founded! I've certainly been burned by services disappearing...
It's a compromise to have a business model that allows us to develop Charm full-time vs. being completely open source. Our plan is to have a business model similar to GitHub: free and low cost services for individuals with enterprise hosting options (colo or hosted by us).
One thing that I think we should do is allow for a very easy export of all of your data. We can build that into glow and charm (vs. having to email us and ask for it or some other draconian option).
This is very much a 1.0 release of the Charm Cloud and one of the reasons we wanted to ship it early was to get feedback from the community and build out the features our users want. Your point is a great one that we should solve for ASAP. A glow export feature that spits out a .tar.gz of all your stashed markdowns seems like a great idea.
We're also open sourcing the libraries we use to build glow and charm, most of which don't require the Charm Cloud. We just made our TUI framework Bubble Tea open source for instance:
https://github.com/charmbracelet/bubbletea
We also have a library for applying JSON stylesheets to Markdown called Glamour that is independent of the Charm Cloud (and currently in use in GitHub's cli):
https://github.com/charmbracelet/glamour
Thanks for your question, it's a problem worth addressing.
FWIW, I get the impression that they want to charge money soon. Keybase never really wanted to bother with that and the implicit SLAs that come with it? Charging money is a great way to actually guarantee it will be around for a while longer than your typical ad or VC funded 'free' service.
I feel like the major problem with E2EE cloud services is getting apps to adopt it. For example, I like bear editor, but I don't like their encryption model. I like standard notes & inkdrop's E2EE encryption models, but I don't like them as editors that much since a bunch of their energy is probably being consumed by encryption stuff. It feels like the engineers who make good UIs have a hard time doing well on the security standpoint and the ones who bother to make things secure have a hard time dealing with making good UIs.
I want something that the bear's of the world will adopt and then they don't have to think about E2EE anymore.
Well-structured exports is certainly something that I think will make a lot of people more confident (and also a necessity if you intend to have first-class support for EU customers with any kind of individual identifiers, with GDPR and all).
I’d still put a dogfed (same as you run yourselves, possibly missing select enterprise-targeted features) FOSS self-hosted server version.
Look at Hashicorp for a great example of that this works in practice when you're targeting a techy audience!
I’m sure a lot of your potential customers are happy to pay exactly because they want someone else to take care of all the technicalities - and on the other end of the spectrum (and I’m sure you’ve thought of this already) you have the enterprises that for legal or compliance reasons simply have to own the whole stack - and there you’ll have free users providing back tutorials, documentation, bug report and the occasional PR that you’d otherwise have to do yourselves or neglect. Just food for thought :)
I absolutely agree that there’s a need for the middle ground.
I’d also say that I think nobody really cares if GH is running vanilla git, or which SMTP and IMAP server Fastmail is using; perhaps the important thing is that there exists at least one maintained fully API compatible open alternative, not necessarily 100% identical to the software run by the hosted, or even primarily developed or maintained by them.
(And, well, the insight that can be derived from metadata is not to be dismissed, so I wouldn’t agree on “almost irrelevant”)