Clojurians Log v2
off-topic
We created a GitHub repo dedicated to brazilian who want to discuss, learn and know from other people 🙂 for more information check #clojure-brasil https://github.com/wandersoncferreira/clojure-brasil
"3. There is no step three"
Hi everyone!
This is not a formal announcement yet as we are currently open-sourcing our repos, revamping our website, and planning for alpha. However, I share it with you since we need a little boost 🙂
A few years ago, Mike Anderson (creator of core.matrix) started working on something that closely ressembles Clojure on this very efficient "blockchain". It aims to be inclusive, green and eco-friendly, seeing all the intolerable excesses this industry has sprouted.
A good chunk of our toolset is written in Clojure and it has brought us a decisive advantage. I am keen to share all that with you "soon" and get your feedback. Meanwhile, we would appreciate a little vote to push us forwards. We are a very small team of devoted people, with limited means, so any help matter. Here in the "Digital Assets" section, there is a demo video as well: https://t.co/wS9idAG1if?amp=1
And if you want to get a hang of it, Convex Lisp is very similar to Clojure and you can hack in the "Sandbox": https://convex.world/documentation/tutorial
Cheers!
Reminds me of the Stuart Halloway tweet about refactoring... https://copilot.github.com/
Copilot sounds horrible... what about licenses? What about code quality?
There's a huge discussion already about licences. Even on Github, they explicitly told that sometimes copilot copy-pasted the full GPL license on an empty file
(that, after saying that copilot will not copy, but derive, code. Typical Microsoft talk, to be honest, to make incomplete, false, or confusing statements)
what did he say?
I avoid twitter. Can you summarize?
Lol, summarize XD Repeat what he said would be better
https://twitter.com/stuarthalloway/status/1234261008560115712
https://twitter.com/stuarthalloway/status/1316030414746771457
that really is something that demands a bit of substantiation. to be specific, I mean the statement “refactoring tools do more harm than good”. anyone found it?
And that reminds me of Gene Kim quoting Rich Hickey describing some refactored code: > It looks like someone took a wedding cake and then smashed it against a wall https://www.cognitect.com/cognicast/159 at ~24:50
I can imagine in practice someone trying to program with this copilot by constantly editing their comments to get it to autocomplete what they want until they give up and use google to find a snippet to copy from stackoverflow.
:rolling_on_the_floor_laughing:
Someone on the internet summarized some of mine thoughts ;; I don't like QI comparison, this should be like Clojure experience or something like this
Macros can transform code, much in the sense that code can transform data. And data is simple to query and transform 🙂
You don't want to use macros most part of the time
It's better to not build another DSL
many developers will give you pushback on using meta programming... I think it is best to keep any DSL contained how a micro service approaches things... small and rebuildable.
versus 100k line code base of a DSL
meta programming can save you though... if trying to refactor a difficult code base
there is a very thin line between a rich data structure and meta programming / DSLs
build effective programs that does stuff instead of making things too generic?
Many of the most powerful libraries use macros. Also it is interesting that runtime generated (DSL) code is very common, just in the form of maps and vectors rather than quoted lists. It’s essentially meta programming in a slightly different shape. Honeysql, malli, datomic/datascript come to mind.