4 ms·

That's a shame. I am reading through this thread because I saw your comments on a NetBSD thread, and couldn't believe there wasn't a place I could go read more

That's a shame. I am reading through this thread because I saw your comments on a NetBSD thread, and couldn't believe there wasn't a place I could go read more in-depth descriptions of what you were saying about how you browse the web over there. https://news.ycombinator.com/item?id=37282334 https://news.ycombinator.com/item?id=37282334 if you've any links or anything to share on the topic, I'm all ears!

1vuio0pswjnm7 2y ago
Note archive.netbsd.org has been fixed since then. 1.6 is retrievable again. I have addressed the value (to me) of HTTP/1.1 pipelining in many comments, despite consistent backlash from HN commenters trying to promote new protocols that make surveillance and advertising easier and add needless complexity to information retrieval. I have been sharing the simple, unsophisticated programs I write to perform pipelining outside the browser. Given that I do not "browse the web" the same way that most HN commenters and programming nerd bloggers/website operators would advocate, it should not be surprising that I also do not subscribe to their idea of websites. (Conglomerations of ads, tracking and resources from disparate IP addresses that require "interaction".) I retain an older, simpler idea of websites: all resources from a single IP address. That allows for easy pipelining to efficiently retrieve information non-interactively. For communication, i.e., 1-to-1 or 1-to-smallnumber, I have always prefered the idea of peer-to-peer overlay networks, long before they became as "popular" as they are today. Not websites. Why are so many people reading "news" on Facebook, why the large audience. Possibly they are actually using its website to communicate or interact with individuals or small groups and that's why they log in. And here I am submitting this comment while logged in to... a website. Is this how I think we should be communicating over the internet. Via websites. No. From what I see, it is the source of many problems. But, here we are. Comunicating via someone else's website. One of the uses I make of HTTP/1.1 pipelining is for retrieving DNS data in bulk via "DoH". Sometime after I started doing this years ago I noticed someone else was implementing pipelining when using "DoT". Not sure what to make of that, if anything. Another use is for "Web APIs". Another use is for retrieving bulk HTML via XML sitemaps. Anytime I want to retrieve and consume data/information efficiently in quantity, HTTP/1.1 pipelining is useful. Web browsers, whether text-only HTML viewers or media players with Javascript engines, can be useful but I tend to do most "browsing" offline from simple HTML pages I have created using text/html/json that I have retrieved from the www and processed outside the browser. I do not use a browser as the primary means of sending HTTP requests nor do I use Javascript as the means to process text/html/json into a format that I can consume. The preferred formats for me are plaintext, CSV and SQLite3, comprised of 7-bit ASCII if possible. I retrieve the bits from the web using command line programs and then process to those formats, if needed, using small filter programs I write. Popular web browsers cannot do HTTP/1.1 pipelining. HTTP/1.1 pipelining is not advertising-friendly. Popular web browsers, namely the company/companies behind them, have developed "replacement" protocols that are very advertising-friendly. They tell the public that HTTP/1.1 was a "problem" that needed to be fixed. The question is who was it allegedly a problem for, exactly: for those companies or for www users. I'm using HTTP/1.1 pipeliing every day for over 20 years. I'm not seeing any ads. Not a problem for me.
- sourcepluck 2y ago
  My networking knowledge is sub-par, I'm relatively new to all this computer stuff. I can mostly follow what you're saying though, and you've given me many useful avenues for further exploration. Even though I was only somewhat understanding what you're saying, I'm very interested in alternative approaches to "obvious" things, and I love when the curtains get pulled back and you go: "oh wow, it's just that". So I'll be looking into this, and have saved these posts. Cheers.
1vuio0pswjnm7 2y ago
You might like this pipelining usage example if have not already seen it. https://news.ycombinator.com/item?id=38777921 https://news.ycombinator.com/item?id=38777921 No browser, no Python, no curl. The later does not work well for piping binary data from the www to other programs. tnftp can do it. nc, tcpclient, etc. can do it with yy054.

1vuio0pswjnm7 2y ago

Note archive.netbsd.org has been fixed since then. 1.6 is retrievable again. I have addressed the value (to me) of HTTP/1.1 pipelining in many comments, despite consistent backlash from HN commenters trying to promote new protocols that make surveillance and advertising easier and add needless complexity to information retrieval. I have been sharing the simple, unsophisticated programs I write to perform pipelining outside the browser. Given that I do not "browse the web" the same way that most HN commenters and programming nerd bloggers/website operators would advocate, it should not be surprising that I also do not subscribe to their idea of websites. (Conglomerations of ads, tracking and resources from disparate IP addresses that require "interaction".) I retain an older, simpler idea of websites: all resources from a single IP address. That allows for easy pipelining to efficiently retrieve information non-interactively. For communication, i.e., 1-to-1 or 1-to-smallnumber, I have always prefered the idea of peer-to-peer overlay networks, long before they became as "popular" as they are today. Not websites. Why are so many people reading "news" on Facebook, why the large audience. Possibly they are actually using its website to communicate or interact with individuals or small groups and that's why they log in. And here I am submitting this comment while logged in to... a website. Is this how I think we should be communicating over the internet. Via websites. No. From what I see, it is the source of many problems. But, here we are. Comunicating via someone else's website. One of the uses I make of HTTP/1.1 pipelining is for retrieving DNS data in bulk via "DoH". Sometime after I started doing this years ago I noticed someone else was implementing pipelining when using "DoT". Not sure what to make of that, if anything. Another use is for "Web APIs". Another use is for retrieving bulk HTML via XML sitemaps. Anytime I want to retrieve and consume data/information efficiently in quantity, HTTP/1.1 pipelining is useful. Web browsers, whether text-only HTML viewers or media players with Javascript engines, can be useful but I tend to do most "browsing" offline from simple HTML pages I have created using text/html/json that I have retrieved from the www and processed outside the browser. I do not use a browser as the primary means of sending HTTP requests nor do I use Javascript as the means to process text/html/json into a format that I can consume. The preferred formats for me are plaintext, CSV and SQLite3, comprised of 7-bit ASCII if possible. I retrieve the bits from the web using command line programs and then process to those formats, if needed, using small filter programs I write. Popular web browsers cannot do HTTP/1.1 pipelining. HTTP/1.1 pipelining is not advertising-friendly. Popular web browsers, namely the company/companies behind them, have developed "replacement" protocols that are very advertising-friendly. They tell the public that HTTP/1.1 was a "problem" that needed to be fixed. The question is who was it allegedly a problem for, exactly: for those companies or for www users. I'm using HTTP/1.1 pipeliing every day for over 20 years. I'm not seeing any ads. Not a problem for me.

sourcepluck 2y ago
My networking knowledge is sub-par, I'm relatively new to all this computer stuff. I can mostly follow what you're saying though, and you've given me many useful avenues for further exploration. Even though I was only somewhat understanding what you're saying, I'm very interested in alternative approaches to "obvious" things, and I love when the curtains get pulled back and you go: "oh wow, it's just that". So I'll be looking into this, and have saved these posts. Cheers.

1vuio0pswjnm7 2y ago

You might like this pipelining usage example if have not already seen it. https://news.ycombinator.com/item?id=38777921 https://news.ycombinator.com/item?id=38777921 No browser, no Python, no curl. The later does not work well for piping binary data from the www to other programs. tnftp can do it. nc, tcpclient, etc. can do it with yy054.