I Wish …
Maybe Someone Will Grant Them
Another reason to actually have a web site is simply a place to publicly proclaim your wishes for yourself and the world. I suppose a wish is not as weighty as a goal or a todo list. A wish is something that you just want, not that you necessarily need to do yourself. In fact, usually quite the opposite. You wish that someone else would do it instead of forcing you to. Inevitably, I end up granting my own wishes rather than complaining or sitting around waiting on someone else.
Coder Coffee Shop Desktop App
I’d love to just be able to pull up an app on a secondary screen and see everyone’s screens while they are coding. The app could rotate through random people and give them a 4x size for a bit just to highlight what different people are working on. Those screens that don’t change could auto-minimize themselves so only people doing stuff would be on the screen. Then we could add like a Discord chat room with it all. The one solid requirement would be that it not crash constantly.
Even something that is just terminal based would be amazing, perhaps something like what TMUX or Screen shared session do right now.
Reformed Educational Systems Focused on Learning
Most public education systems (and most private ones as well) are not focused on learning. They are focused on meeting quotas, perception management, making money, pushing people through, and glorified child-care. The most motivated teachers in such a system simply cannot accomplish true learning in their students. When learning happens it comes at the cost of dependency on the teacher and system instead of producing true autodidactic individuals ready to face the world.
Teachers who have gone through the formal education to become one either love or hate that when I say this. I’m convinced the science surrounding education is seriously flawed and focused on the wrong outcomes. Ken Robinson agrees. This is why I want nothing to do with traditional academia and teacher certification. It turns well-intentioned teachers into drones with their aspirations burned out by an objectively failing system with upside-down priorities.
The path to reform means significantly changing the culture of education. It means turning teachers into mentors and learning facilitators, not subject-matter experts who have actual experience. It means obliterating ego and promoting empathy in the learning environment. It means having SMEs and facilitators working together. The SMEs produce the content, challenges, and activities. The facilitators organize learning lab environments where the learner is in charge of their own learning. Responsibility for learning is placed on the learner from the very beginning rather than creating an antagonistic teacher versus student approach we have today. We need the courage to work toward motivational skills and let those unwilling to learn simply not learn, to hold them back for as long as it takes until they decide on their own terms to begin and own their learning instead of trying to force them to learn.
These techniques have been proven successful over and over again in environments where the traditional systems gets completely out of the way, from the groups of “lost learners” who are essentially thrown into detention during school all day, to the “gifted and talented” deemed to smart for the current system. Both of those systems flourish because the teacher is not a teacher, but a motivator, a facilitator, and yes, even a responsible friend.
All of the plights of 2020 can be traced back to a lack of good learning. Those who learn and love to learn discover empathy and facilitators working in this way must be above all experts in emotional intelligence and empathy, councilors more than teachers. Most people who remember their “best” teacher will recall all of these qualities in that person who operated this way within a seriously broken system.
Better PEG Grammar Notation and Tooling
Update: I’m actually making this if you want to help.
PEG (Parser Expression Grammar) is by far the most important advance in language grammar development in the last 20 years (even though many CS departments still don’t even talk about it in their curriculum). The ecosystem around PEG is fractured and full of over-engineered code generators that are implementation language specific. What we need is a standardized notation based on Bryan Ford’s original (but incomplete) “example” in his paper that started it all. Then we need to show that tooling can exist — including code generators — that simply use the grammar and don’t intermix PEG with additional syntax to render in a specific language. That way the same grammar files can be used to generate parsers of different kinds in different languages. The same grammar file could be used to generate a highly optimized parser within a single file or another that uses very flexible AST approach and parse function library.
PEG is particularly useful in computer science education not only to create grammars, but to document existing ones and get the mind to understand what syntax is as well as any specific syntax. Capturing an existing language in PEG is a great educational exercise.
Fast, One-Pass Parsable Markdown (Ezmark)
Markdown is amazing, but not even the CommonMark initiative has recognized the value in creating a version of Markdown that can be parsed in one pass instead of two. This is because of reference links and other stuff in blocks at the bottom of a Markdown document that must be parsed before anything above can be rendered that depends on them.
All the advances in Markdown — mostly in Pandoc — have been toward become more flexible with adding extensions and providing more and more additions to the syntax. This is entirely against the original goal of a standard, simple language that is better than HTML than anyone can learn in 20 minutes. We need a Markdown that enough people could use without a lot of training not just because it is simple, but because HTML needs to be replaced with some version of Markdown.
A Decentralized Knowledge Network
This is one I have been working on for several years, something to replace or supplement the world wide web that maintains the motivations of the original web: knowledge exchange. It will leverage the infrastructure of the web without depending on it. It will provide localized content searching through semantic structuring, parsing, and local caching. It won’t even need a network. It will provide digitally-signed attribution of authorship. It will use trust to improve content discovery.
Development is moving along slowly here.
Bash Shell Rewritten Under Better License
Seriously, Bash is so bloated and the code base so old and crusty that the whole thing needs a serious rewrite under a license that will allow it to ship on embedded hardware and such instead of Bash because of its completely stupid GPLv3 license. (No Zsh is not the answer.)
Kick-Ass Educational Chemistry 3D Shooter
I’ve long wanted someone to take me up on the idea of creating a really rated M educational chemistry shooter that does smell like a “for education” game. It could be a full triple-A game that requires you to prepare different ammo and weapons to fight an wildly different set of monsters and mobs using the right stuff. There could be a chem lab where you see the periodic table in 3D, play around with simulated combinations of all the chemicals to learn about their properties and more.
Hell just a realistic chem lab that taught about what happens when you combine different chemicals would be amazing.
Full YAML 1.2 and JQ Parser Command
jq command is absolutely phenomenal. It just only exists for JSON. There needs to be a
yq as well that is good — not that Python PyYAML shit. It also needs to be written in Go, C, or Rust and maybe have a lot of the unnecessary query stuff removed. It could be installed as both
jq and detect how it is called in order to process data differently. Fuck the old
jq which never plans to support more than the “JSON internal structures”.
Such a tool would be so monumentally useful it might even put a serious dent in the use of databases in general. Something that supported that level of power, including the full remote files idea, would completely dominate the flat-file data management space.