03:05
librasteve_ left
04:03
arkiuat left
04:15
arkiuat joined
04:19
arkiuat left
04:33
arkiuat joined
04:37
arkiuat left
05:06
arkiuat joined
05:11
arkiuat left
05:41
arkiuat joined
05:48
arkiuat left
06:01
arkiuat joined
06:07
arkiuat left
06:20
arkiuat joined
06:25
arkiuat left
06:44
arkiuat joined
06:50
arkiuat left
07:19
arkiuat joined
07:24
arkiuat left
07:45
arkiuat joined
07:54
arkiuat left
07:57
human_blip left
08:24
arkiuat joined
08:29
arkiuat left
08:47
arkiuat joined
08:52
arkiuat left
09:13
arkiuat joined
09:18
arkiuat left
09:32
arkiuat joined
09:37
arkiuat left
10:00
arkiuat joined
10:05
arkiuat left
10:34
arkiuat joined
10:38
arkiuat left
11:05
arkiuat joined
11:09
arkiuat left
11:37
arkiuat joined
11:41
arkiuat left
11:59
human-blip joined
12:06
arkiuat joined
12:11
arkiuat left
12:41
arkiuat joined
12:46
arkiuat left
13:07
arkiuat joined
13:12
arkiuat left
13:29
wildwestrom joined
|
|||
wildwestrom | I hope you'll forgive me sort of coming out of left field with something non-raku related. imgur.com/a/wm9KIPb | 13:35 | |
I was interested in structural editors and other visual programming ideas. Basically, instead trying to lex and parse text when we could directly edit language semantics and represent the source code however we please. | |||
What do y'all, members of the raku (and maybe by extension perl) community think of these ideas? I'm interested in your takes. | |||
lizmat | librasteve might have some ideas about that | 13:36 | |
13:41
arkiuat joined
13:48
camelia left
|
|||
wildwestrom | Nice! I'll look at the logs for librasteve if I end up logging out. | 13:52 | |
13:56
arkiuat left
14:03
lizmat left
14:04
lizmat joined
|
|||
disbot | <antononcube> @wildwestrom It is not clear what you were asking about. Please expand or clarify. | 14:13 | |
wildwestrom | I'm just asking about people's opinions on structural editors (and less so visual programming in general). | ||
14:17
arkiuat joined
|
|||
disbot | <antononcube> Well, anything goes if it is productized is enough. LLMs are big equalizers of the field. Quite a lot of approaches become much more accessible and feasible. | 14:20 | |
<antononcube> for example, it is easy with LLMs to generate flow charts from code; and the other way around: to generate code from flow charts. | 14:21 | ||
<antononcube> Hence, I personally, wouldn't spend time learning some visual programming application, how it works, and know how to manipulate the mouse clicks in order to produce code. | 14:22 | ||
<antononcube> But some people do like this kind of stuff. For example, the so-called agentic-AIs always get some visual interface associated with them. | 14:23 | ||
<jubilatious1_98524> @wildwestrom it's an interesting idea. Visual programming has seen it's heyday, but what about Grammar trees? That would be exciting. FYI the Raku Discord has an #other-programming Forum where Raku/Perl people chat about other languages/approaches. | |||
wildwestrom | Where's this discord bridge? I normally hang there in the Rust server. | 14:24 | |
If voice is active that'd be wonderful. | |||
14:25
arkiuat left
14:35
stanrifkin joined
14:52
arkiuat joined
14:57
arkiuat left
15:03
stanrifkin left
15:17
arkiuat joined
15:23
arkiuat left
15:24
arkiuat joined
15:40
stanrifkin joined
|
|||
disbot | <librasteve> wildweststrom: welcome to the fun house | 15:40 | |
<librasteve> here's how to write that in in raku my @fib = 0, 1, *+* ... *; | 15:41 | ||
<librasteve> m: [0, 1, *+* ... *][9].say | |||
<Raku eval> 34 | |||
<librasteve> discord.gg/VzYpdQ6 | 15:43 | ||
<librasteve> ^^ the url above is a Discord invite | |||
<librasteve> sorry - couldn't help showing off the raku fib code - not really relevant to your question ... tbh I am not sure about visual coding - personally I prefer to actually code with text ... I am slowly following the path to LLM and that seems to me to make visual tools less interesting since they are not in the LLM coding loop as easily | 15:46 | ||
<antononcube> @librasteve > [...] following the path to LLM and that seems to me to make visual tools less interesting since they are not in the LLM coding loop as easily Until recently, say, a year ago, that was true. But with the so-called agentic-AIs visual programming somewhat resurrected. | 15:51 | ||
<antononcube> See here, for example: n8n.io | 15:54 | ||
<antononcube> cdn.discordapp.com/attachments/768...22a1c& | 15:55 | ||
<librasteve> .oO | 16:53 | ||
16:57
wildwestrom left
|
|||
disbot | <wildwestrom> That's a valid point, but then again there's weirdos (somewhat like me) who will make their own particular emacs config and maintain it for 15+ years. My point is, you could still have a keyboard-focused interface, but the logic is not tied to a concrete string of bytes. | 17:01 | |
<wildwestrom> That's a valid point, but then again there's weirdos (somewhat like me) who will make their own particular emacs config and maintain it for 15+ years. My point is, you could still have a keyboard-focused interface, but the logic is not tied to a concrete string of bytes. | 17:02 | ||
<wildwestrom> noice :) | |||
<antononcube> but the logic is not tied to a concrete string of bytes. completely agree. | 17:03 | ||
<antononcube> But for me this means using, say, natural language DSLs, not visual programming. | 17:04 | ||
<wildwestrom> Tools like n8n could be an interesting way to gradually go from fuzzy logic to something more well-defined. | 17:05 | ||
<antononcube> Yes, this is how it's supposed to be used. (Same with Alteryx DataIKU, etc.) | 17:06 | ||
19:26
sjn joined
21:24
arkiuat left
21:35
arkiuat joined
21:40
arkiuat left
21:48
librasteve_ joined
21:54
arkiuat joined
23:52
lizmat_ joined
23:55
lizmat left,
librasteve_ left
|