Talk:Web server/Archive 1

Market share, the second
There are a few web servers which serve one website on thousands of domains. (These which are running for domain resellers for example.) Netcraft try to remove them in their Activ-Webserver statistics. So this would be a much better source. (And then they gain totally different results. IIS nearly triples its market share.) But even this is a totally wrong statistic: On other domains like running hundreds of websites. An at last there are many websites which are not available for the public. Which would reincrease the market share of IIS an it would add many strange web servers like these of remote-controls for smart-TVs, stoves and home-routers. So I think it is nearly impossible to make stats of all market share.

So I would prefer to replace this statistic with the market share of the million busiest websites. This would be much more reliable. And is btw. a really interesting number. Fabiwanne (talk) 11:27, 28 August 2016 (UTC)

W3Tech has completely different numbers, and could possibly be used as a second source? I don't know the methods used by any of them. https://w3techs.com/technologies/overview/web_server/all — Preceding unsigned comment added by 2001:470:1F0B:56E:7E5C:F8FF:FE32:5302 (talk) 20:21, 2 April 2017 (UTC)

Hmm, I just stumbled over this page researching something else. The picture with the market share seemed exceedingly wrong. Netcraft changed to actually showing active sites instead of the stats used here showing mainly inactive, parked sites. The number that is shown in the above w3techs link is close to the number shown in Netcrafts active sites: IIE is close to 10%. Interestingly on the page liked to from this wikipedia article on the netcraft stats says: "The active sites metric is more appropriate for some applications, as it counts websites but excludes those that contain automatically generated content such as domain holding pages. Apache also has the largest share of this market (45.8%), with its total number of active sites now reaching almost 80 million – comfortably ahead of nginx, which takes up second place with 34 million active sites.". I therefore fail to see why Wikipedia is using the metric where each domain is measured equally as I understand it (google.com = UncleAlfredsFirstWebpage.info) instead of the 1 million busiest sites or at least only active sites (=not parked domain pages). Could this stats image and the table in this article possibly be changed to a more useful/relevant metric? — Preceding unsigned comment added by 31.10.31.200 (talk) 17:37, 29 August 2017 (UTC)

Thank you for including a less ridiculous statistic regarding market share. If the image/graph could reflect this it would be even better. — Preceding unsigned comment added by 2A02:168:4614:2:C536:1DC6:735C:1095 (talk) 07:54, 6 October 2017 (UTC)

Market Share QQ?
User:Lkt1126 added a number of things to the Market Share section.. One of the inclusions is that TenCent's "QQ" software (near as I can tell, a Chinese Instant Messenger client) is one of the most popular. I may be crazy but that doesn't seem possible. Looking now and doing more research, it seems that QQ uses QZHTTP as a server.. not "QQ". Although I don't feel confident about this. It doesn't seem like people can identify really whether QZHTTP is actually its own proprietary server, as Tencent claims it is. Dabizi (talk) 20:55, 16 November 2009 (UTC)

Market share

 * The most commonly-used web server, Apache, with over 60% of market share as of March 2003, is available from the Apache Software Foundation.


 * This kind of seems like advocacy/POV, especially since this is about all web servers, and not Apache in particular. Should this be removed? --User:Khym_Chanur 07:06, 25 Oct 2003 (UTC)


 * I agree that this is on the hairy edge of advocacy, but I don't think it was intended by the writer. It's actually useful information.  I came here  trying to find out precisely this: what to expect in the way of web server software on various systems.


 * I think we can swing this back towards the middle of NPOV by adding a little table giving statistics about what percentage of web servers use what software. Something like  Apache, 61%; BonzoWeb, 14%; HTTeePee, 11%, and so on (sorry, I made up those names).  It should also be broken down by operating system type.  That would be a great table; it would show the state of play in a neutral and informative way.


 * Trouble is, of course, that I don't know how one would go about collecting those statistics. If anybody adds such a table, please leave a hint about where the stats come from, so people can fact-check your ass.


 * Great article &mdash; I'm about to add some stuff that everybody knows, but we should say in case they don't.


 * ACW 16:44, 17 Oct 2004 (UTC)


 * ACW: the External links -section contains a link to the stats by Netcraft Web Server Survey. --ZeroOne 20:53, 17 Oct 2004 (UTC)

Question from 61.246.245.2
Is tomcat is an application server? If so, why? (posted to article by 61.246.245.2 on Dec 11, moved here at 06:29, 15 Dec 2004(UTC))

Is this page about web servers or about Apache?

Answer. It is about web servers, Apache is named for historical reasons (A Patchy Web Server derived from the first popular web server).

HTTP cookie
I have submitted the article HTTP cookie for peer review (I am posting this notice here as this article is related). Comments are welcome here: Peer review/HTTP cookie/archive1. Thanks. - Liberatore(T) 16:56, 14 January 2006 (UTC)

Server models
Would it be better to move all the stuff about concurrency and server models (threaded, process-based, etc.) to its own topic ?


 * I think so. I think that information is of very narrow interest and doesn't belong in this article. ---Elving 01:49, 11 April 2006 (UTC)


 * Perhaps an article on "web-server architectures"? --Allan McInnes (talk) 02:45, 11 April 2006 (UTC)


 * Maybe, but the information isn't really web-specific. It could be more generally described as simply server architectures. --Elving 05:35, 11 April 2006 (UTC)


 * Good point. A "server architecture" article would also be a good reference point for articles on other types of servers, as well as things like the software architecture article. Any objections to moving the content on concurrency and server models to a new article called "Server architecture" (or some better name, if anyone has any suggestions)? --Allan McInnes (talk) 16:37, 11 April 2006 (UTC)


 * No real objection, but I would try to disambiguate the server name because it can mean an hardware machine or a software program (it depends on the context). What about server models (software) ? --Ade56facc 07:50, 13 April 2006 (UTC)


 * That sounds good to me. --Allan McInnes (talk) 16:58, 29 April 2006 (UTC)


 * OK, better late than never, moved chapter Concurrency and server models to new article and renamed the chapter to Performances. --Ade56facc 16:25, 7 October 2006 (UTC)


 * Update: sadly a wiki administrator (UtherSRG) has removed the above article "server models (computer software)" because it has been understood as an HOWTO instead of an explanation of how things work. --Ade56facc 19:00, 8 November 2006 (UTC)

Web servers
I have serious problems with this article. I've been in IT for 20 years and a Web Server is the software that serves web pages NOT the physical hardware. The physical hardware is simply a server. To prove the point, a physical server can not serve a webpage unless it has a Web Server installed such as: Apache, Nginx, IIS Therefore, a physical server is just a server, while a web server is the software that does the job. I realize that the article is covering the subject fairly well and states that it can refer to the physical hardware or the software but showing an image of a Dell server and calling it a web server is misleading & confusing & propagates a misunderstanding. Referring to the hardware as anything more than a server (other than to make note that an IT short hand reference to the hardware as a web server sometimes occurs) is misleading.

In nearly all cases hardware doesn't serve web pages it's the software that is actually acting as the web server. NOTE: even embedded systems typically run a version of Linux, or even occasionally heavily trimmed version of windows... this is an Operating System ... and they require a Web Server to be installed in order to serve web pages... to disambiguate this article it should be made very clear that a web server is software and that sometimes, for the sake of simplicity, the physical server hosting the software is referred to as a web server. - --75.174.48.115 (talk) 19:32, 6 June 2016 (UTC) (actually just lost my Wikipedia login ... though I've only rarely contributed)

What is a web server, explained so that even dad gets an idea what they are?

I'd like an indication of what is and how to install a minimal OS to run a web server. presumably a kernal and what else? mysql and apache too. the point being that apache is an application and presumably requires an OS, but what is essential?

What is a "managed server"?

Large file support
Large file support to be able to serve files whose size is greater than 2 GB on 32 bit OS. I really can't think of why my web server would be better or more of a web server if it served files more than 2 GB. Is this really an inherent part of the definition of webserver? It doesn't seem comparable to any of the other qualifications: HTTP support, dynamic support, etc. --Mrcolj 21:54, 4 September 2006 (UTC).


 * A1. It's not part of the definition, it's only one of the most common features that nowadays are requested by web masters in order to serve huge multimedia files, backup archives, etc.; just as an example Apache 1.3.x does not have any large file support whereas Apache 2.0.x and 2.2.x do have large file support. --Ade56facc 11:32, 7 October 2006 (UTC).


 * Support for large files is a software aspect and not a feature per se. A software aspect enables a feature to function as desired, in this case serving files. Same goes for configurability and extensibility. I removed all three. 217.132.73.191 16:35, 26 May 2007 (UTC)


 * I agree that large file support is not a fundamental quality, and that a negligible fraction of web hosts actually need to serve files anywhere near 2 gigabytes. Meanwhile, someone has added it back saying "Adding LFS support back in, as it is certainly a feature to be able to download DVDs images ... and it was only recently that the most common web server could actually do so" -- intgr #%@! 20:14, 27 May 2007 (UTC)

English grammar used in this article
The tendency toward the exclusion of the article ("the" or "a") disrupts the reading and gives the impression that the article was written by a foreigner, not a native english speaker.

Common features
repeated twice in the page, deleted —Preceding unsigned comment added by 86.199.91.7 (talk) 21:29, 11 February 2008 (UTC)

The section of standardization is added...
Please feel free to edit it if it is inappropriate. Currently, I'm not able to find the review report on the standards. It would be beneficial if people could provide the citation of the review/justification/validation reports for those stds —Preceding unsigned comment added by 64.62.138.97 (talk) 10:49, 26 February 2008 (UTC)

The following are two examples of the critical review on ISO 17025 and validation report on ISO 11290

http://www.atypon-link.com/AOAC/doi/abs/10.5555/jaoi.2003.86.5.1038?journalCode=jaoi

http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6T7K-42HFWX2-8&_user=10&_rdoc=1&_fmt=&_orig=search&_sort=d&view=c&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=284b649a4b53306ce03fa34911e0fc27

—Preceding unsigned comment added by 64.62.138.97 (talk) 13:11, 26 February 2008 (UTC)

I would suggest that ISO people could refer the standards, at their website, to the publications or documenents from which the standards have been derived, if it is not confidential.

Such practices are very common with many well-known academic journal publishers.


 * The standardization process of many www technologies is clearly leaded by W3C (WWW Consortium) and handled together with IETF;  a note could be added to clarify the words "through a standardization process" but I suspect that all this might want its own mini-article.  Adding good references about what is written in the article will take sometime, I guess, because many old sources may be considered obsolete or incomplete nowadays, so be patient (this task could be easily done by everybody who has enough time to collect those references);  anyway, here we are not talking about a method to do something or a physical object (excepted for the concept of computer server) but about generic software implementations (that can be very simple and incomplete or very sophisticated and complicated) of well defined standards (de facto) described by RFC(s) (Request for Comments) that are publicy available, so in those RFC(s) you can find all needed technical references. --Ade56facc 09:13, 9 April 2009 (UTC).

Revert
I've restored the article to its state three edits ago, someone decided it was a good idea to remove lot's of spaces and paragraphs, breaking the wiki syntax and thereby making the article unreadable. No valuable changes have been lost in the process Floker (talk) 04:52, 28 February 2009 (UTC)

Unencyclopedic phrase needs rewrite
From the opening paragraph:

"Most people think a web server is just the hardware computer, but a web server is also referred to as the software computer application that is installed in the hardware computer."

The phrase "most people" is unencyclopedic. The intent is to draw the reader's attention to the distinction between hardware and software, which is an important distinction worth indicating. It needs a simple rewrite, that's all.

If no one else jumps in to do it, and no one objects, I'll feel free to do it.

Karl gregory jones (talk) 06:00, 14 December 2010 (UTC)
 * Concur. That is an important distinction, but the sentence as currently worded was apparently written by a young child under the age of ten. --Coolcaesar (talk) 07:29, 14 December 2010 (UTC)

Dell advertising
The two first illustrations are Dell equipment and its photos... Please diversity, not  advertising Krauss (talk) 20:52, 18 August 2017 (UTC)