Open Source Summit, days 2-4

More niches! Some convergence.

The days passed quickly!

Because being precise in the midst of such a blur is impossible, I'll relay that I had some really good discussions! Only some of them concerned linux or open source.

Somehow, my Hyprland desktop came up and someone mentioned that the lead coder has some... people issues, as so many of us do.

Because I missed the day 1 description of Cosmic DE, I changed a couple lines in my Nix configs and after a long, conference-wifi download I was up and running. The keybindings make basic sense, and I'm making progress as usual.

Somehow System76 has no presence in the sponsor's playground, so I spent a piece of an hour speaking to Framework! They had their new touchscreen on display, alongside their desktop and a sleek rack-mount 4x desktop lab. The 16" input modules are some peak engineering; I had an engineer run lsusb as he plugged and unplugged them. I'm super impressed, needless to say.

Incidentally, the Framework lads had been amused to pull up my old Jira ticket - on Jan 4, 2022 I had been the first order shipping to Washington, DC; I asked someone to add the district to their e-commerce form.

In the middle of the conference I paused to message one of my collaborator groups on Slack:

Hi, clicking in from the Linux Open Source Summit in Denver; A couple people are discussing distributed (dispersed) data lakes (record collections). These include:

• https://lfedge.org/projects/edgelake/
• https://gravitino.apache.org/

these options are groundbreaking ways to solve a problem we may have someday: connecting different databases and sources.

Our lab-location problem is more properly called "High Availability" or "Disaster Recovery". The usual approach is to combine "Replication" and "Failover". This normally requires reading long incomprehensible explanations:

• https://www.enterprisedb.com/postgres-tutorials/postgresql-replication-and-automatic-failover-tutorial

We can be unusual though!

This open-source DB is designed to be multi-machine, and a drop-in replacement for postgres. The other apps like Airflow or Tableau should have no idea the underlying program is changing.

• https://www.yugabyte.com/
• https://github.com/yugabyte/yugabyte-db

And speaking of Slack, I am a new member of some buzzing new groups: OpenSearch and CHAOSS.

Now, in the 4th-day mini-summit, I'm on the special CHAOSS track. After a couple days of high-IQ discussions, I found some refuge in discussions of docs, inclusion, onboarding, and the more human sides of open source management.

CHAOSS is the main group appraising open-source health on a global basis. They produce education programs and open-source apps to help us explore the science of coding, using the public records of authorship inherent to open-source.

Today is a deep examination of how to measure open-source community.

Phase one in the process is to mine the rich resource collections: you can use GrimoireLab or Augur. Augur is the source behind 8knot, a dashboard core to CHAOSS.

One of the speakers mentioned having authored a blog post for opensource.com, which seems to be experiencing many Drupal and AJAX errors!

I'm in the middle of an employment application as a Drupal manager, so I may seek to help on their app if possible. Oddly enough, I see no source-code links among their pages.

These authors mentioned the Cardano blockchain as a peer-reviewed blockchain geared towards paying open-source coders. Yay! These speakers also authored Mozilla & the Rebel Alliance and similar reports examining community relationships.

I'm marginally sad to have missed a discussion on Ecosyste.ms, and eager to explore this domain more as soon as possible. I had absconded from the CHAOSS track for that talk, to go see how to apply Guardrails for AI apps.

Hopping back to Tuesday as I examine my open pages, I had been super pleased to see the author of Wireshark explain StratoShark, an analogous program he has made to explore system calls.

The key idea behind this program is to be essentially a clone of WireShark; the .pcap capture file for packets uses the same encoding as the .scap files for system calls. The audience asked around the possibility of building an app, perhaps a web UI, to display both captures in chronological order.

"Sure", he said. "In the base of the repo there is a sharkd daemon, and someone has had success communicating to that process from a WebUI before." [paraphrased]

Now, I'm sure more is going to emerge from this conference in the days to come; I'll spend many more public-library sessions in my email inbox and blog, perhaps uploading images from my camera.

Which reminds me - I had a chance to ask a Sony patent engineer about my Alpha DLSR; we discussed how nearly all serious camera companies are made in Japan, and we double-checked where [Hombu Dojo] is (Shinjuku).

I need to spend much more time learning Japanese, and besides, this may be the primary appeal of the Open Source Summit; the highly international audience in attendence.

I'm making connections, and am eager build a much broader dialogue around these ideas in the days and months ahead.