freenode/#lisp - IRC Chatlog
Search
8:51:34
beach
Duuqnd: Right, we don't really have a good editor for Common Lisp code. Nor a debugger, nor ...
8:51:45
mfiano2
There are plugins to do syntactic and semantic analysis server-side with deep learning in real time, but this is something different and exciting. Been keeping an eye on it for years.
8:52:20
no-defun-allowed
Sadly "good enough is the enemy of what is actually needed" is not as frequently uttered as "perfect is the enemy of good".
8:54:05
beach
Cymew: Think syntax highlighting based on semantics rather than regular-expression parsing.
8:54:59
beach
The dynamic nature of Common Lisp makes the task much harder than for static languages.
8:59:21
Cymew
Was there any functionality in the old editors like Zwei that would ne interesting to resurrect? I have only glanced at the code briefly?
9:00:23
beach
That sounds likely. On the other hand, computers are way more powerful now than was the case of the Lisp machine.
9:02:04
Duuqnd
Sometimes I wonder how the tech industry would look if Lisp Machines became the norm.
9:02:55
shka__
For instance, it would be possible to have structural search to find expressions in the code.
9:12:27
scymtym
beach: thanks. it was inspired by an emacs lisp rewriting system presented at a recent ELS
9:18:49
shka__
for extra spicy example, one can imagine function that will return function bodies that are similar to other function using SVR metric
9:19:29
shka__
it would be a computationally intensive task, but with current hardware it is possible
9:22:23
lieven
Kent Pitman's "what was lost" post is 20 years old and these functionalities haven't been regained yet
9:25:52
lieven
I was merely reacting to the view that it is hardware performance that is/was keeping us
9:31:30
beach
Part of the problem, of course, is that many people are convinced that our current tools are absolutely fantastic.
9:33:01
Cymew
I might add my former comments are from the perspective of a grumpy old man, knowing his limitations. ;)
9:34:06
lieven
and part of the problem is that the goalposts are continuously shifting. Pike hit it with his also 20 year old remark that systems software is irrelevant. In order to have a functional desktop experience you need so many huge standards implemented that it's not feasible to do them from scratch and using existing implementations ties you to the current state of the art
9:35:26
lieven
for example the remark here earlier that so few pdf viewers do a good job with forms. to do forms properly you need to have a full ecmascript implementation in your viewer.
9:37:56
Duuqnd
Putting garbage languages into document formats! Truly, we live in the best timeline.
9:38:44
lieven
since acrobat stopped developing acroread for linux/unix, there is no pdf viewer that supports the 1.7 version of the PDF spec for these platforms
9:40:55
beach
One year, my (admittedly small) family actually had to go buy a Windows computer to do her US taxes because of that. Now, they have a website for it.
9:41:49
lieven
yeah a friend of mine had to go look for a windows computer to use for a visum application in a pdf+forms format
9:41:57
jackdaniel
"PDF 1.7 (...) ISO 32000-1, includes some proprietary technologies defined only by adobe"
9:43:10
lieven
jackdaniel: it may not count as an open standard but it is needed for a common desktop experience. likewise video codecs, drm incluced, etc
9:43:15
jackdaniel
"PDF 2.0 (newest standard) does not inlcude any proprietary technologies as normative references", so we have at least that
9:44:23
lieven
yeah and the bureaucracy that just send you one of those things will be hugely cooperative if you complain
9:45:09
lieven
last year there was a big ruckus in my country because tax professionals had to have a microsoft account to get access to the new tax legislation on an official government site
9:45:21
Duuqnd
Another example of this sort of thing is governments providing land border maps in a proprietary AutoCAD format.
9:45:58
aeth
we should specify the border between #lisp and #lispcafe in a proprietary AutoCAD format.
11:21:29
scymtym
shka__: no, i just wanted to try and make something like the emacs lisp thing using our libraries (eclector, mcclim, etc.)
11:24:00
shka__
do you intend to work on this? can I help? I would want to learn clim but it is much easier when you are working with someone that knows stuff
11:28:18
scymtym
shka__: i mainly work on the underlying libraries. in this case eclector, the pattern matcher and the s-expression syntax library. regarding clim, i enjoy using it and recommend it, but it is not at the core of the prototype in the screenshot
14:52:56
jackdaniel
a new name for a macro style often found in the wild: dimwit(?) -- do I mean what I tell(?)
14:56:51
phoe
minion: memo for pfdietz: the comments at https://gist.github.com/phoe/335fecfdc195bddd47ab0928b0e62e52 are good and outline errors in my reasoning. We're back to the drawing board, except now I wonder if commenting out the ANSI-TESTs that test the LOOP FINALLY variable values wouldn't be the best option if their value is to be treated as undefined.
17:17:51
Bike
yesterday pfdietz and shinmera and some other people were talking about utf8 strings in lisp, so i tried implementing them https://github.com/Bike/utf8string
17:18:13
Bike
that is, they're lisp sequences stored as utf8 encoded strings, so they're more compact than the utf32 the implementation strings probably are
17:22:19
Shinmera
Bike: t-e-s offers a fallback implementation of the protocol on implementations that do not natively support it, which may or may not be better depending on your POV.
17:26:13
Shinmera
t-e-s hasn't been tested anywhere except SBCL either (both with and without fallback).
17:26:18
Bike
you also need code-char and char-code to work with codepoints, and for the implementation to have unicode characters to begin with
18:00:01
mfiano
Bike: Nice, I checked out that project yesterday when it was pushed and it's very cool. I also think you trumped sjl with swearing in comments :)