←back to thread

412 points conanxin | 4 comments | | HN request time: 0s | source
Show context
kkfx ◴[] No.41085082[source]
To be (re)read together with the Unix Haters Handbook https://web.mit.edu/~simsong/www/ugh.pdf to realize that what we need is re-made LispM, Smalltalk workstations or the OS as a single application, framework opened down to the most low level part, in the user hands, fully end-user programmable, discoverable, and fully integrated. A 2D and even 3D/4D CLI as the UI, witch is the DocUIs with eventual 3D and video/multimedia elements.

As a conceptual framework http://augmentingcognition.com/assets/Kay1977.pdf

replies(1): >>41085282 #
lproven ◴[] No.41085282[source]
Riotous applause from the cheap seats
replies(1): >>41085635 #
kkfx ◴[] No.41085635[source]
Thanks :-) but allow me to enlarge, in my poor English: "us" the westerns, we came from those who invented the IT, the modern industry. We are evidently in a declining phase. Evidently someone else try to emerge and actually one of the most prominent new power, China, emerge doing what we have done when we was at the top of the world in industrial, public research, education terms. Meaning such "technical model" works.

Evidently the financial capitalism have worked for a certain period but does not work anymore. So, why keeping committing suicide? We have started to suicide with the WWI. We kept going with WWII and we continue now.

We are still the leader for IT, and we know what does it work, the classic FLOSS/at least open IT model, the one where some sell iron not bits, the one where customers own their systems and bend them as they wish, the one where communication is not owned by some walled gardens but open like on Usenet, classic mails (as opposite to webmails, hiding the decentralized ability for most users who do not own the WebMUA). To continue with the China comparison I've recently bought some battery tools, very cheap crap but enough for domestic usage and I've observed that batteries have a standard connector, I can swap them from different brands issueless, batteries, chargers are "standard". I also own some high end domestic battery tools, who happen to have a damn nfc tag inside the battery tying it to the device, even if inside the battery are classic connected li-ion batteries. The same I observed for BEV, some friends and I have three Chinese BEV from different brands/models and they have a gazillion of common parts. So to say "open/standard pay back" yes, it might erode SOME OEMs margins, but pay back the society at a whole and as a result the OEM itself. The same is valid in software terms. My desktop is Emacs/EXWM, I use as a search&narrow framework consult/orderless/vertico, they simply "plug in" any other packages because the system is a unique application end-user programmable at runtime. You do not need to "specifically support" a package to integrate it. You do not need third party to explicitly support your package to use it together. And what we do instead? Our best to create walled gardens, we have had XMPP, SIP/RTP and now most are on Zoom/Teams/Meet/Slack/Skype/* all walled gardens. Even many push to substitute emails with some new colorful walled garden. Similarly any app try to add features someone else have since it's impossible just using it, like a sexp in a buffer.

As a result modern IT from Unix where the user at least can combine simple tools with some IPCs in script we are limited by cut&paste, drag&drop and not much more. Anything is damn complicated because there are no shared data structure in a shared environment, but any "app" do everything on it's own, often with a gazillion of dependencies who have a gazillion of deps on their own, with a big load of parsers of any kind and even "some standard to pass data" (like JSON) who try to emerge but are still not a shared data structure in an unique integrated environment.

All of this is the old Greenspun's tenth rule and actually is killing our IT innovation, killing one of the last sector we still rules, for the same issues that have killed our industry essentially.

replies(1): >>41085830 #
lproven ◴[] No.41085830[source]
I 100% agree, and your points about Chinese tools are particularly incisive.

As an aside, but I think relevant and you might find it interesting:

A decade or so I discovered Oberon, the last masterwork of the great genius of programming languages Niklaus "Bucky" Wirth. A complete OS, UI and compiler, in about four and a half thousand lines of code.

I have it running in various forms.

I introduced it to the Squeak Smalltalk community, and when I told them what I was looking for:

« a relatively mature, high-performance, small OS underneath, delivering some degree of portability -- something written in a type-safe, memory-managed language, with multiprocessor support and networking and so on already present, so that these things do not have to be implemented in the higher-level system. »

That is how I found Oberon. They told me such a thing did not and could not exist.

I gave them some links and people were amazed.

It seems deeply obscure in, as you say, the West.

But I have found an active community working on it and developing it. It is in Russia.

It may be that in other countries now the target of Western sanctions, we may inadvertently be fostering some very interesting tech developments...

replies(2): >>41086659 #>>41087628 #
1. kkfx ◴[] No.41086659{3}[source]
I know Oberon only by name, I've used (and honestly not loved it at all) some Pascal dialect back at high school, but back then was not real programming and was an introductory very bad organized course so it's hard to tell, I've encountered probably Oberon for a river navigation applications around 10 years ago but I wasn't really involved so I can't say much, I essentially do not know anything but the name, if you have some interesting links to share I'll skim them with pleasure.

In more broad terms I do not put much attention in a specific programming language even if clearly an OS-single-application is tied to a specif programming language, in the sense that there are many, and they are many factions loving one and hating others, the point is offering something usable at user level, like "just type a sexp and execute it" also from an email, because integration means also an immense small diversity and heavy dialogues so innovation. With such model we can keep our supremacy and more important we can't lose it because essentially anything float in a common see.

The main issue to reach such goal I think it's general cultural of the masses, today most peoples, many programmers included, do think that IT means computers, like saying that astronomy is the science of telescopes. Of course computers are like pen and paper, an essential tool, but they are a tool, the purpose of IT is information and that's not a specific technical task but a broad aspects involving essentially all disciplines. Until this is clear for anyone there is little hope people understand the need, power and issues of IT, they'll keep looking at the finger pointing the Moon instead of at the Moon.

The rest came after, even the most basic computer skills came after because to learn we need to be motivated, learning "just because you have to" as a social rule is not productive.

replies(1): >>41099569 #
2. lproven ◴[] No.41099569[source]
> if you have some interesting links to share I'll skim them with pleasure.

I do not advise skimming.

I've been a full-time tech journalist for 2 & a half years now (I was in the 1990s as well but the 21st century is very different) and I find the majority of readers who angrily disagree with my articles did not in fact understand the article because they tried to skim it and they didn't get the gist.

(In a previous job I was a TESOL/TEFL English teacher. "Skimming for gist" is a skill we test for, and many people don't have it and don't know they don't have it. I an not accusing you here -- but you did mention your own English in negative terms.

For example, I was on a talk at FOSDEM in February -- https://fosdem.sojourner.rocks/2024/event/3113 -- and it seemed to me that most of the audience angrily arguing about what the GPL meant and implied had not really genuinely read and understood all 6 pages of the GPL.)

Executive summary of Oberon:

https://ignorethecode.net/blog/2009/04/22/oberon/

13 page academic assessment, but very readable and accessible:

"Oberon – The Overlooked Jewel" https://dcreager.net/remarkable/Franz2000.pdf

replies(1): >>41102172 #
3. kkfx ◴[] No.41102172[source]
> I do not advise skimming

That's very right, but modern life is complicated so accurately study something demand much time, quickly see the concepts might helps and well, the concept of textual-UI is definitively not alien to me, since my desktop is EXWM, with almost all my digital life in org-mode, org-roam-managed notes, it's still very different than Oberon (or Plan 9) desktop but the textual concept and org-mode links that can execute sexps on click (a feature I use much, for instance to link specific mail/threads in notes and create interactive presentations) it's similar. The 2D "spaced" desktop concept It's something I see in the far past, SUN Looking Glass LG3D concept desktop https://en.wikipedia.org/wiki/Project_Looking_Glass and yes while the above help they still can't tell me what's inside the package, meaning what's behind the UI concept and the language grammar. I still miss the architecture.

However I suppose for what I've seen so far that essentially it's not really usable in real life so it's a nice to know project but stop here, like Lisp M Genera or Plan 9. Emacs at least can be used for real today. It's sad the IT industry have pushed what I call the glorification of ignorance, but more than preserving knowledge for a more civilized world I think we can't do.

So far in the last decades most of the old valid ideas get anyway accepted, for instance widget based UIs have essentially failed and are more and more substituted by WebUIs witch are read-only DocUIs or NotebookUIs witch are limited 2D CLIs, something close to a DocUI. They are mostly text-based as well. So well, maybe in 10+ years we will finally have something like an Oberon or LispM desktop, surely returned with many anti-users aspects but still offering something of the past glory, and there memories will help to keep correcting the aim and reducing the loss. Anyway until people realize the substantial importance and role of IT there is little hope for a more civilized era...

replies(1): >>41102977 #
4. lproven ◴[] No.41102977{3}[source]
I really do feel your pain. ;-)

No, it's not viable as a general-purpose OS these days. At one time it was and was deployed to non-technical staff inside ETH.

The last development in the line, not from Wirth himself, has a zooming GUI, resizable overlapping windows, SMP, a TCP/IP stack, an email client and a very basic HTTP only web browser. It is closer than you might expect.

I believe the core OS is on the order of 8000 LOC.

You may enjoy my FOSDEM talks if you're interested in this kind of thing.

I did one involving rebooting the local OS stack based on Oberon and Smalltalk, or maybe Newspeak:

https://archive.fosdem.org/2021/schedule/event/new_type_of_c...

I turned it into an article recently:

https://www.theregister.com/2024/02/26/starting_over_rebooti...

And this year a more Linux centric one based around 9front:

https://fosdem.org/2024/schedule/event/fosdem-2024-3095-one-...

That became an article series:

https://www.theregister.com/Tag/One%20Way%20Forward/