Almost done with revision of my book. Will be publishing tomorrow and then take the rest of the week off from all things related to writing, editing, reviewing, etc.
Hope to read a few fantasy books and watch some nice anime shows. Heard good things about Tengen Toppa Gurren Lagan, so I’ll check that out first.
dropgit sounds like a really interesting project. I’ve been thinking a lot lately about how documentation is done at companies and I’m beginning to believe that effective company-wide (or project-wide) documentation would be helped by making it easy for technical and non-technical people to collaborate in the same version control system. Would dropgit allow a developer to continue to use the CLI flows they’re familiar with while a customer support person uses a friendlier interface?
I started to talk with some developers about the problems in documentation and prototype a solution. I thought about relying a lot in git, mainly because I like the idea of having documentation near the code as comments as javadocs or jsdocs, but it also enables other features: most developers I talked said the code is the ultimate true and they can’t trust the documentation because it gets out of sync, so imagine a tool that shows you the last commit date for a code block and the last commit date for that code block documentation. You would be able to easily see if the documentation is outdated and check what were the changes since the last time they were synced.
I’ve often wished contracts in docx form and other legal documents could be in git. Something like dropgit enables that, and can encourage people to use alternate formats like richtext for easier diffing…
Thanks! The idea is definitely to allow CLI and non-CLI people to work together. How that’s going to look exactly isn’t settled. I have some ideas but we will need to experiment and see what works. At this stage I am trying to see if I can gather enough interest in it to be able to put a serious amount of work towards it.
There is a strong chance for self-hosting and it being completely open source. It’s very early stages for this idea but it’s grown out of working on kitspace.org and gitbuilding.io and part of the motivation is to be able to write software that can be re-used for these open source projects. I wrote a few more details on the GOSH forum if you are interested.
Oh and about the commit messages: the idea is that you can change them if you wish. It will give you some time before they become “permanent”. So you can do it later or you can do it as you go. Dropgit is still very early though, mostly ideas at this point, I think I need to tweak the landing page to make that clearer.
I definitely need to make this more obvious on the page, but this is a UI concept which is at the idea stage. I have some software written for kitspace.org that I can re-use for this but Dropgit as presented doesn’t exist yet.
My current thinking around the commit messages involves three different scenarios:
This is a shared repository, you are committing directly to master.
Dropgit acts as your staging area that “settles” an hour after you make your last commit, then it syncs to the remote. You can only edit your commit messages before it’s “settled”.
This is a shared repository, you are committing to your own branch.
Synced right away to the branch, Dropgit will push --force-with-lease amendments you make to the history of this branch.
This is not a shared repository, this is your repo and you are working on master.
Synced right away and push --force-with-lease to master
Whether we really want to support all three scenarios and how to keep the interaction and interface simple enough bears some further design and experimentation.
Work: More documentation work. I just finished a mini-editor to help me track which tables/columns in the database still need describing in our docs.
Personal: I just got https://idea.junglecoder.com and https://feed.junglecoder.com ported from Nim to Elixir/Phoenix. Now the next project for learning how to work with Elixir/Phoenix is building site to act as a game arcade for the various Godot projects I plan on building over time. I’d like to make it look more stylish than my typically programmer-looking projects, so that will be interesting, but first comes the skeleton, and figuring out how to customize the Godot HTML export well.
I’ve read this book more than a decade ago and appreciated it. However with new life experiences, I’m compelled to reach for it again, and this I’m I would understand it differently this time.
Aside from that, continue working on the Cerveau project (zettelkasten related).
Work: Working on a few essays, and writing my own take on a zettelkasten-like app. Brain extension software seems like one of those things you gotta write yourself. Situated software and all that.
As for the zettle… I think part of why it works so well is you see random cards while hunting for the tag you want. I think that aspect could be valuable.
That’s one reason I’m writing it for myself: I can implement whatever features I think work best for me. I’m storing everything in sqlite so I can do think like reverse link lookups and hierarchical tagging, plus it’s easier to back up a single data file than someone else’s app.
Interesting. I’ve got one laid out but I’m finishing some other projects other people want too before I’ll be able to justify the time spent on this. Mine is a Clojure app backed by rocksdb with Crux on top.
I set up a Minecraft server and finished my course on Scala programming this past weekend. Will probably be playing a lot of Minecraft.
I think I’ll return to reading The Little Typer. I got to around Chapter 9 before going on an indefinite hiatus due to Life Stuff(tm), and that book gives me the same feeling that I had while reading SICP. It’s insane.
Work: duct taping things together, working on going through the headache that is open sourcing things. Trying to open-source something I wrote in Rust.
Personal: Worldbuilding, Dungeonbuilding and working on a loosely ended plot for the Big Bad Evil Guy for my campaign. Also writing up some ideas for enemies and other classes for that system. I kinda want to do something with a witch class that has a more literal “magic has a cost” type of system. The witch would need to sacrifice materials and maybe cut their hand and give blood for spells. Still working it out.
Work: Trying to finish up the last things before launch so that we’re not gesturing at things in the dark. It’s been a long stretch and not quite done yet.
Life:
Reading more about Buddhism,
finishing up Doom: Eternal,
working more on my RSS reader so I can finally ditch Feedly (Need to move it over to a SQLite solution as every clever trick I’ve had doesn’t really work or is just more work)
Starting up a reading habit again after being knocked out of it by Life.
Over the weekend, I made much more progress than expected on integrating filesystem extended attributes support in FreeBSD’s package manager. I’m hoping to finish that up this week.
At ${DAYJOB}, I’m adding more value-added features to our proprietary fork/build of OPNsense. I’m hoping to start building the public-facing infrastructure so that we can push out updates to customers.
I’m still learning programming. I started with Go and i’m liking it so far.
First time seeing slices and maps and i’m getting ultra confused, but am slowly starting to grasp. I understand it, but i get all confused when writing the for range loops for those.
I’m learning from udemy, but i’ll probably go after some text course after finishing this one.
Trying to figure out exactly which distro to run on my Talos II ppc64le server. I was surprised to learn that Gentoo has fewer packages ready for this architecture than Debian, despite being source-based.
Debian works quite well on ppc64, except for qemu user-mode emulation for running amd64 binaries. The Debian kernel is using 64k pages out of the box, whereas qemu needs 4k pages for parity with amd64. I struggled a bit with compiling a custom Debian kernel before switching to Gentoo, which actually uses 4k pages by default and didn’t require configuration on my part.
Due to package scarcity on Gentoo I might switch again, this time to Ubuntu. I generally prefer Debian, but 20.04 looks like a very interesting release: 5.4 kernel, LXD, and ZSys are all desirable to me.
I would love to have a 100% ZFS-based system, but the built-in bootloader petitboot is not able to understand ZFS partitions. Perhaps I could set up a kexec-based chainloader using Alpine.
It’s perf review time and due to reasons (joining midway through a cycle, having my two previous teams dissolved, general product misadventures, and the background radiation of 2020) I’m just steeling myself to go through a process that I know is going to be anxious and depressing. Yay.
I’ve had two development plans thrown out, and we’ll see how the third one goes.
I’ve been getting to have some more fun with Postgres than I’d intended. Row-level security is very neat.
¬Work:
I’ve got a whole slate of new games to chew through, if I find the time and werewithal. After those, I’ve got the massive itch bundle.
I’m looking to move out sometime in July, but need to find a place to move to. It’s a pain in the ass.
I’ve got a bunch of books on the way to pick up. Finished “Kill All Normies”, which was a thing, and am deciding which thing to read next.
Rebuilding a community platform with Nixops.
Slowly pulling out of various communities and backchannels…it’s all so tiresome. Part of it is definitely the year getting to me, but a lot of it is just gradually coming to the realization that I’m not getting the value I thought I was from it. Time spent arguing is better spent creating, and the argument to discourse ratio has been broken for a while now.
At work, I’m continuing to work on building a variant of Bottlerocket for Amazon ECS (and writing bad Rust, as I am very much a beginner in the language).
Outside work, I’m starting to play Animal Crossing.
Upgrading & scaling our Prometheus LTS solution (using M3), fixing a couple of bugs in our Nomad logging pipeline that have been bothering me for a week now.
!$work:
Improving my edge worker for collecting some stats about my personal blog & the blog from a friend/colleage (much more popular than mine). This has shown that I need to fix a couple of things regarding the data/labels that we collect.
I’ve opensourced a small golang proxy that takes & verifies the JWT token from Cloudflare Access and automatically signs in the user into my personal Grafana instance (grafana running with auth.proxy enabled). This removes the need of having to sign into 2 places (Cloudflare Acess & Grafana) it is a very minimal, but quite handy for my experiments.
Hopefully writing a small blog post about a couple of ideas about how we’re using SLOs in our logging pipeline (this might be work & !work, let’s see how it goes).
I’m busy rewriting documentation, planning videos, blogging, and filing bug reports for my new gig as Open Source Evangelist at Jina AI. Hoping to travel soon to meet the rest of the team. I’ve been stuck in a small country for too long and my feet are itching to get out.
Other than that, hoping to catch up on The Twilight Zone and Salt. Fat. Acid. Heat to further my cooking projects.
Project A: Aaalmost done with the implementation, which includes what is, essentially, an ASN.1 parser, so the testing will be fun…
Project B: Squeezing a menu into a system with 2 KB of RAM. But I got a whole Flash bank just for that so it’s doable, and I’m actually looking forward to it. It’s a selection menu that allows me to toggle various options. It was on the todo list anyway, but for various reasons, the options were implemented before the menu, and activating each one of them required changing a compile flag, which made testing pretty unpleasant.
Non-work: I’m juggling two projects so non-work during weekdays has been canceled until December, at which point I am looking forward to sleeping outrageously long hours during an outrageously long vacation. (Fair warning to any younger readers: this isn’t healthy to do in the long run. 6 months of it, if you have a good reason, is one thing, but unlimited 10-14 hour workdays are bad for your health and anyone trying to convince you that’s just how success happen is swindling you). I just got a Windows machine so probably half an hour of Doom Eternal? I keep at least 1-2 hours in the evening free, to spend them with the wife, so if you add up the hours, there’s not much time for anything else.
Almost done with revision of my book. Will be publishing tomorrow and then take the rest of the week off from all things related to writing, editing, reviewing, etc.
Hope to read a few fantasy books and watch some nice anime shows. Heard good things about Tengen Toppa Gurren Lagan, so I’ll check that out first.
Going to do a few things:
spywareWindows anymore!All over the place, as I am most weeks:
dropgit sounds like a really interesting project. I’ve been thinking a lot lately about how documentation is done at companies and I’m beginning to believe that effective company-wide (or project-wide) documentation would be helped by making it easy for technical and non-technical people to collaborate in the same version control system. Would dropgit allow a developer to continue to use the CLI flows they’re familiar with while a customer support person uses a friendlier interface?
I started to talk with some developers about the problems in documentation and prototype a solution. I thought about relying a lot in git, mainly because I like the idea of having documentation near the code as comments as javadocs or jsdocs, but it also enables other features: most developers I talked said the code is the ultimate true and they can’t trust the documentation because it gets out of sync, so imagine a tool that shows you the last commit date for a code block and the last commit date for that code block documentation. You would be able to easily see if the documentation is outdated and check what were the changes since the last time they were synced.
I’ve often wished contracts in docx form and other legal documents could be in git. Something like dropgit enables that, and can encourage people to use alternate formats like richtext for easier diffing…
Thanks! The idea is definitely to allow CLI and non-CLI people to work together. How that’s going to look exactly isn’t settled. I have some ideas but we will need to experiment and see what works. At this stage I am trying to see if I can gather enough interest in it to be able to put a serious amount of work towards it.
My only wish for dropgit would be better commit messages… But it’s a really really cool project and idea! Signed up.
Any chance of self hosting? I don’t put many repos on github/lab so I’d need to be able to point the back end to my own hosting…
There is a strong chance for self-hosting and it being completely open source. It’s very early stages for this idea but it’s grown out of working on kitspace.org and gitbuilding.io and part of the motivation is to be able to write software that can be re-used for these open source projects. I wrote a few more details on the GOSH forum if you are interested.
Oh and about the commit messages: the idea is that you can change them if you wish. It will give you some time before they become “permanent”. So you can do it later or you can do it as you go. Dropgit is still very early though, mostly ideas at this point, I think I need to tweak the landing page to make that clearer.
How is it that you support adding commit messages later? Do you re-write history on the remote?
I definitely need to make this more obvious on the page, but this is a UI concept which is at the idea stage. I have some software written for kitspace.org that I can re-use for this but Dropgit as presented doesn’t exist yet.
My current thinking around the commit messages involves three different scenarios:
push --force-with-leaseamendments you make to the history of this branch.push --force-with-leaseto masterWhether we really want to support all three scenarios and how to keep the interaction and interface simple enough bears some further design and experimentation.
Work: More documentation work. I just finished a mini-editor to help me track which tables/columns in the database still need describing in our docs.
Personal: I just got https://idea.junglecoder.com and https://feed.junglecoder.com ported from Nim to Elixir/Phoenix. Now the next project for learning how to work with Elixir/Phoenix is building site to act as a game arcade for the various Godot projects I plan on building over time. I’d like to make it look more stylish than my typically programmer-looking projects, so that will be interesting, but first comes the skeleton, and figuring out how to customize the Godot HTML export well.
Read Status Anxiety by Alain de Botton.
I’ve read this book more than a decade ago and appreciated it. However with new life experiences, I’m compelled to reach for it again, and this I’m I would understand it differently this time.
Aside from that, continue working on the Cerveau project (zettelkasten related).
Work: Working on a few essays, and writing my own take on a zettelkasten-like app. Brain extension software seems like one of those things you gotta write yourself. Situated software and all that.
¬Work: Apartment hunting.
As for the zettle… I think part of why it works so well is you see random cards while hunting for the tag you want. I think that aspect could be valuable.
That’s one reason I’m writing it for myself: I can implement whatever features I think work best for me. I’m storing everything in sqlite so I can do think like reverse link lookups and hierarchical tagging, plus it’s easier to back up a single data file than someone else’s app.
Interesting. I’ve got one laid out but I’m finishing some other projects other people want too before I’ll be able to justify the time spent on this. Mine is a Clojure app backed by rocksdb with Crux on top.
Work: Crunch time, probably.
!Work: Playing with my nephew, hopefully.
I set up a Minecraft server and finished my course on Scala programming this past weekend. Will probably be playing a lot of Minecraft.
I think I’ll return to reading The Little Typer. I got to around Chapter 9 before going on an indefinite hiatus due to Life Stuff(tm), and that book gives me the same feeling that I had while reading SICP. It’s insane.
I recently setup a vanilla Minecraft server for my family. We’ve been having a lot of fun on there!
It’s running in a screen session at the moment with cron-jobbed world backups.
Work: duct taping things together, working on going through the headache that is open sourcing things. Trying to open-source something I wrote in Rust.
Personal: Worldbuilding, Dungeonbuilding and working on a loosely ended plot for the Big Bad Evil Guy for my campaign. Also writing up some ideas for enemies and other classes for that system. I kinda want to do something with a witch class that has a more literal “magic has a cost” type of system. The witch would need to sacrifice materials and maybe cut their hand and give blood for spells. Still working it out.
Work: Trying to finish up the last things before launch so that we’re not gesturing at things in the dark. It’s been a long stretch and not quite done yet.
Life:
Personal:
Work:
I’ve been doing some freelance writing for the past coupe of weeks. It’s been fun but I’m 99% done and it’s time to go back to working on PLUMA.
Over the weekend, I made much more progress than expected on integrating filesystem extended attributes support in FreeBSD’s package manager. I’m hoping to finish that up this week.
At ${DAYJOB}, I’m adding more value-added features to our proprietary fork/build of OPNsense. I’m hoping to start building the public-facing infrastructure so that we can push out updates to customers.
@work More release testing.
@home Over the weekend I wrote this package for using the RideWithGPS API from the Lisp REPL. This week I’d like to build something on top of it.
And I’m finally getting around to reading the boxed set of “His Dark Materials” that I received as a Christmas gift.
I’m still learning programming. I started with Go and i’m liking it so far.
First time seeing slices and maps and i’m getting ultra confused, but am slowly starting to grasp. I understand it, but i get all confused when writing the for range loops for those.
I’m learning from udemy, but i’ll probably go after some text course after finishing this one.
Trying to figure out exactly which distro to run on my Talos II ppc64le server. I was surprised to learn that Gentoo has fewer packages ready for this architecture than Debian, despite being source-based.
Consider sysbench. It’s completely missing on Gentoo (not even marked
testing!) because luajit is not available. However Debian has it available in buster-backports. Luajit doesn’t officially support ppc64 yet, so I’m curious how Debian is pulling this off exactly. Perhaps they’re packaging the ppc64 port as-is.Debian works quite well on ppc64, except for qemu user-mode emulation for running amd64 binaries. The Debian kernel is using 64k pages out of the box, whereas qemu needs 4k pages for parity with amd64. I struggled a bit with compiling a custom Debian kernel before switching to Gentoo, which actually uses 4k pages by default and didn’t require configuration on my part.
Due to package scarcity on Gentoo I might switch again, this time to Ubuntu. I generally prefer Debian, but 20.04 looks like a very interesting release: 5.4 kernel, LXD, and ZSys are all desirable to me.
I would love to have a 100% ZFS-based system, but the built-in bootloader petitboot is not able to understand ZFS partitions. Perhaps I could set up a kexec-based chainloader using Alpine.
Work:
¬Work:
At work, I’m continuing to work on building a variant of Bottlerocket for Amazon ECS (and writing bad Rust, as I am very much a beginner in the language).
Outside work, I’m starting to play Animal Crossing.
$work:
!$work:
auth.proxyenabled). This removes the need of having to sign into 2 places (Cloudflare Acess & Grafana) it is a very minimal, but quite handy for my experiments.I’m busy rewriting documentation, planning videos, blogging, and filing bug reports for my new gig as Open Source Evangelist at Jina AI. Hoping to travel soon to meet the rest of the team. I’ve been stuck in a small country for too long and my feet are itching to get out.
Other than that, hoping to catch up on The Twilight Zone and Salt. Fat. Acid. Heat to further my cooking projects.
Work:
Non-work: I’m juggling two projects so non-work during weekdays has been canceled until December, at which point I am looking forward to sleeping outrageously long hours during an outrageously long vacation. (Fair warning to any younger readers: this isn’t healthy to do in the long run. 6 months of it, if you have a good reason, is one thing, but unlimited 10-14 hour workdays are bad for your health and anyone trying to convince you that’s just how success happen is swindling you). I just got a Windows machine so probably half an hour of Doom Eternal? I keep at least 1-2 hours in the evening free, to spend them with the wife, so if you add up the hours, there’s not much time for anything else.
Work: Its the last week at my current position so I am cleaning up loose ends etc.
Personal: Looking at working on some WPF and .NET Core things because this is where the new focus is.