doing as much as I can.

gram:do is a new exercise in recording my daily progress. I'd like to make a dedicated rendering layer to place these on a calendar. Perhaps! Anything is possible if you ignore how much energy is required.

I've been focusing on keeping calm, as I acclimate to changing domestic scenario; going from pickup-nomadism (18 months) to proper camper-home nomadism (2 weeks). I do experience panic attacks, especially when becoming sleepy, which make sense in a pickup; sleep implies the approach of each day's freezing dark hours. Oddly, I also feel an uneasy sense of guilt from being able to sleep inside, warm and cozy, when so many of the surrounding neighbors are unable to secure a home.

I suppose the issue is that I spent the last 18 months, and in a larger sense the last 6 years, learning so much about nomadism and proper homelessness - including crisis management infrastructure designed more for their capacity to suck up peoples' money than to inflate their ego or (common) senses.

All this said, there really is no equal to a proper secure home. I slept until noon for the first time in ages, a couple days ago. After that I spent a few nights nocturnal, which made sense because I was hopping cars around Maryland, to manage registration and insurance concerns. I had some late-night bicycle and bus rides to follow some damp rainy nap days.

Today, I'm back to early-hours, in essence because I "phased around" by lagging through a sleep cycle. This also seems good; I like the quiet hours in the hackerspace when I can click record with no background machinery running.

This week, I'll need to spend those early morning hours preparing to speak at Iffy Books in Philadelphia on Thursday.

I do indeed hope this is all coming through in the daily progress log, and I'll be upgrading this to be more explorable, soon as possible.

---

Morning Recording

A couple days ago I had been able to record coding session, including my screen and accompanying audio. This morning, I spent a similar duration to record more progress on gram:fiscal.

Once I neared the end, I looked across to see that my audio recordery had been keyed to a "monitor" source rather than the USB mic. The audio dial was immobile, meaning no signal appears in the recording. I finished up, and loaded the two screen captures into kdenlive. Oddly, the video sources had the proper dimensions (I'd been recording a vertical screen) although the signal had also been blank; a dark screen.

There is one more option for producing the session recording; I had asciinema running to capture the console session directly.

Embedding an asciinema player in this domain is a little too much for my purposes, so I'll come back and drop that here if I find some spare hours. For now, there is too much to do so I'll simply run through a summary of the changes.

---

gram:fiscal changes

In my original session, I had produced a cache full of separated reports, by month. As I looked back through these this morning, I recognized that each page had a header, and possible tailing lines. I made up some line-based parsing logic to shed this baggage.

I looked ahead some, and found that the paging scheme differed from my expectations. Page numbers are described as, e.g. PAGE: 1(1), wherein the pages are scanning in both rows and columns, like vertical and horizontal-scrolling a spreadsheet, with the first row and the column header pinned (frozen) on screen.

This is a good realization, and a consequence of small page sizes in legacy ASCII reporting programs. Text-based tabular reporting is common across industry largely because these reports underlie processes so crucial, they were modernized... before the modern era.

Hard to imagine that governments were once the leading edge of procurement. Once they upgraded themselves into the mainframe era, no one realized how challenging an escape had become.

I'll spend one more morning on this, and then you'll be able to grab .csv or .xlsx directly from the cache, or run the nu code on your computer to reproduce the process locally, learning as you go.

For now, you can find the intermediate .noheader files in the cache (.tar.gz or .zip).

---

Employment on Horizon.

As much as I have done to learn the boundaries and rules of nomadism, I do feel many days as though I am simply running from place to place, unable to lend anyone use of my skills or experience, because I am simply too preoccupied with the demands of keeping up.

Recently, I've been subscribing to employment postings that are aligned with my background, so I am considering once again how to apply to the corporate sphere. In case I'm able to secure something soon, I'll be crunging numbers to see about refilling my 401(k), paying off my credit card, and perhaps paying back the camper that I begrudgingly (as in all aspects of our relationship) am assigned to by my mom and dad. Let's keep from dwelling on this too much.

---

Resume, in progress.

The new roles mean I need to become more serious about the application process, and I'm going to re-build my resume as needed. I have some ideas that can help make this unique and exceptional; in the past I've coded my resume by hand as a web page, including the option to render as a PDF.

On this occasion, I'm going to take that approach again, and as I re-phrase many sections to more align to my career goals, I'll seek to make the resume easier to compose and re-arrange on the fly. I'm going to make a codebase, using the new elixir nex library, that is able to compose a resume on-demand from pre-made pieces, or the basic "claims" I make regarding my career experiences.

---

More secure coding isolation.

Because I, like anyone else, am new to the nex package - or, more precisely, because the package is new to us, I'd like to minimize the risk of exposure to supply-chain attacks. These insiduous security holes come around when someone embeds a problem inside a crucial code package, thereby reaching coders on their local machines.

Because of several insecure practices I'm cleaning up on my local machine, I'm pausing the addition of any new programs until I clean up. Rather than run new code directly, I am using VM isolation for programs that I'm unsure of.

Some people are already seeing me explain that, sure - I can join you on Zoom, although I remain unable to patch my webcam through to the VM. Of course, if Zoom's closed-source app is not scary enough, a quick glance through their terms and conditions makes me quickly recommend Jitsi in any of those scenarios.

I'm rebuilding [gram:pool] as much as need be, to begin doing much of my local coding inside an isolated minimal VM. I'm already pleased with how easy the networking has been, and for web-apps this is a clear and sensible direction for sampling complex libraries, including niche and hypereffective graph databases, and keeping my base machine as secure as possible in the process.

Thinking ahead, these changes to pool are likely going to cause ripples, when I begin publishing code packages again. Any process I make up for building or testing a package inside a VM is going to be far simpler to automate on a dedicated build machine, because my dependency on the nix flake is going to include each necessary build recipe, regardless of any decisions for the build pipeline.

Sure sounds like a bunch is cooking here... all you can do is all you can do.