aff... I understand that this is a good idea, but posting a project in HN without a release, and saying it will have, but have not implemented the cool features at the frontpage, like cloud integration, photos, search, tags, extensions, cloud integration(gdrive,dropbox,etc), key manager, etc... This is the dictionary´s definition of vaporware. Maybe all you want is to attract more developers to the project, but if all big project wannabe start posting in HN... what the heck.
Yes. Why would I trust all my files to this? Needs to be a lot further along. why is "Rust" a feature? I could care less the implementation language as long as the thing works as advertised.
This is open source software. Perhaps at this phase they're not interested in having you as a user?
You seem to have a lot of expectations & be uninterested in the underlying technology. Open source is about more than "things work as advertised;" it's also about picking good tech & insuring we have open paths to the future we can advance on.
Your belief that "Needs to be a lot further along," applies to who? Yourself? Everyone? This seems like a great pitch to me, at a great time to get onboard.
Rust is a feature because as anything written in rust is subject to one of the smartest typesystems on the planet, designed to make you do the right thing. It's also easier to maintain and refactor than many other languages.
I definitely see it as a feature, but I don't choose software solely based on this.
How is it fascinating? The compiler forces you to do things the right way, if someone's convinced the compiler that their program is valid it's proof of all the good thing the compiler enforces, which is more than "any other language".
It's also likely that if you're using rust you care about performance and have thought about it while building your application, this is also a feature.
What's fascinating to me is everyone who says written in Rust isn't a feature. It definitely is, and it's going to keep eating C and C++'s lunch in new areas every year.
After installing it from source, it doesn't appear to be functional. Running from main could be unstable though, so hard to determine if broken vs missing.
As one of the core team has mentioned, they didn't post it themselves, but rather a community member posted it.
It's definitely something that I'll be bookmarking for later though. It's a tool I could see myself using.
This post wasn't made by me or anyone else on the core team, but by a community member. I definitely agree that this post was made too early, as the app is missing a lot of core features as you point out.
The actual problem I need solving: multiple copies of files in multiple places, older versions of documents not linked to to the latest versions, disorganized backups.
Also - as a potential user, I would much prefer to read about useful features, rather than what language it's written in and why. (Rust! Webviews!)
That's why I use git for basically everything of value.
My friend, an engineer for a Fortune 500 was lamenting to me recently having had someone overwritten one of his designs on a shared drive. He had to get IT to pull a backup.
It absolutely perplexes me that DVCS isn't the norm outside of programming.
It perplexes me that application developers don't leverage git for their data model. Asking users to be the tip of the spear in advancing better data management seems like a big lift. But I fully expect by now we'd see more people develop apps that so happen to back their data-models with git.
There's irmin[1][2], a git-like database written in Ocaml. I was always confused by how git-like it was... at points I thought there was outright interoperation. But I'm not sure what if any software was ever written on irmin.
I have my own project pr0x that has a file-persistence engine, which I need to spend a weekend adding git support to sometime. Very high importance to me, but haven't made it a priority.
Again,
> It absolutely perplexes me that DVCS isn't the norm outside of programming.
Absolutely perplexes me that programming hasn't made DVCS the norm in the outside world. We're steering this ship. People don't even have to know what's going on... we should still be using the best class tools on this planet to be managing data, and what data do we better manage/is more important to us than source code? It's up to us to seize this initiative.
Maybe a silly question but: is this by accident or a dark pattern? I hate it when this happens in the wild, but I never know if I should be suspicious of a dark patter or not.
This looks like an accident. Even many of the suspicious sites that do it seem to be an accidental side effect. The ones that annoy me the most are the news articles that trigger a history event on scroll. I think they do it as part of a tracking mechanism to detect scroll position, but the result is you have to press back twice to leave. I can't see the benefit, not like anyone is going to press back then say, "oh i guess i can't leave so I'll click an ad".
The site wealthygorilla[.]com alters your history by erasing the current contents and replacing it with more of their articles. It irritated more way more then it should've the one time I went to their site. I went through and found the JS where they implemented that "feature".
Safari on ios too. I assume it is a bug because it would be a very counterproductive feature to use when presenting a project. As others have said: good idea, but with no release the risk of forgetting about it is high. I have starred many github projects but only gone back to see progress on few of them.
Please don't complain about tangential annoyances—things like article or website formats, name collisions, or back-button breakage. They're too common to be interesting.
I want to point out that this is being developed by some guy called Jamie Pine on twitch, but this doesn’t look to be posted by Jamie. I absolutely don’t think its ready to showcase on HN.
Maybe it's less showcasing the product and more fishing for sponsors/contributors/first adopters? Whatever it is, it seems to be moderately successful, at least judging by the attention it's getting...
I'd be more interested in a modern file system that I can use on xBSD / Linux, Windows, macOS. ZFS comes pretty close except Windows support is still beta quality and ZFS lacks a "tags / label" feature.
Pardon my ignorance, I'm slightly technically challenged. But isn't this something similar to ceph cluster or maybe HAMMER2 with an explorer/fileman utility wrapped around it?
Disclaimer: I think this project is silly in the face of tools like rclone + borg.
Hammer2's wikipedia article lists dragonfly bsd as it's solely supported os and ceph requires one to... Manage a ceph cluster.
Now the question of whether this is similar to those technologies becomes easy to answer: no, they aren't similar. This tool aggregates external storage providers while the tools you reference require you to be your own provider.
My personal recommendation (if this project interests you) is to check out initiatives like upspin.io and Tahoe lafs
I tried something similar with a project of mine[1], with a focus on API rather than frontend so I could build tailored apps for each device/user. At some point I should get around to finishing it -- I mostly focused on S3 and local filesystem compatibility, but being able to explore filesystems from other devices (maybe through an agent) is a neat idea.
Of course it's bad for accessibility, but for me only the "Coming soon..." text on the start page is hard to read. Maybe you have your screen set too dark? Anyway, it feels a bit harsh to discount the whole project because of an unfortunate design choice...
I was reading in the dark, so had brightness was low, but fine for any normal site.
This mistake is a common thing for people to complain about on accessibility forums. If you've lost lost any visual acuity, which a large percentage of the population has, your website is not usable.
I'm not sure if it actually does work this way, but this would be amazing if I could install a small daemon on my NAS/fileserver to do the scanning and indexing, which the graphical client could then pull the results of.
I love this, though I'm hoping your demo on the main page scrolling opposite my mouse settings is a bug of the in page demo and not an actual feature of your UI. I'll definitely be checking this one out.
Hum, try to imaging a new files manager: instead of force you traversing a tree it gives you a search&narrow UI, not just in files names but in a sort of metadata-rich notes/headings witch may also have a file or more attached or may not. Imaging you hit a key and type things like:
- "ph b" and just phone bills results appear
- "car ins" and car(s) insurance(s) appears
- "ph sum" and some summer photos "notes" with location and year tags start to show
- $your_host_name and it's NixOS config is there
- "zsh" and the relevant commented zsh config arrive
- "foo bar" and foo bar contact note arrive, with links to mail search, projects up with him/her, few noted threads, some files, contact info, ...
Well... That's how I manage 99% of my digital stuff, Emacs/org-mode(notes)/org-roam(UI to access|manage them)/org-attach(to have files there without having to manually manage them in a hierarchy), for mail notmuch and ol-notmuch to link searches and threads/messages etc and so on.
Why? Because my brain is a chaotic graph, not a specialized tree so while some treeviews are nice sometimes I need many of them for the same contents, depending on the context, because sometimes I have to note small things, not enough to be in a single file, with a name, properly placed somewhere for easy retrieving. Also I have only one brain, no reason to partition things like: this is where I look for notes, this for docs, this for emails, this for contacts. Did you know Google Search? Well if a unique UI is good to look for anything why not having the same on my desktop, locally, for my info? Why having countless separate apps with only context switch (Windows Management) and cut/copy/paste for IPCs? Unix at it's time at least offer decent CLIs IPCs, older systems offer more in GUIs, Emacs is an old and modern one who offer such IPCs and more important a total integration. Why have a collection of scripts, so files, some shell functions (more than one per file), aliases etc if anything can be just elisp?
Well... That's the file management of the very past (Xerox workstations, LispM etc) AND for the present AND for the future. All modern apps try to achieve this goal a small step at a time, like this project with tags, some have already goes farther (see tabbles.net, tagspaces.org and so on) but all quickly hit their architectural limits of the OS/environments they run and even for achieve just a small part of the game their complexity is extreme.
Long story short: we are told that we need separation, an app a job, we start following that model and quickly any apps try to incorporate and duplicate functions of others, they are hyper hard to integrate each others etc => separation works to sell software as products and services and surroundings a need for some against the others. We need unique systems, an application is the entire operating system and all the rest, easy to be mold by it's user. Witch happen to be a user, not a human-robot able just to click around.
Sorry for being rude, but IMVHO it's about time rediscover such classic concepts instead of investing countless resources in new stuff with same limits of the new classic model...
Because every single JavaScript UI I use is a bloated, slow mess - even “optimized” apps like vscode feel like sludge to me compared to Sublime Text or Nova.
Last thing I want is a sluggish file manager that uses a crap tone of memory with a bunch of “ui renderer processes”.
Check out Pathfinder for a power user file manager done right.
VSCode devs do some magic trickery to improve Electron's performance, but it is by design not in the league of Neovim or even fully fledged Qt Creator in terms of responsiveness.
I use all these editors, and I'd say that for purely editining purposes Neovim feels incredibly responsive, it's something that requires the whole UI framework to be fast. What keeps me from switching from VSCode are its great plugins
