My wife was skeptical at first, but the ability to add labels, search, etc made it really easy to work together and accomplish the tasked we needed in time for the wedding.
The hardest part was creating a bookmark that links directly to the issue tracker.
Oh, I’ve also used GitHub issues to organize all the boxes in my most recent move. I would create an issue and the description would list all the contents of a box. Then I would write the issue number on the box. After moving, I could search GitHub to find that one thing I was looking for and know what box it was in.
Anyway, a fun solution but I think it’s more effort than I would have been willing to put in even if I would have appreciated the outcome.
Turns out one kitchen box got placed at the bottom of in a pile of book boxes in the living room.
If you unpack in a day, it's no big deal, but if you spent a week unpacking, you may find yourself having to eat something other than spaghetti for lunch, which is normally fine, but not when you really want spaghetti and the lack of spaghetti merely makes you more determined to find it.
In addition, I’m numbering the boxes, and when packing them keep a list mapping the numbers to what’s in each box. So when later searching for something, I know it should be in box number x. This can be helpful even years later when you don’t unpack all boxes.
Indeed, this is one of the biggest reasons I tracked this information to begin with.
I think you’re supposed to unpack 80% on day one, and keep the rest boxed up for the next move?
Look at you! I still have boxes packed from my move a decade ago. :-)
It only makes sense if you plan on unpacking over a year but if you unpack everything in a couple days then the system is not as useful.
I've used LibraryThing for book boxes. Using smallish boxes (30-40 paperbacks each) so that carrying them is not a backbreaker. Scan the ISBN barcodes with phone app, fix old ones/whatever on web app, tag with box number written on at least two sides. No problems found so far.
This is a fun anecdote to share, but everywhere you can find people with absurd workflows that are better dealt with using proper tools. FWIW I used Org mode to organize a move to another country. I really cannot stand the idea of feeding my personal information to Microsoft.
But at my current job Gitlab could easily take over Youtrack, already took over Upsource.
{
viewer {
issueComments {
totalCount
}
issues {
totalCount
}
}
}
I got 9,413 issues and 39,087 comments, for a grand total of 48,500 combined!
A lot of companies pay GitHub a lot of money to look after their source code and related artifacts. That's GitHub's business model. I don't think they would jeopardize that trust for the sake of training a model on private data.
I can only assume that companies paying for GitHub also pay for enhanced levels of privacy. Just because a company can pay GitHub not to train on their data, doesn’t mean they’re not going to train on your data that is being hosted for free. They are almost certainly crawling all free repos.
privacy needs to be verifiable (Apple has show this is possible with private cloud compute)
Maybe the non-training only applies if you pay protection money? But then you run into the whole if it's public there's nothing stopping some other AI that isn't MS from accessing the repository and training on it.
Related discussion:
How is it excellent when current logs could do with a bit of redesign doesn't find the comment (requires quotes to find it)
And then a tiny typo "current logs could" do with a bit of redesing also fails you
Their code search supports regular expressions now, which I find pretty incredible. https://github.blog/engineering/architecture-optimization/th...
language:Go path:/flake.nix
Can you tell me where I've gone wrong?
This seems to work;
path:/flake.nix go
Searching for "go" as a substring turns up a lot of false positives, but at least now I know the search is not flat out broken.
Others have mentioned this but if you want to keep this workflow, the best app I've found is Obsidian + Git Plugin. It works fantastically well on desktop though it does require a little work to get it working on iOS.
Heck, maybe I should just use code for notes. One plus would be web access with code server, since Obsidians only docker image that I know of uses VNC.
Anyone compared these two tools and have a decent write up? The biggest item which comes to mind would be referencing other notes and the features built on top of that?
Yes...always visible in the status bar. Fyi: https://github.com/Vinzent03/obsidian-git
If I want to keep this workflow why shouldn't I just continue using this same workflow?
"Automatic bidirectional sync between all devices" scares me. How does it deal with merge conflicts? How am I sure I'll be able to revert to a previous version? Can I see the full history of a file? I don't know, perhaps it'd be ok. I certainly wouldn't learn git just for note taking! But, I know how to use my hammer, so everything look like a nail...
I also store in a onedrive folder for automatic sync and backup in case I have a crash before I do a git commit.
https://github.com/cybersemics/em/
"em is a beautiful, minimalistic note-taking app for personal sensemaking."
I'm not trying this app.
The readme is more about the technical details of the code than the actual features of the app. Where do I go to see what this thing can actually do?
Do people expect me to run the program just to see if I want to run the program?
I know this is common with projects that think Github is a replacement for a website, but I genuinely wonder how does it get so bad that a 5 year project with 9000 commits and 60 contributors doesn't have a single screenshot. Nothing personal or particular about this project specifically, just... the whole open source culture of dropping something on Github and not even doing the bare minimum to have other people get to know the project.
It feels like such a waste. It could be an amazing project but who is going to bother with it if they can't see what it looks like?
I like how future proof a folder of markdown files is. But I like the design, simplicity, and deep features for capture and media support offered in Apple Notes.
The more a markdown app supports extra stuff, the more proprietary it starts to feel, as any app to read it will also need to support those things.
A while back I told myself I was going to stick to Apple Notes, as going back and forth to other things is painful, and doing it proactively means more pain, rather than maybe having some pain in 10 years of the app goes away. However, where I am again, in the middle of a largely manual migration back to Obsidian for my folder of markdown files. I used an export, but the formatting is so bad that I need to clean up every single note.
Basic formatting is a plus for me. Although now notes has really advanced formatting and even sketching.
I just tried it and it still works:
uvx apple-notes-to-sqlite /tmp/notes.db
# in another terminal while that's running
uvx datasette /tmp/notes.dbBut trouble free sync between machines, the ability to ‘scan’ documents, adding basic maths support, the ease… it always sucks me back in.
I wish it kept the date of creation and edits readily available, and supported markdown. But it’s damn close to what I want.
Why can’t Apple Mail do search as well as notes?
Or neovim with FzfLua (on laptop)
This makes every app that saves into iCloud files behave like Notes, i.e. work offline with automatic online sync.
* The ability to write an issue summary separate from comments. When you are debugging some hairy bug, some manager doesn't really want to wade through all the comments to get an idea; an editable summary at the very top of the page communicates high-level points to stakeholders while others continue to comment on details. People work around this by editing the initial comment of the issue but it's better if there's something more dedicated.
* Sophisticated access control. More than once when someone writes a bug report they are referring to a bug experienced by a single user. For user privacy reasons there needs to be a per-issue permission system to restrict access beyond the permission implied by the repo.
* The ability to add personal notes to an issue without publishing it. Whether it's a draft form of a comment or something else, it gets rid of the need to maintain your own notes.
It wouldn’t be too hard to add that with a 3rd party plugin. You could set an event hook to run through the comments, and add it to the top comment.
For bonus points, you could use an LLM to summarise. Every company loves a bit of AI these days, so your manager can gloat with his manager buddies that you now have AI-powered issue tracking.
2. That sounds like an anti-pattern to me. There shouldn’t be PII in your dev issue tracking system.
There problem here isn’t RBAC, it’s the workflow. If you’re in a situation where you need to make notes of sensitive information then you should store that in the same data store as the source information (eg Salesforce, et al). And I say this as someone who hates Salesforce.
3. I’ve not seen this feature in any issue tracker. Sounds like a nice feature but I wouldn’t have thought it as essential.
Then how do you track such issues? I'll give you a real example I've experienced: a high-value customer writes to support and complains that his UI is broken. None of the other people's UIs are broken. Do you not use the issue tracking system just because you need to get that customer's private settings and PII in order to debug the bug?
It's common to require PII to be stored elsewhere, but people will still make mistakes and copy paste PII for convenience. In the end isn't it better for your issue tracking system to be flexible enough to store PII?
> It's common to require PII to be stored elsewhere, but people will still make mistakes and copy paste PII for convenience.
That’s a training problem.
And It’s also common for people to fuck up RBAC. The latter is a harder problem to fix with training than teaching them to keep PII out of issue tracking systems.
> In the end isn't it better for your issue tracking system to be flexible enough to store PII?
I’ve worked in some very sensitive domains. They managed just fine keeping customer data out of the issue tracking systems.
If training is such a problem, I don't understand why we solve the problem by obviating such training and make the issue tracking system an acceptable storage medium for PII?
I would hope this will soon be written by AI automatically; summing up high-level points in an issue discussion seems like a perfect task for an LLM.
Here's an example where I condensed 50+ comments (accumulated over a year and a half!) into a summary comment and added it to the thread: https://github.com/simonw/sqlite-chronicle/issues/7#issuecom...
I created the summary like this:
llm \
-f issue:https://github.com/simonw/sqlite-chronicle/issues/7 \
-m openai/o4-mini \
-o reasoning_effort high \
-s 'Output markdown explaining this issue and the proposed resolution succinctly and clearly'The saving grace for now seems to be that Microsoft sees the "Advanced" PM tools in GitHub as value adds and is either rolling them out slowly to charge "Preview Customers" for as long as possible for that for that feature before making it public and/or putting them into GitHub pricing tiers such as "If you have to ask how much it costs, you can't afford it" and "Offer you cannot afford to refuse".
I don't want to risk my notes on a configuration error or billing mishap.
For example, stuff like this could happen: https://news.ycombinator.com/item?id=23832437
Wouldn't hosting my own private notes on there be against the spirit of what they're trying to achieve?
I have no problem at all freeloading off GitHub!
But ~ 2 years ago I switched to Obsidian for that
Eventually I event started used Obsidian for my project management, and ditched GitHub Issues / GitHub Projects
With caveats, that I use that for the greenfield project with lots of unstructured exploration + AI agents for keeping design docs and figuring out detailed tasks
For the established and legacy projects, I would probably use GitHub Issues for bugs, enhancement requests. And GitHub Projects for all reactive work (support, ops, bugs, etc.)
Lastly, I disagree that it's "almost the best notebook in the world". It my might be a best ticketing system or a note taking system, but not a notebook in the sense of Jupyter or LiveBook (but nothing stopping them to make code blocks executable[1], and even add some LLMs).
Also it's easy overwrite the content of the issue, even by a single person working from different tabs (at least that was the case in the past).
---
1. GitHub Blocks
Last week, Microsoft cut off the email account of the chief prosecutor of the international criminal court, after being ordered to do so by the Trump administration. This is merely a week or two after Microsoft did a big PR campaign, telling the European Union how they were to stand up against this administration if the need arises. The irony!
Trump isn't hiding his opinion about the EU. What's stopping him from ordering Microsoft to block the EU from accessing github one day? Right now: Nothing at all. Think about how it would impact your business and open source software in general. We must not put all our eggs in a single basket.
I’m responding more to the comment that it’s extra steps. It’s clearly not as good a solution as obsidian, but there are no extra steps for it being accessible over the web or to all your devices - not just those for the devices built in cloud provider.
I think GitHub/gitlab issues is totally viable. Obsidian/Logseq too.
Or just use GitHub/gitlab issues.
(I wouldn’t say Obsidian is perfect, but so far I haven’t found any note-taking application that I like more.)
Then for unlimited collaborators in 2020: https://docs.github.com/en/get-started/learning-about-github...
Granted, you do have to install it versus using the default editor that comes with your operating system, so that is an extra step. But most people don't use the builtin editors from their OS anyway.
I'm thinking, host it locally and just let one drive act as back up for the mark down files it produces?
I knew about using just git, but having Windows, Linux, macOS, iOS and Android, it was just way easier to use this solution instead of fighting with git-like apps for each OS.
* It doesn't extract the title from the issue on bare linkes, instead the url will become something like: <group>/<repo>#15 Which isn't as nice. So issue #15 in repo tootie/mynotes would look like: tootie/mynotes#15
* It also doesn't do offline sync.
I think that's the only thing it doesn't do. I use Git Touch on iOS and it seems decent enough, you can get to issues and update them, etc.
The bonus is, assuming you run your own Forgejo, or trust whoever runs it for you(say Codeberg), you don't have the MS privacy concerns.
Next up:
- "SQLite is almost the best notebook in the world"
- "Claude is almost the best notebook in the world"
- "An SQLite database containing only Markdown files in a Git repo self-hosted on an SD card in my Raspberry Pi served by a Node.js web app accessible via 56k modem is almost the best notebook in the world"
- "I created a startup to take the previous thing, reinvent GitHub Issues on top of it, call it AI Notes, and make almost the best notebook in the world, and it's now valued at $50B"If you have discipline to do a monorepo for the entire organization, your issue and code tabs are effectively the entire universe in one place. This is the only thing I've seen that can pull middle managers out of email - a single bucket that has everything of concern in it.
You can put it online so you can get to it anywhere
You can run it off a stick to keep it with you (Tiddlywiki)
You can cross link pages to be able to collect things together
You can search. Some even have auto link builders so you can build a page of links to pages that match a category.
Some support Markup so you have formatting the way you want.
I haven’t run a wiki in years because markdown+git is good enough and simpler. I never really need anonymous edits.
But I get the OP, if you live in GitHub, it makes perfect sense.
Any local notebook, including pencil and paper, is better.
I've experimented with a bunch of tools for extracting copies of my issues and pulling them down locally, which is easy thanks to the GitHub API.
One that's particularly fun is a GitHub Actions setup that copies the full contents of each issue thread into a file in the repo itself. Then I can "git pull" to fetch my backup!
I still don't have any of them running on a cron at home but I might set that up some day.
There will always be better apps, but you need to stick to an app with its flaws and limitations.
I have been through this rabbit hole... (paid for Bear, IAWriter, noteplan, FSnotes and probably more that I forget) and have just stuck to local markdown files and use VS code ... I even used LLMs to generate a tiny plugin to solve my micro needs.
Extremely doubtful. The fact you haven't yet hit limits does not mean no limits.
Nor does the fact the provider claims unlimited.
I take a lot of notes but I don't think I'm going to end up with the largest issue collection on GitHub.
I could always spin up a second repo or org if I do run out. I have my notes scattered across several hundred repos (and a hunch of orgs) already.
Yet still, one more could hit the limit.
Step 2. ...?
I don't mean to be facetious here. It's just that if GitHub doesn't end up going through the usual enshittification process, it'll be quite the anomaly.
Now, maybe it'll be an anomaly, exceptions to rules exist. But it just seems to be an endless cycle, trusting these companies, getting burnt, trusting again, getting burnt again. Is there no end to the cycle? Is there some new pressure on these companies to behave better that was not there before, and therefore we should all give them (yet) another chance?
I don't see why we would, personally. It even seems obvious to me that we shouldn't. Like in an abusive relationship, no matter the regret and the promises, eventually you've to say enough is enough, and ideally sooner rather than later. Nonetheless, it's very common practice for people who appear at least superficially like smart people to trust these companies a second, third, fourth time.
Either these apparently smart people know something I don't, or they like getting paid large amounts of money so much that they'll stay quiet and use the popular services and simply use their tech savvy to jump to a new thing quickly when the shit hits the fan. Or maybe I just prefer being cynical, and it's a personality thing. Someone, educate me here.
I wouldn't trust any system for my notes that doesn't have this.
- I can transform it to other formats should I need to
- The GitHub variant of it (GFM) has feature like syntax highlighting for different languages which is incredibly useful to me based on the kinds of notes I take
- Diffing clearly genuinely is a benefit, sometimes I want to know exactly what I changed!
- I can parse it with regular expressions, so useful information doesn't end up locked in some weird binary format
- LLMs are great at reading and writing it
These things may not matter to people who are not nerds. I'm a nerd!