Now, my postscript stuff is still interesting.

Custom Graphviz Nodes

Tell you what!??? I've had a lot of fun.

Ever find old code you wrote and it is so ugly it is useless? My-o-my, I was just starting out and figuring Python was the way. Ugh... not sure I can even use this:

Data Flow Diagram GRa(F) generation

My local-first vs. offline-only vs online-available needs a flowchart. Of course, I need to create a knowledge graph, so I had to look up some old work of mine from almost six years ago. I probably have a hundred thousand files to search through, so sometimes the Wayback Machine is easiest.

Snail Mail Import

This one is growing on me. #NowPlaying #AnasaziKaha #CodingMusic

Internet Archive

Space in my inner mind : Anasazi Kaha : Free Download, Borrow, and Streaming : Internet Archive

Anybody else sing all three every time? (Stevie onward)

My insistence on monolithic things over ecosystems has created some bad design decisions. I need to step back and rethink my stance in light of my interest.

fetch messes up offline-first. I can hack it with a recent Electron binary release and a handful of JS or manual browser config (dangerous to suggest!), but it is still nice to be pure default browser. Today I'm figuring out if I can bypass the fetch based on a timestamp value in local storage without messing with the overall performance on initial load.

I've long been aware that intellectually timid/lazy folks whine about semantic triples, and yet they overload keys with semantic information. If you do the hard work of the knowledge graph first, you can eventually arrive at concise and powerful key-value pairs and get both worlds.

Don't interrupt the sorrow Darn right In flames our prophet witches Be polite A room full of glasses He says "Your notches, liberation doll" And he chains me with that serpent To that Ethiopian wall Anima rising Queen of Queens Wash my guilt of Eden Wash and balance me Anima rising Uprising in me tonight She's a vengeful little goddess With an ancient crown to fight Truth goes up in vapors The steeples lean Winds of change patriarchs Snug in your bible belt dreams God goes up the chimney Like childhood Santa Claus The good slaves love the good book A rebel loves a cause I'm leaving on the 1:15 You're darn right Since I was seventeen I've had no one over me He says "Anima rising So what Petrified wood process Tall timber down to rock" Don't interrupt the sorrow Darn right He says "We walked on the moon You be polite" Don't let up the sorrow Death and birth and death and birth He says "Bring that bottle kindly And I'll pad your purse I've got a head full of quandary And a mighty mighty thirst" Seventeen glasses Rhine wine Milk of the Madonna Clandestine He don't let up the sorrow He lies and he cheats It takes a heart like Mary's these days When your man gets weak © November 17, 1975; Crazy Crow Music --- Off of my favorite Joni album Hissing of Summer Lawns

You can see both the explosion and shift in culture around 2012. This was when the WWW showed its capability for collective intelligence. At the same time opportunists were figuring out how to capture and hijack the wealth for themselves. The story is not over. The tech is free; the power shift is a harder problem. The infrastructure for collaboration is harder still, but unacknowledged.

The more I think about it, the more a cultural change to facilitating adaption pathways, rather than being self-appointed center stage silver-bullet tech overlords of the biosphere, makes sense. Don't overthink nature. She will find a way. With this cultural shift, we might tag along.

My theory on heptad element placement is that I should sort the data I stick in the PNG so that the main text at the leaf nodes (markdown, html) is close to each other for versioning. So, I'll squash graph/subject/predicate into the index/key to sort on, with secondary sort on date for versioning as the object changes. (JSON inside PNG data segment)

An Explanation of the DEFLATE Algorithm

Finally, before I STFU, my biggest mistake was sharing as I learned. I'm still learning. I wasn't a developer when I started. I had to figure that out. I knew nothing about knowledge graphs. I only knew Gane and Sarson DFDs were uniquely powerful yet simple. The absolute best guide is tracing simple programs to see where I'm failing in my understanding. Modern software ecosystems and tools hide the truth, particularly within an ecosystem of tools meant to deal with a giant knowledge graph: WWW.

One small example for those that think "instance" is easy. The model, the ontology, forms a nice map. But, we think in instances *or* model. So, if you have a journal app, an entry is an instance of the model where you have relations (tags and subcategories). The node of the subcategory in the model has instances of a journal entry. Most want to make it into a taxonomy, which does kind of work, but it messes up cognition of the model vs. instance. I don't want to do RDF/OWL. I want a shortcut.

A huge mistake I've made is trying to explain the design of knowledge graphs up front to non-academic folks. The difference between attributes, classes, relations, and instances is something that I am finally grasping after 7 years of immersion. Further, I use a foreign atomic particle of knowledge graphs that is a morphed semantic triple, so not even academics understand what I'm doing. It was bonkers and futile to try. It must be a working app first, something that is recognizable and useful.

I am refactoring heptapp completely to take advantage of the PNG compression better by re-arranging the heptad elements, and making the app more personal, which makes it more relatable.

Hello everybody out there using minix - I'm doing a (free) operating system (just a hobby, won't be big and professional like gnu) for 386(486) AT clones. This has been brewing since april, and is starting to get ready. I'd like any feedback on things people like/dislike in minix, as my OS resembles it somewhat (same physical layout of the file-system (due to practical reasons) among other things). I've currently ported bash(1.08) and gcc(1.40), and things seem to work. ... Linus