Talk:HTTP/2

HTTP 2.0 (estimated) release date

Hi all, I was thinking that, if it is known and if it comes from a reliable source, it would be a good idea to add in the article the estimated release date of HTTP 2.0. I am pretty sure this addition would add value to the article. Does anybody know if there is an estimated release date of the new protocol? Thanks --★ Pikks ★ ^MsG 08:31, 10 August 2012 (UTC)

I added the milestone table from the charter. The current plan is for standardization in late 2014. This could have general adoption likely beginning in 2015, though you'll probably get experimental implementations in Firefox and Chrome sometime in 2014 for testing and use with early adopting sites. Note that it is entirely possible that this timetable could get delayed for a lengthy time if no consensus is reached during its development period, so there's no way to state any release date with any certainty at this early stage. 108.16.222.138 (talk) 15:23, 2 June 2013 (UTC)

A "Criticism"-section should be made

I'm sorry, but this page seems somewhat one-sided to me. Having operated webservers in-depth for some time, I would like to point out there is resistance towards HTTP 2.0 / SPDY: https://www.varnish-cache.org/docs/trunk/phk/http20.html

I'll make a "Criticism"-section if I get the time, but until then, I'll just leave this comment. NickyThomassen (talk) — Preceding undated comment added 20:02, 20 March 2014 (UTC)

Hello there! Hm, this is a quite interesting quotation from the above linked page:

"Our general policy is to only add protocols if we can do a better job than the alternative, which is why we have not implemented HTTPS for instance."

So, how did they do a better job regarding HTTPS, for example? — Dsimic (talk | contribs) 04:14, 22 March 2014 (UTC)

By not implementing it and leaving it up the user to proxy HTTPS and strip the SSL. Perbu (talk) 06:33, 27 May 2014 (UTC)

Given the latest comment by a member of the WG, it seems that a criticism section is something that people will be looking for (http://lists.w3.org/Archives/Public/ietf-http-wg/2014AprJun/0815.html) Sgaragan (talk) 11:50, 27 May 2014 (UTC)

A separate "criticism" section is strongly discouraged, see WP:CSECTION. Multiple viewpoints should be described in context. -- intgr [talk] 16:14, 27 May 2014 (UTC)

More references, less opinion

Organization of this article is well done and the content ordering seems to flow. A more neutral stance could have been taken, especially in the “Background of HTTP 2.0” section. You write the phrase, “it is important to understand…” several times, and this should be reworded to seem more neutral, and read more like an encyclopedia article rather than an opinionated one. I’d love to see more connections to other links. Your “Protocol” section has many, but the other sections are lacking. More outside sources may provide readers with a better understanding of the topic and allow for even more research. This correlates with citations/references also. You’ve got a good start on citing your work, and I think even more could be done. More citations at the bottom of the page, and in-text citations throughout the article would prove the article to be more authentic and well researched. Keep up the good work! Goblue2013 (talk) 21:59, 28 March 2014 (UTC)

This phrasing has since been changed, and I improved the sourcing. -- Beland (talk) 16:12, 27 August 2014 (UTC)

Structure of the article

The article contains a lot of information, but could have a better structure. For example you have a subsubsection goals, but also talk in three other sections about the goals in general. And this is just on example that I noticed immediately. What I did not find immediately are the actual changes from HTTP 1.x to 2, while those should deserve a subsection. Also there is a lot of redundant information, probably because the article grew along with the work on the standard and related stuff is all over the place, including information that is only historically relevant (the development should get an own subsection as well probably). Now that HTTP/2 is almost done, the article should get an overhaul. Also the new name is [HTTP/2 instead of 2.0] --85.16.76.63 (talk) 14:04, 14 August 2014 (UTC)

I'm fixing both structure and naming now. -- Beland (talk) 15:24, 27 August 2014 (UTC)

Rename to HTTP/2?

The following discussion is an archived discussion of a requested move. Please do not modify it. Subsequent comments should be made in a new section on the talk page. Editors desiring to contest the closing decision should consider a move review. No further edits should be made to this section.

The result of the move request was: page moved, self-closing per WP:SNOW. -- intgr [talk] 10:17, 14 September 2014 (UTC)

---

HTTP 2.0 → HTTP/2 – Given the very clear official preference at [1] and the fact that a "http 2" Google search mostly also turns up references to "HTTP/2", I think we should rename this article. -- intgr [talk] 08:15, 8 September 2014 (UTC)

Support as per nom Gregkaye ✍♪ 10:54, 8 September 2014 (UTC)

Support as per nom. 2014Best (talk) 16:50, 8 September 2014 (UTC)

Support even a search for "http 2.0" leans toward the official name. —innotata 21:12, 13 September 2014 (UTC)

---

The above discussion is preserved as an archive of a requested move. Please do not modify it. Subsequent comments should be made in a new section on this talk page or in a move review. No further edits should be made to this section.

Differences from SPDY

The section on differences from SPDY has some rather unfriendly ambiguous text:

SPDY communicates separately with each host, which means that multiplexing happens only at one host at a time, no matter how many connections are open. This means that SPDY can only download things from one host at a time. The improvement HTTP 2.0 makes on this is that it allows multiplexing to happen at different hosts at the same time. This makes downloading multiple web pages or content from the Internet significantly faster.

This doesn't make any sense to me, and the reference is simply pointing at the spec. Frankly this sounds like a load of bollocks. HTTP/2 still specifies that there should only be one connection to a host per client. "SPDY can only download things from one host at a time" is--AFAICT--false.

Agree. It's just bs.178.23.225.78 (talk) 16:17, 5 December 2014 (UTC)

NPOV Dispute

This page has a NPOV banner, but I don't see any discussion of POV on this page currently. Should it be removed, and if not, what is the dispute?

mnot (talk) 01:05, 10 February 2015 (UTC)

I believe the reference was to the already existing "More references, less opinion" section (still above). FWIW, the article does seem better now. Rwessel (talk) 04:17, 10 February 2015 (UTC)

I say we vote on the removal of this tag. It doesn't seem to push a point of view at this moment. If anyone can point out where it does, I'd be happy to attempt to clean up the article! Meşteşugarul - U 16:23, 13 February 2015 (UTC)

While the article does still seem a bit positive on HTTP/2, it's hard to spot any criticisms in the real world other than the encryption/security related ones, which are covered in the article. Rwessel (talk) 08:29, 14 February 2015 (UTC)

Encryption: mandatory TLS 1.2

User:MureninC has added material sourced from what appears to be their own blog/mailing list posting. This has been reverted before, and I have reverted it again. Let's discuss the addition, specifically any sources meeting WP:RS supporting it, here. Rwessel (talk) 06:35, 16 February 2015 (UTC)

Thanks Rwessel: some of that text simply is incorrect ("consensus pro encryption"), the other part just copies stuff that was sent to the WG's mailing list and IMHO was properly replied to over there. This is not a place to copy mailing list discussions from a unilateral point of view. Reschke (talk) 07:16, 16 February 2015 (UTC)

The veracity of the source itself shouldn't be brought to question. It's the World Wide Web Consortium's mailing list. What should be brought to question is how the source is represented and whether the source itself presents something worthy of inclusion in Wikipedia. Reschke's statement about "consensus" being incorrect is very important to address. It is ultimately difficult to draw up a consensus from a mailing list. I see issues with how the source is described in the article, but not the source itself as material that could be useful. Meşteşugarul - U 08:56, 16 February 2015 (UTC)

I agree, and my last edit, the one that was reverted, did address it. MureninC (talk) 21:48, 16 February 2015 (UTC)

Your latest edit still claims that there was a consensus pro-encryption, which is not the case. Reschke (talk) 06:35, 17 February 2015 (UTC)

I don't see such statement in the latest edit that was reverted; if you refer to the in lieu of consensus for mandatory TLS paragraph, then that one was a part of a bulk revert from a previous revert; if you think it's incorrect, I'm fine with it being fixed. MureninC (talk) 19:35, 17 February 2015 (UTC)

Well, posting something to a public mailing list and then using your own archived mail as support for your POV seems to be problematic. The mails have been replied to, so if this topic is to be included here the text would need to represent those replies as well. 217.91.35.233 (talk) 16:47, 16 February 2015 (UTC)

No, 217.91.35.233 / Reschke, what's problematic is that you as a non-trivial member of the standards body in question, seem to revert the opinion about the failures of the group as a whole, and also do so from your non-registered account first, probably trying to hide your affiliation with IETF. MureninC (talk) 21:48, 16 February 2015 (UTC)

I did revert without logging in the first time, for which I apologize. If I wanted to hide something, I definitively wouldn't have done that using my account the other time. Reschke (talk) 06:32, 17 February 2015 (UTC)

Rwessel, the last edit, which you've reverted, did actually/finally discuss the points raised by others in their replies on the mailing list (e.g. note how it's different and longer than the prior parts that were reverted), so, I do believe that it's NPOV now, and your revert was not needed. As for Reschke, who has reverted the text twice (first under 217.91.35.233, possibly to hide his close affiliation with HTTP/2 WG), prior to me finally adding the POV from the replies I did receive on the list, he appears to be non-trivially involved in these IETF standard bodies, and, as such, it's unclear why he keeps reverting the POV of non-members here on Wikipedia, which they, as a standards body, fail to address. Rwessel, as I mentioned in my last summary prior to your revert, the article isn't so large as to warrant removing valid points, and the points I've added remain valid, and they aren't sourced by my "blog", either, and have not been addressed further in the mailing lists beyond the latest edits that were already incorporated in my last edit which you've reverted. In light of all of this, please kindly revert your revert. If necessary, feel free to improve the text. HTTP/2 development as a whole has had a less than stellar track record of keeping everyone happy and addressing all of the present issues of the WWW, yet this article seems very skim on the precise details of as to why. MureninC (talk) 21:48, 16 February 2015 (UTC)

The references are simply not adequate. They're not a WP:RS, and they don't support the responses. And adding links to the responses won't work because that's still, at the very best, a primary source, and even then combining all that runs you with into WP:SYNTH. User:Mnot's link might suffice, though. The size of the article is immaterial, the valid points should be sourced appropriately. Rwessel (talk) 04:15, 17 February 2015 (UTC)

Rwessel, but the text quoted is not something that an expert cannot verify independently, so, I'm not sure a reference is even necessary. Can you explain why exactly are you reverting the text, if you merely disagree with the source of the reference? Do you disagree with the text, too? Aren't just the facts have to be referenced? The statements in question doesn't really introduce any new facts to a well-versed reader, and the analysis of the existing facts is quite trivial, yet important to understanding the subject, so, I'm unsure why the whole text is being reverted, if you merely disagree with the reference. Yes, size is immaterial; however, both the article is rather small and the section in question itself is extremely small, so, why do you keep reverting this text as a whole, instead of providing additions to represent another POV? MureninC (talk) 19:35, 17 February 2015 (UTC)

See WP:WHYCITE. "In particular, sources are required for material that is challenged or likely to be challenged – if reliable sources cannot be found for challenged material, it is likely to be removed from the article." As the statement has already been challenged, that requirement is already met. So we need a source that meets WP:RS and WP:V ("All content must be verifiable. The burden to demonstrate verifiability lies with the editor who adds or restores material, and is satisfied by providing a citation to a reliable source that directly supports the contribution."). In general verifiability by an expert (how could anyone verify that?) is not sufficient, a proper source is required. And frankly, an expert ought to be able to provide citations supporting for their opinion. Rwessel (talk) 22:25, 17 February 2015 (UTC)

I do not believe the statements were actually challenged; I believe they simply have been removed because the original text didn't summarise the other PoV of the replies that the original text has received (fails to mention the feedback that was provided / the text fails to mention the replies the author got). This has since then been addressed in the latest edit, so, there is no more a need to still keep the whole text and section reverted. MureninC (talk) 02:13, 18 February 2015 (UTC)

FWIW, the membership of an IETF Working Group is defined by the set of people posting to the Working Group's mailing list. So MureninC is a member as well. It escapes me why one member would have the right to misrepresent what happens in the Working Group, while other members would not have the right to fix these claims. Reschke (talk) 06:39, 17 February 2015 (UTC)

Reschke, but you aren't fixing the claims. You're simply removing valid claims made by someone else. If you think claims are overrated, you should provide another point of view. My last edit (as referenced earlier) did provide such point of view. MureninC (talk) 19:35, 17 February 2015 (UTC)

FWIW, my shepherd writeup might be a useful source for this text. YMMV. mnot (talk) 00:01, 17 February 2015 (UTC)

Mandatory Encryption - language

"It is noteworthy that even though the mandatory encryption that has been criticized by an agent of the industry leader in 2013 has not been made mandatory as part of the standard, as of early 2015 the agenda has nonetheless been carried over by the leading browser makers, who have thus far refused to implement HTTP/2 without encryption."

This sentence is very hard to understand. Trying to communicate that the leading browser makers implementing with encryption... or not? I am guessing yes, ("the agenda") but that's after reading several times, and by slightly knowing the subject. Too long of a sentence and too many negations to make quick sense out of it. — Preceding unsigned comment added by 80.94.146.19 (talk) 07:03, 25 February 2015 (UTC)

Browser Support

Holizz recently removed the HTTP/2#Browser support section, because most of the content was on the Comparison of web browsers page and this had been tagged for a merge. ‎128.211.193.44 has reverted this because the table contains information about TLS support. I near as I can tell, *all* web browsers supporting HTTP/2 support TLS, but some major ones (Chrome, Firefox, IE), so not support (unencrypted) non-TLS session. While a non-TLS HTTP2 column could be added to the Comparison of web browsers article, or perhaps a multi-way indication of support in the one existing HTTP/2 column, I'm not sure that's needed. The (lack) of non-TLS support appears to be a non-issue, as very, very little would actually use that mode, and so very few servers will even be set up to support it. So my suggestion is to just go ahead and remove this section again. Rwessel (talk) 01:24, 31 October 2015 (UTC)

I found a source for all the browsers listed in the "Browser support" section not supporting HTTP/2 without TLS and added that to the encryption section, and also explicitly listed those browsers. Though personally I feel that the fact that the spec allows HTTP/2 without TLS is a piece of esoterica that will soon be largely forgotten as all the major implementers have totally rejected the notion. I also realised that MS Edge was not in the Comparison of web browsers and added its support for HTTP/2 along with a citation. I believe these changes should address ‎128.211.193.44's concerns and so I unreverted my edit. --holizz (talk) 14:21, 1 November 2015 (UTC)

Commentary removed

I've removed the following from the article:

This can be named a "total encryption". [...] The total encryption may have some side effects. Technically, it is contrary proxy-servers and data inspection by nature. This mean lack of opportunity to reduce overall traffic by, at least, proxy-caching and to applying policy by organizations to its employees. Other effects lay in concept of encryption and privacy. For many years encryption stay for a technology of keeping privacy when acting with a sensitive information. This implicitly imply a definite meaning of the information and the act. The classic example is a financial transaction where the information is your identity and the act is a money transfer. Nowadays and further with total encryption the information and the act is undefined but encrypted. An end-users applications and, globally, devices will do encrypted data transfers but what the data to be transfered and where is unknown for a user. It is impracticable to document every application interactions. And it is hard to check application manifest due to encryption. So, a sensitive information may go out of control.

This appears to me to be a mixture of commentary and speculation, which is not allowed in articles: see WP:OR, WP:CRYSTAL. -- The Anome (talk) 12:47, 23 December 2015 (UTC)

HTTP/2 New Version since HTTP 1.1 Mismatch

Hello,

I was reading the wiki article on HTTP and following I read the one on HTTP/2. I noticed there is a discrepancy. In HTTP/2 it reads:

"HTTP/2 is the first new version of HTTP since HTTP 1.1, which was standardized in RFC 2068 in 1997"

But the article on HTTP reads:

"The first definition of HTTP/1.1, the version of HTTP in common use, occurred in RFC 2068 in 1997, although this was obsoleted by RFC 2616 in 1999."

Does it mean that HTTP/2 replaced the version from 1999 rather than the version from 1997 the HTTP/2 article states?

All the best. — Preceding unsigned comment added by 8bit traveler (talk • contribs) 18:54, 8 August 2016 (UTC)

W3Techs should get a page

The introduction on this page mentions W3Techs in a way that implies they are an Authoritative Source of Information. If that is true, then Wikipedia should have a page describing W3Tech. Gwrede (talk) 19:36, 15 September 2018 (UTC)

Yep. WP:DIY. — kashmīrī ^TALK 20:01, 15 September 2018 (UTC)

Lead section should focus on what HTTP/2 is, and adds to 1.1

Currently the lead section is the history of HTTP/2's development and introduction, its current browser support, and the history of its successor's development and introduction. It would be more useful to move most of that detail to separate sections, and concentrate in the lead on what HTTP/2 is. The best way to do that would seem to be to refer to HTTP for the general idea, HTTP 1.1 for the starting point, and then spend the time talking about what HTTP/2 added or changed.

I'm not the right person to write this, but as a reader it looks like the main points are those under "decrease latency" in the goals. The rest of the goals are just '...without breaking anything'. Steven Kelly (talk) 13:47, 18 November 2020 (UTC)

I fully agree. I came to visit this page to write an opinion article on one of my blogs and was looking for information regarding HTTP/2 and HTTP/3. What immediately caught my attention was the claim '[...] while there's non-default support in stable versions of Chrome and Firefox and HTTP/3 can be enabled'. Uh, no. As of 2020/11/18, and under macOS Big Sur, HTTP/3 support is only present in the Canary version of Chrome (and Microsoft Edge) and the Nightly build of Firefox, which are technically non-stable versions of those browsers, intended to test out radical new technology before they even go into beta (or release candidates). Although the flags for activating HTTP/3 actually exist on the current version of Firefox, they seem to be non-functional. MS Edge for the Mac, even though it's built upon Chromium, does not seem to have the ability to activate HTTP/3. Safari 14, however, seems to work 'straight out of the box' with HTTP/3 support, and it's reasonably stable. Such comments, however, would be appropriate for the HTTP/3 page, not here! Gwyneth Llewelyn (talk) 18:08, 18 November 2020 (UTC)

References