Plan 9 / 9front - Netsurf (browser) porting effort

I would get more use out of 9front with a better browser. Netsurf might suffice.

http://ninetimes.cat-v.org/news/2020/04/16/0/

I have compiled a small TODO list for the port of Netsurf to Plan 9 from things that I can come to think of. I also made the list so that it would be easer to select a task for anyone that would find it fun to help out. Maybe you would like to play with something on the list? The support for various image formats could be one thing that is easy to do in parallel with other tasks. I am not expecting changes in 9front just to compile upstream netsurf without patches, neither do I believe that the Netsurf team would change their code base too much in order for it to compile out of the box in Plan 9. All comments are welcome, however. As Ori and Kyle mentioned they were willing to work on some things, I put their names in brackets for those tasks :-)

This is actually nice. I saw the NetSurf contrib a few days ago and I was wondering if it is not a joke.

Have you tried it yet?

Originally Posted by YesItsMe This is actually nice. I saw the NetSurf contrib a few days ago and I was wondering if it is not a joke. Have you tried it yet?

No, not yet. I did some experimenting with 9front a couple of years ago and liked it, but lack of a full-featured browser was the drawback. So I may revisit it now.

That's good news. This will probably benefit all of the plan9 based OS's as well. If I find time I'll have to see if I can get it going.

Freaking google

https://support.google.com/accounts/...h2p9ZDU8Qq-bxU

Good luck.

All the more reason to use an alternative browser... Google's malware is too bloated to run on it

I imagine people messing with such an obscure OS couldn't care less...

Originally Posted by rokytnji Freaking google https://support.google.com/accounts/...h2p9ZDU8Qq-bxU Good luck.

One of many reasons I don't use Google products.

Originally Posted by Mill J This will probably benefit all of the plan9 based OS's as well.

Given that most of them are not really being developed anymore (except 9front - where the NetSurf fork seems to originate - and probably Plan B), this is an easy task though.

re:Google and its supported browsers: Meh.

I hike around on this site in Dillo and Links2.

That was why my previous post was mentioned.

Originally Posted by YesItsMe This is actually nice. I saw the NetSurf contrib a few days ago and I was wondering if it is not a joke. Have you tried it yet?

The article you cite is very interesting. I agree with most of it.
https://news.ycombinator.com/item?id=10273656

But to be fair, Dennis Ritchie and Brian Kernigan wanted to do research in data processing systems. They had to fight for budget. The only way they could get budget to fund their proposals was to show "something practically useful" to ATT. And that was the birth of nroff and troff. They developed it and trained staff to use it for document preparation. Actually Steve Bourne of Bourne shell fame told me this personally way back in 1991 when he joined Sun Microsystems.

Dennis Ritchie did not start out intending to make something commercially viable. Nor did ATT want to market it as it turned out. Ritchie used the small platform to work on OS architecture and the idioms he thought were useful: what we all know today as the Unix philosophy of focused, simple tools that do one thing well but can be "integrated" so that they are used as building blocks to form more sophisticated applications.

I think like many other things (history keeps repeating itself) someone took it and tried to make money out of something that should have been made more robust even as a starting point for commercial versions. And thus created the chaos.

Also, Unix was partly a child of the failure of the monolithic Multics which tried to be everything to all people. But it did spawn a lot of good research and ideas.

Unfortunately, the computing field is rife with the ghosts of failed "technology" and quick-and-dirty attempts to make a fast buck. And, referencing the second article you cited (also a good one)
https://queue.acm.org/detail.cfm?id=2349257

I must agree. I am not trying to be glib or flippant here, but the bazaar is the quintessential marquis of the failure of human "cooperation." Winston Churchill famously quipped "The best argument against democracy is a five-minute conversation with the average voter." Might I add "with the average architect or programmer in the bazaar."

Bill Gates, sneaking around quietly like a mouse, took over the world because the Unix goliaths were fighting amongst themselves. They were too arrogant to do what we now know the Japanese, Chinese, Koreans and others did... start small, break into the market, then grow in sophistication and market share. Look at the automobile industry. The Japanese (and the Koreans learned well from the Japanese) didn't take Detroit head on. Do you remember the little Honda "toolbox" that came to the US in 1970? Now look at where Honda and Acura are.

IBM and Lexmark made the same mistake. They made printers that could print a thousand pages a minute for huge publishing operations. They haughtily laughed at the idea of home printers or even low- to mid-range sized printers for small businesses. And who won? HP....

Sadly, Linux is falling into the same quagmire as the articles you cite indicate. And the whole world wide web and the programming platforms we have today are a sad commentary on how we've regressed. More, new but not better. I'll stop here before I write a book.

Sorry, one more piece of information... I remember in the early 1990's (I want to say 1993 or 1994 but it might have been 1992) Dennis Ritchie gave a speech at the Usenix conference in San Francisco's Moscone Convention Center (he even wore his beanie with propeller -- seriously -- to make everyone laugh).

The talk was about the history of Unix and an introduction to Plan 9. I recall clearly he said that Plan 9 was never intended to be a commercial offering. It was his latest research project only. He talked about security and encryption a lot in that talk, discussing how no password is ever passed unencrypted inside the OS even (not talking about networking here).

I honestly don't remember much else about that session except that many, many people in the audience kept asking if he thought Windows was a threat to Plan 9. He quipped "no, why would it be if it's not a commercial system?" He then joked "Well, Windows might be a threat to other things."

Just some info.... I'm not trying to make a point.
;-)

The main problem with Linux is that it appeared at the scene when Plan 9 was a thing, and still it clinged to Unix "standards" for no good reason at all, except that it was everything Linus knew.

Originally Posted by YesItsMe The main problem with Linux is that it appeared at the scene when Plan 9 was a thing, and still it clinged to Unix "standards" for no good reason at all, except that it was everything Linus knew.

I think the timing that you mention is certainly a reason for the problems with Linux. And certainly you're right that Linus only knew a little. But that doesn't excuse the disastrous waste land that the Linux arena has become. In my opinion, the more overarching problem is always the impetus of the many parasites to make a fast buck--in the Linux world and in every other arena.

I recall these distro founders laughing--almost in ridicule of the community--how they made millions of dollars on their distros. Apparently, they were not primarily motivated to do anything right, good or for "the greater good" like Multics. And they made no effort to standardize, interoperate, improve. The varied opinions of "which distro is best" is a perfect exposé of the problem. I'll even go further. I believe that there are always multiple forces at work. In the Linux landscape
A. The developers were mostly immature, young, inexperienced, lacking in the perspective of history
B. The distro founders wanted to be "unique" (read: proprietary) to either distinguish themselves or lock you into their "platform"

And in that context, the Linux community as a whole is really no less guilty of creating chaos than Microsoft, Apple or the mainstream Unix commercial companies.

For example, I see no good engineering reason why Linux cannot standardize above the kernel: system utilities, user utilities, desktop environment and window managers. There is no good engineering reason why any of these platforms cannot exhibit loose coupling and high cohesion at the interfaces between the OS utilities, X-Window system, window manager, desktop environment.

I understand that the BSD and System V families are fundamentally different. But within each family, is there any good reason why we don't see truly modular, well-designed abstraction, loose coupling and high cohesion across "modules"?

The situation is really no better than Microsoft making its file formats proprietary intentionally, or Apple or any other company. If you want to compete, make a better thick client mail client, a better search indexer, a better disk defragmenter, etc. But don't stifle interoperability by going proprietary. For Linux, make a better desktop environment or window manager. Do we really need so many package managers?

The effect is that things are re-discovered and re-invented. And the community as a whole really makes incredibly slow progress if any. And it's not just the Linux community. Look at the disaster of the programming language and environment arena. Look at the disaster of the Internet: TCP/IP is universally regarded as the inferior protocol stack. But they did it quick-and-dirty and beat the OSI committee. Over the years, there has been the need to do much redesign of TCP/IP and the ancillary protocols. And still there is no session, presentation or application presence.

Then enter WWW and the disasters of HTTP, HTML, XML, XHTML, etc. A ghastly abortion of the best principles, knowledge and practices of computer science.

One poignant example is the quote from Brendan Each, author of JavaScript:

JS had to 'look like Java' only less so, be Java's dumb kid brother or boy-hostage sidekick. Plus, I had to be done in 10 days or something worse than JS would have happened.

So why has the industry embraced this garbage? Because the impetus of profit-chasing trumps the goal of quality, good engineering, standards and the struggle to take the best and build for the future.

And I think the human factor is a major reason why there are these disparities, incompatibilities, lack of interoperability, etc. And that affects improvement. There are hundreds of "programming languages" but are they really an improvement on what came before? And that's why those articles you referenced resonated so much with me.

I've been accused of being a critical zealot many times on LinuxQuestions. I'm really not. But I have seen good things bastardized for selfish, pecuniary motives. I'm an engineer, not a businessman. But that's why the businessmen scoff and deride me.

Originally Posted by rhimbo Apparently, they were not primarily motivated to do anything right, good or for "the greater good" like Multics.

I don't think that most of today's "developers" would even understand what made Multics better than Unix. (It was not only the safer programming language, it was not only the superior architecture, it was not only the commercial funding.) Plot twist, though: In 2020, languages emerge which claim to be "modern" and "a language for the 21st century" by introducing features which PL/I had in the mid-60s.

Times they should be a-changin'.

Originally Posted by rhimbo For example, I see no good engineering reason why Linux cannot standardize above the kernel: system utilities, user utilities, desktop environment and window managers. There is no good engineering reason why any of these platforms cannot exhibit loose coupling and high cohesion at the interfaces between the OS utilities, X-Window system, window manager, desktop environment.

Linux formally has a "better POSIX" (because who needs working standards), namely: the Linux Standard Base, which includes rules like RPM being the standard package format et cetera. I know of not one single distribution that follows it even almost completely. The latest attempt to "unify" Linux was systemd and it was a horrible idea as well.

Originally Posted by rhimbo Then enter WWW and the disasters of HTTP, HTML, XML, XHTML, etc.

The WWW is a more advertisement-friendly document presentation system than Gopher, and so it won the race quite easily. Because who wants to use a document viewer without ads, right? And one day, the typical Linuxers appeared on the scene and said: All I can do is write ugly websites. Let's make ugly desktop websites and waste everyone's computing resources! Hooray!

And those who spend the money are never those who know a thing about technology. Worse is better.

We need more 60s spirit, I guess.

I think we are in agreement sir… :-)

The design of MULTICS was excellent. Languages such as Algol and PL/1 were excellent. I learned a lot about MULTIS as an undergrad computer science major at MIT because two of my professors had been heavily involved on the MULTICS project. According to them its failure was that it was overly ambitious and unmanageable as a result of complexity and size and the hardware technology limitations of the day. And it tried to be everything for every application. But many great things came from it, including, in a way, Unix.

The name Unix was a jab at MULTICS that derided the bloat of MULTICS, perhaps unfairly because, as we know, MULTICS intended to be a robust, capable, commercial platform for serious data processing and research. MULTICS inspired Dennis Ritchie's (or Ken Thompson’s, I can’t remember) inspiration to design an architecture where utilities do one thing well but still “fit together” as a result of good modular design. The philosophy of “everything is a file” enabled this cohesion across utilities both at the OS and utilities levels. Moreover, Ritchie was influenced as a result of his formal training as a mathematician. I read an essay he wrote a long time ago that mathematical “functional composition” was one of the archetypes he tried to adopt in the Unix architecture.

At that same Usenix conference I mentioned in my previous post, I distinctly recall that he said that “a lot of things that look like Unix are not Unix.” And he was referring to the fact that the internal architecture had departed so much from his model that he didn’t consider those systems to “be” Unix systems.

I don’t criticize Unix because it was just a research project. But someone who attempts to make a commercial, robust, general purpose operating system needs to put in all the elements that support its purpose—that make it fit for purpose. I’m not sure I could say with confidence that any Unix or Unix-like system boasts that level of robustness or sophistication—not Solaris, not HP-UX, not Irix, certainly not Linux, not even RedHat enterprise.

I recall when I was still with Sun Microsystems in 2002, I was sent to consult at Charles Schwab. They openly laughed at the idea of replacing their IBM mainframes with Sun E68000, E-10K or E-15K “mid-range” systems. The mainframes were an order of magnitude more robust, more secure, more scalable. And they had transaction processing, fault tolerance and so forth that even the sophisticated Sun mid-range systems didn’t have. The IBM mainframes had on-die instant fail-over and secondary and tertiary fail-over.

Yet everyone denigrated the mainframe mostly as a marketing ploy. Everyone is peddling their wares. And this is the criticism I have of the Linux community. I’m not saying it’s all bad. But it’s tough to find something robust to replace for example my Mac. That being said, it might take more effort, but hopefully I’ll experience less vendor lock-in with Linux. Still, there is a lot of room for improvement.

But as a technologist, I’m just concerned overall with the industry in general. The same pattern exists in every domain from chip architectures to “advanced” web programming platforms, browsers, etc.

Don’t even get me started on programming languages. Barbara Liskov's CLU programming language from 1975 was wonderful, and it was a hell of a lot simpler than C++ or Java or Objective-C. Look at Lisp from the ‘60s, SmallTalk, and others. Actually, there has been no real advancement to speak of in language design. There have been advancements in compiler design and optimization techniques and in the “engineering” of microprocessors. But the fundamental foundation has not advanced.

I wrote real-time software for the Space Shuttle’s inertial and celestial navigation system in Jovial… a language from 1959…! Today, where is the improvement? Real-time Java? C++? Kotlin…? Rust…? (I think I’m going to throw up).

So enter C++, Java, Python, Perl, Ruby, JavaScript, and a thousand others… what do they bring to the table? And their programming and runtime environments are no better than Common Lisp—in fact worse.

I agree intensely with your comment about today’s Linux programmers. There are a few forces at work contributing to this. First, to quote Charles Darwin “Ignorance more frequently begets confidence than does knowledge.” Remember the adage: “You don’t know what you don’t know.”

Secondly, the youth are not being taught the foundations and lessons of the past. And certainly they’re not getting the historical perspective in the real world once they’re out of school.

Thirdly, the latest trends are unlikely to expose the youngsters to the knowledge, to enable them to accidentally stumble upon what they should be learning. The BSD code looks like it was written by software engineers. The Linux code looks like it was written by a bunch of high school students.

Final point… no one is teaching concepts. I sit my colleagues down—even those who are well into their forties—and I ask them what is the difference between CGI and today’s frameworks such as Java servlets, or .Net or Ruby or Grails or Python or Node.js…? They have no clue. I explain how “the web” programming model has come full circle and we’re exactly where we were in 1995. They look at me dumbfounded but don’t want to admit that they don’t understand what CGI was.

It’s the same with Linux. It’s been 20+ years since I worked on the SunOS and Solaris system utilities, drivers and some kernel stuff. But many of my colleagues who actually wrote NFS, SMP, etc. say that they are met with belligerent resistance by the Linux wet-ass fan boys who think they know better.

I hope the situation changes. I hope there is some respectable standards effort. LSB has turned out to be a calamitous failure. I’ve recently learned about systems so I read a lot of folks in the forums lamenting about that as you do.

OK, sorry, this is too long.