Jump to content

Talk:HTTP

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Usage? Support?

The table in 'Summary of HTTP milestone versions' has columns 'Usage in August 2024' and 'Support in August 2024', but there's no description of what usage and support mean. IMO, the meaning is far from obvious so therefore needs to be explicitly described. Stevebroshar (talk) 11:52, 2 November 2025 (UTC)[reply]

Receiving no opposition; I removed these columns. Stevebroshar (talk) 10:45, 14 November 2025 (UTC)[reply]

Versions are wonky

The article starts talking about "HTTP" which clearly is the topic. Then paragraph #4 starts "HTTP/1" with no explanation of what the heck that is. The reader is expected to know that "HTTP/1" is the wonky way that we refer to: version 1 of HTTP. It's weird so it needs to be explained. Saying "HTTP/1 was finalized and fully documented (as version 1.0)" is confusing unless you are already in the know. Should be more like: "Version 1 of the protocol, known as "HTTP/1", ... Stevebroshar (talk) 11:19, 10 November 2025 (UTC)[reply]

I added a versions section that describes how versions are labeled (HTTP/#). I put it early so that we define terminology/conventions before using it. Stevebroshar (talk) 09:36, 24 November 2025 (UTC)[reply]

What exactly is the topic?

Title is HTTP, but ... there are articles for HTTP/2 and HTTP/3, so is this article supposed to be for earlier versions? If not, where's the articles for those versions? ... I _think_ this article covers history and differences between all the versions, plus the details of .9, 1, and 1.1, but not the details of 2 and 3. Stevebroshar (talk) 11:24, 10 November 2025 (UTC)[reply]

I added a section called Versions that lists each version and notes that this article is more about versions before 2; that there are articles for 2 and 3 that cover them in more detail. Stevebroshar (talk) 12:46, 30 November 2025 (UTC)[reply]
Very few people need to know about HTTP/2 or HTTP/3, particularly their internals, and so the limited amount needed by most people can be handled by a section in this article. For people who need real detail on a protocol, we don't replace consulting the primary sources. Even fewer now need to know about HTTP 1.* and earlier.
What is needed here, in this article, is an overview of HTTP, a guide to how it is used, the format of how it handles requests, responses, document bodies and response coding, and description of the various methods within it (GET/POST etc.). That's broadly in this article, although the section ordering is poor and it's editorially weak. We don't need that long history section upfront, and a History of HTTP would be a worthwhile split. A hefty rework could be beneficial here. At >1,000 pageviews a day, it's certainly worth doing. Andy Dingley (talk) 15:41, 30 November 2025 (UTC)[reply]
I don't know how "Very few people need to know about HTTP/2 or HTTP/3" is important. How many people need to know about HTTP? What does the number of people have to do with my OP (or the price of tea in China)? What does need to know have to do with it? ... I know you cannot read my mind, but my point is that I see inconsistency in the articles: there's HTTP for versions before 2 and then there's one for v2 and one for v3. When there's inconsistency, one often wonders: why? and further: should we make it consistent? You say the why is that people don't need to know about v2 and 3 and my response is that makes no sense. ... I often find that WP info is factored weirdly. The MO seems to be: this article is too long so let's split it it up into chunks; like history and a list and whatnot. IMO that's not good. Let's keep related info together. In particular, we would not benefit from factoring out History of HTTP. There are good reasons to refactor that result in smaller articles, but _just_ making it smaller is not a good reason to refactor. Stevebroshar (talk) 11:07, 2 December 2025 (UTC)[reply]
Maybe you didn't notice or maybe you don't think I added value, but I edited the article modestly recently. The overall layout/flow was not good. It had good content, but was organized poorly. I reworked some of the high-level layout/structure to improve it. In particular, I considered moving the history section to the end. I think it's disruptive to the main thrust of the article. IMO a history section should be either first or last in an article and that placement depends on the nature of the info. If history is likely to be the primary thing folks want to know about a subject, then it should go first. But I think most would find the history of HTTP to be boring; so it should go last. Stevebroshar (talk) 11:14, 2 December 2025 (UTC)[reply]

The article talks about HTTPS, but what exactly is the relationship to HTTP? In particular, there are versions of HTTP (.9, 1, 1.1, 2, 3). Is there an HTTPS for each? For some? Or are these things unrelated? Are there versions of HTTPS? Stevebroshar (talk) 11:25, 10 November 2025 (UTC)[reply]

HTTPS has its own article (which is pretty clear) but could use better description in this article, to clarify the very valid questions you raise here. I'm no expert – I just use these things, and leave their implementation to others.
HTTPS, weirdly, doesn't mention rfc:2818 (2000) until late on and a History section. This is wrong – that primary source should be prominent. It's also the source (and still is) for the broad structure of HTTP / HTTPS / TLS. Which is somewhat complex, in order to make things simple otherwise (the wisdom of the old IETF at work). HTTPS is a thin (very thin) layer atop HTTP which doesn't try to replace HTTP or do the same task, but instead merely sets up the HTTP layer to use the underlying TLS to provide a secure channel for it.
Then much later on, in 2018, changes happened beneath TLS which could then allow HTTPS to become as ubiquitous as it is today; even without needing to change HTTPS itself. HTTPS 2818 was obsoleted by rfc:9110 in 2022 (and a bunch of things rolled into one document because they were now 'finished' and wouldn't need to evolve. Hmmm.), but 2818 is still the much easier read than 9110. Andy Dingley (talk) 15:53, 30 November 2025 (UTC)[reply]
I know there's a separate article for HTTPS. My point is that HTTP and HTTPS are highly related concepts, yet this article (HTTP) does not describe the relationship in a way that satisfies my vague understanding or my curiosity. My post is a request for someone to add this info. I guess the info I want could added to HTTP or HTTPS, or maybe, to some extent, to both. ... This bit: "HTTPS is a thin (very thin) layer atop HTTP which doesn't try to replace HTTP or do the same task, but instead merely sets up the HTTP layer to use the underlying TLS to provide a secure channel for it." is exactly what I'm wanting to know! You buried that in talk of citations for HTTPS!?!? ... I think, that your comments about citations in HTTPS belong in the talk page for that article; not this one. Stevebroshar (talk) 10:51, 2 December 2025 (UTC)[reply]