https://gist.github.com/egeozcan/b27e11a7e776972d18603222fa5...
Live:
https://gistpreview.github.io/?b27e11a7e776972d18603222fa523...
Selecting via file-picker works too. Dragging usually does not. When all works, images are inserted inline as blobs.
After adding images, if you save the page (literally file->save), the blobs are saved together. don't want a part when saving (for example, removing images)? inspect element, remove, save page.
throw the page on some server or just double click on your computer/mobile.
https://github.com/cadars/john-doe gives me the same feels.
I also use google/facebook takeouts, reformat the results, and store+index all my human-facing correspondence in there. Text is cheap and I avoid most images. Its still under 200mb and instantly searchable with a nice UI and as a bunch of markdown files it is easily portable.
You're really going to drop these three words without any context?
Zettelkasten is a methodology of organizing a LOT of notes.
I index by topic, date and people involved. I can look up a friend and re-read every shared IM, email, and event I logged almost instantly. Faster than any website can. It's my own personal pile of papers future historians will be excited to find because they can actually read it.
One of my biggest frustrations is that most of my note-taking tools are not permitted in my workplace for security reasons. I have to keep all my notes on their infrastructure. I'm going to loose a chunk of my brain when I change jobs someday.
Make sure you setup basic version control in syncthing, I had some issues with my daily notes getting clobbered because they were autogenerated by multiple obsidian instances.
I think more general tooling to "convert your assorted takeouts into a local database" is higher on my todo list. I have a bunch of python scripts I cobbled together to convert things. If we can get it all into an easy to use database, everybody could do their own things with them more easily.
The key feature it adds compared to your own setup is mapping subfolders to subdomains (+ dynamic websites, but you don't seem interested in that).
ex: ~/smallweb/example => https://example.localhost
We have a little discord community at https://discord.smallweb.run if anyone is interested.
will list your directory tree as a html file..helpful?
Seems like an ideal use for a very simple DIY static-site generator. Write it in Bash or Perl and it will be future-proofed forever.
The only tricky part has been dealing with git on iOS. I have to use a particular app (Working Copy) and some shortcuts to get the syncing behavior consistent. But it is doable!
Nowadays along with html i try to archive using MHTML format instead of manually embedding
Run a simple http server and start browsing archives
FOR IMAGES I DO IS
---> Store all images in Folder
---> Open localhost server
---> Open folder in browser
---> Using javascript convert links to <img> tag with src=link
--> Once browser fetches and displays all images Save as and i have embedded MHTML archive
Or simple bash script can be used to create html with img tag and links to folder
Or you can manuaaly template a MHTML
BUT i let my browser do the heavy work why go manual,
Also instead of BASE64 EMBED, EMDEDDING DIRECTLY BINARY IMAGES IN MHTML IS QUITE MORE EFFECTIVE AND LESS MEMORY CONSUMING
Eg i have 15 images MHTML (binary encode) -> 4MB MHTML (BASE64 ENCODE) -> 5MB
Another method i use is, Run python -m http.server on any folder
Or linux : tree -H http://localhost:8000 Set recursion depth
Then open folder link from server or tree created HTML IN BROWSER
in cmd execute wget -rkpN -e robots=off http://localhost:8000
It will recreate folder with index.html for you to browse, you dont need server then for viewing
Same as export from google or twitter or youtube
I also wrote a script to convert Kindle highlights into Markdown files. If anyone’s interested, I'd be happy to polish it a bit and share.
For public-facing content, the Static Site Generator ecosystem keeps improving. I started with Jekyll (since it's the GitHub default), moved through Gridsome, and eventually landed on Nuxt 3 Content, which feels like the sweet spot for me. If I were starting now, I might have chosen Astro.
In any case, the barrier to entry has never been lower. We can host sites for free on GitHub, and if custom styling is needed, AI models are incredibly helpful with CSS.
Markdown is like JavaScript for text formatting. Despite its quirks, it just works.
Ah, that is quite sad. Do you write general impressions of the work day, at least, when you get home? I guess none of us remember all of the details of our workdays anyway.
[1] https://github.com/IAmStoxe/obsidian-markdownr
[2] https://addons.mozilla.org/en-US/firefox/addon/markdownload/ - Theres also a chrome extension
is author/entrepreneur Derek Sivers' script for reproducing his bare-bones, low-overhead, long-term "Tech Independence" stack.
Regarding Facebook, I do it manually only for selected posts. I tried to do it otherwise, but Facebook exports don't have data about likes (it would be helpful to filter popular content) or comments (often more important than the original post itself).
Ever since WordPerfect I've preferred more deterministic, lightly-formatted documents with some way to see formatting characters directly. Markdown is brilliant, basically a DSL (domain-specific language) for HTML.
The key to plain text is tooling! A couple Markdown tools I haven't seen mentioned here yet (even though they've come up on HN before) are:
https://addons.mozilla.org/en-US/firefox/addon/markdown-view... - pretty-render Markdown right in the browser
https://casual-effects.com/markdeep/ - standalone web-friendly Markdown formatter with many features
Not considered stable just yet, but it works well-enough for me.
You could add a "Download this page" button on the page which does some tricks to produce an HTML file with the images baked in that the user can download, which could work on mobile.
Here's a quick prototype: https://gistpreview.github.io/?14a2c3ef508839f26377707dbf5dd... - code here: https://gist.github.com/simonw/14a2c3ef508839f26377707dbf5dd...
Since I most of the time like to store articles, tutorial or nifty tricks, I like to store the entire website. For this task, my favorite Tool is SingleFile[1]. With SingleFile you can save a Website with embedded images. Also, you can add annotations, and cut away annoying Ads etc. Besides, it supports a distraction free copy of the website. I can highly recommend taking a look.
I find symlinks work for this, which is what I do. I have big directories with the raw pictures dumped from my devices, then categorized directories linking to them.
> “It’s been so long since I had to [use a holo-map], it didn’t even occur to me. It’s like an eidetic chip you can hold in your hand. It even remembers things you never knew before. Wonderful!” He unfastened his jacket, and pulled a second device from an inner pocket, a perfectly ordinary, though obviously best-quality, business audionote filer. “She gave me this, too. It cross-references everything automatically by key word. Crude, but perfectly adequate for ordinary use. It’s nearly a prosthetic memory, Miles.”
> The man hadn’t had to even think about taking notes for the past thirty-five years, after all. What was he going to discover next, fire? Writing? Agriculture? “All you have to remember is where you put it down.”
> “I’m thinking of chaining it to my belt. Or possibly around my neck.”
-- Memory (1996) by Lois McMaster Bujold
That last "audionote filer" is looking increasingly practical in real-life, cross-referencing and all.
It can definitely be used in such a local website, giving the convenience of just writing plain markdown.
=> https://www.tducret.com/pure-markdown/
(source code) https://github.com/tducret/pure-markdown
I like that zim is not automatically a hosted solution but a local app. I would love to see more local apps for archiving solutions and PKM. I just have some issues with the Zim app itself. It works nice for some of my use cases, but not for all. And I wish it would just use markdown (I know it has limits). Stuff like that.
I think Zim does not really fit into the discussion because it does not rely on easily exchangeable standard software like a file explorer and browsers.
That said, I believe that notetaking applications mostly exist because file explorers do an extremely bad job and integrating applications with them is too limited or at least too reunified. Look at what these applications offer. 80% of it is actually the task of a file explorer.
I've read somewhere that Telegram exports work this way, you get a bunch of raw files somehow organised with directories and browsable by themselves, with a tiny local static website to browse them more conveniently.
So different from the last such mass export I used: Google Takeout, which produces a dumb dump of cryptic xml and raw files named in some nonsensical (to the user) scheme. To this day I'm not even sure I got all the data I asked for before deleting it cloudside.
Zim is very much based on the file system. Each note is a text file and if it has attachments or embedded images they go into a folder named after that text note.
Whilst not in markdown the markup used is easy to understand and convert. Zim itself allows you to copy a note to the clipboard in pandoc markdown and export the note to markdown and/or html (though admittedly the styling for html it atrocious).
I am not against zim or obsidian. In fact, I currently use plain markdown wit vs code, which boils down to a similar situation. But vs code and its extensions may be gone in a while and then I will have to look what to do.
Google/Search tends to send quite a lot of people looking for how to take notes with Plain Text, and they seem to have benefited from my simple write-up.
Dumbeldore touched a wand to his head, pulled out a thought and put it in his pesisive.
The book 'Memory' is optional reading for a year long class in development. I still get things out of it.