Notices by Daniel Kochmański (jackdaniel@functional.cafe)

@sznowicki @Watchdog_Polska @icd

Niedługo trzeba będzie mieć konto GMail żeby głosować :3

I know that people snark at HN for many things, but I'm more concerned that it becomes boring. Rare pearls between valuation, ai and opinionated snubby blog pieces. I remember that there was way more hacker than news there at some point and that was fun.

Look what I've found -- CLIM II tutorial from '93:

https://www.cs.cmu.edu/afs/cs/project/ai-repository/ai/lang/lisp/gui/clim/clim_2/luv93tut/slides.ps

#lisp #mcclim #retro

@weekend_editor According to @amoroso 22 years ago:

'Some of the CLIM tutorial's source files indeed include the comment
"Written by John Aspinall at Harlequin, Inc". Does he still work at
Xanalys? How can I contact him?'

https://cl-pdx.com/comp.lang.lisp/2003/sep/118165.html

@aartaka technically readtables are file-local variables, so nothing wrong with using a custom readtable to implement internals.

@aartaka I agree that shilling ones own preferred readtable is unnecessary; I don't buy into the argument that "another dependency layer to manager" is wrong -- the author decides what is the most convenient for them to work with -- this argument could be turned against any dependency, even these that are not introducing exotic syntax.

@solene @thinkberg it is not that openbsd is particularily friendly to common lisp runtimes. W xor X being one of silly limitations from the computation perspective - like having a garden of dried flowers

I've created a liberapay account as an alternative to patreon:

https://liberapay.com/jackdaniel/
https://www.patreon.com/c/jackdaniel_kochmanski

I'm creating Common Lisp Free Software, most involved with #ecl and #mcclim development, but also contributing to other projects when I see fit.

Please consider supporting me financially so my work is sustainable. Cheers!

#lisp #commonlisp

@helge

As a person living in the old FOSS world I feel obliged to answer:

Sure, I'm aware that the code is copied by AI slop companies, loosely compressed and resold with the final aim to eliminate a pesky middlemen between the manager and a working program.

That said I'm not willing to give up on principles because there are burglars in the town. That'd sound silly: "I gave up on writing because someone would certainly take it and regurgitate it. With a profit!". Or -- "well, since they are violating my rights anyway, I'll give up on asserting that it is wrong and leave my doors unlocked".

Licenses clearly matter to users and signal the author's intention (giving a peek at their motivation). Moreover there are pending lawsuits, and hopefully they will succeed, or the bubble will burst eventually.

Another thing is that non-regurgitated software, while never bug-free, will be more often than not free of nonsense.

@tiang common lisp and multiple implementations. Generally because it gives tools fit for any paradigm and let's you choose how you want to write the program, without imposing on you preferences of the language designers about what is the best for you (oop, functional, imperative etc). I call this quality being an 'unopinionated language'. #lisp

Today I've wake up to a couple of LLM-generated made up bug reports in McCLIM repository. What a waste of time.

Thank you slop companies for enabling the erosion of the collaboration platform known as the Internet.

@CodingItWrong float is just a wrong type to do that. You may ude language with rational numbers (like common #lisp), or use ratios provided by gmp.

1/10 + 2/10 = 3/10

LLMs are certainly a cool hack. Not cool enough though to justify gazillion of dollars wasted per cutoff training, burning the planet and putting knowledge workers and artists in obsolescence distress en masse.

@ramin_hal9001 @screwtape @slgr

I'm planning to put scheme as an optional runtime in ecl, if that succeeds then using directly McCLIM from scheme will be feasible. Just saying.

Hey!

If you consider supporting my work on #FOSS here's the most recent update regarding McCLIM. It contains a bunch of videos showing progress (I've shared all of them on this profile too):

https://www.patreon.com/posts/hello-patrons-123700109

And for those who already do support me -- thank you! :)

#lisp #mcclim

Phew, I had to backtrack to Dynamic Windows documentation to understand the operator specified in CLIM II (because the specification was botched).

#mcclim

@mapcar @nsrahmad @simon_brooke @praetor

Common Lisp certainly has numerous warts -- think that it is unavoidable at certain size. That said I think that it is a very practical language -- not in a sense that it is packed with features, but because these features are orthogonal and useful.

This and the fact, that the language does not impose opinions on how your program should be designed, is a quality rarely found on languages that know the one true way of programming.

@mdhughes @screwtape

I've looked into it and was surprised at first, because it works here (tm). Then I've tried the older version in Quicklisp and I could reproduce the issue!

In other words the issue was there, and it was fixed. I don't recall what was it and I think that at this point bisecting /when/ it was fixed doesn't make much sense.

A remedy: clone McCLIM from the branch master to ~/quicklisp/local-projects and try again. Happy hacking!

@screwtape @mdhughes

CLIM features graphical output records that are drawn at fixed positions, and records that are put on the text line.

If you want to adjust the cursor position, then (setf (stream-cursor-position stream) (values cx cy))

Note that you can draw on the text line when you use with-output-to-graphics, then the drawn graphics will be placed on the text line.

Tadaam! thread safe streams in #McCLIM. #lisp #wip