Talk:FreeBSD/Archive 2

Merge Sysinstall into FreeBSD?
I believe that Sysinstall is redundant with FreeBSD, and should be merged with into this page. sysinstall is inly used on FreeBSD, correct? Gigglesworth (talk) 23:58, 21 July 2011 (UTC)
 * Added "proposed merge" markup to both pages. Further discussion should be here only, per Help:Merging. The Crab Who Played With The Sea (talk) 13:50, 24 February 2013 (UTC)


 * Suport Sysinstall includes a lot of discussion about the other installers so does not stand alone. -—Kvng 14:42, 27 February 2013 (UTC)

"regarded as reliable and robust"
I went to Google Books and search the source for "reliable", and then "robust". I wasn't able to verify the information "regarded as reliable and robust". -- Jorge Peixoto (talk) 15:34, 17 May 2012 (UTC)
 * I searched for it but couldn't find any sources so I've removed it. --Kondi (talk) 14:22, 4 August 2012 (UTC)

TrustedBSD
The TrustedBSD reference to DARPA claimed a citation was needed. I added "Under DARPA/SPAWAR contract N66001-01-C-8035 ("CBOSS")", and removed the "Citation Needed" bit. I assume that the contract number is adequate citation. — Preceding unsigned comment added by 97.127.182.235 (talk) 19:54, 24 March 2013 (UTC)

Kernel type
Shouldn't this be Monolithic instead of Modular? — Preceding unsigned comment added by 66.108.42.112 (talk) 23:55, 28 May 2013 (UTC)

8.4 is not an 'unstable release'
The template 'latest preview software release' mistakenly lists FreeBSD 8.4 as unstable release. This clearly is wrong; 8.4-RELEASE is not 'unstable'. If we are to apply FreeBSD terminology, it is a release of the 'legacy' branch 8.x. Some people claim even that 8.x is (currently) more stable than 9.x. Please correct this. 84.50.246.19 (talk) 09:46, 18 June 2013 (UTC)

favourite amongst ISPs during dot com bubble
Hotmail was originally hosted on a mix of FreeBSD and Solaris: http://en.wikipedia.org/wiki/Hotmail#Launch_of_Hotmail

Microsoft has made use of FreeBSD internally: http://betanews.com/2001/06/18/microsoft-we-use-freebsd/  — Preceding unsigned comment added by 203.39.99.163 (talk) 04:05, 24 April 2013 (UTC)
 * it's a victory! Microsoft had a look on bsd when removing it from hotmail about 12 years ago... Ok, suppose you're the best fan of freebsd, and you acquire hotmail, running by that time on AngriestCoMostProprietaryAndTheWorstThingNotUnixOS. So, question is, do you realize that that system would run now your service long time after you purchased the service in any case, despite your fan level? If you do, then what your comment for? This is just ridiculous. When migrating they still used to run bsd. Amazingly! Who could even imagine that such a thing may happen! ... For every maker of the OS no matter corporate or comunity driven it is the very important principle - use what is made by himself. This is a prestige and reputation! I think freebsd project is hosted on the freebsd servers. And all fansites too. Why the same approach has to be wrong in the case of microsoft? Of course they have migrated to Windows and it is a normal thing. And of course this took some time. There is no any not ridiculous reason so actively point to this fact and moreover - include this in the article. This says nothing valuable about bsd, just about fans complexes.

77.52.154.34 (talk) 01:00, 15 July 2013 (UTC)

Timeline not rendering
I'm seeing this error: Timeline error. Command line was: '/usr/bin/perl' '/usr/local/apache/common-local/php-1.22wmf16/extensions/timeline/EasyTimeline.pl' -i '/tmp/timeline_59bf9bba7aca-1' -m -P '/usr/bin/ploticus' -T '/tmp' -A '/wiki/$1' -f 'FreeSans' Teslacuted (talk) 00:07, 18 September 2013 (UTC)

Timeline not showing FreeBSD 9.2
FreeBSD 9.2 has been in development for almost one month now, and release candidates are being built. Is it time already for the timeline graph to show it? src: FreeBSD 9.2 Release Process FChurca (talk) 20:16, 3 August 2013 (UTC)


 * FreeBSD 9.2 hasn't been released yet, cf. WP:CRYSTALBALL. I added a commented-out entry in preparation for the impending release.  DES (talk) 09:25, 19 September 2013 (UTC)

FreeBSD 8.4 not in timeline
FreeBSD 8.4 was released on June 7, 2013, and will be supported until June 30, 2015.

I would make the change myself, but I was not able to figure out how to edit the timeline.

tjameson 16:14, 16 September 2013 (UTC) — Preceding unsigned comment added by Tjameson (talk • contribs)


 * Fixed. DES (talk) 09:26, 19 September 2013 (UTC)

Reliable sources
Most of the references that used in this article are self-published and according to Wikipedia guidelines, self-published references are not reliable. I know FreeBSD documents are excellent, but they are not suitable for an encyclopedia. We need third-party resources. With the current references, the article can't be nominated as good or featured article. If someone knows any reliable, third-party reference, please suggest that. Replacing self-published references with third-party references can greatly increase the quality of the article. Bkouhi (talk) 16:07, 1 October 2013 (UTC)
 * Yes, the article needs secondary sources. AadaamS (talk) 16:13, 1 October 2013 (UTC)
 * A large number of reliable/third-party resources can be found here: FreeBSD in the Press, please help by adding them. Bkouhi (talk) 06:46, 19 October 2013 (UTC)

"Unix-compatible" in the intro
The intro contains the following line: Although for legal reasons FreeBSD cannot be called "Unix",[5] as a direct descendant of BSD (many of whose original developers became FreeBSD developers), FreeBSD's internals and system APIs are largely Unix-compatible. A few problems here: (1) This is previously unpublished synthesis, (2) claiming "Unix compatibility" is vague and meaningless, and (3) FreeBSD is descended from BSD, but all code from AT&T had to be stripped out. This means that the basic heritage of FreeBSD comes from BSD-specific code, along with code that was rewritten following the lawsuits (e.g. nvi replacing traditional vi). Huihermit (talk) 12:47, 31 January 2014 (UTC)
 * I've rewritten this part of the intro. Huihermit (talk) 16:50, 31 January 2014 (UTC)

Windows TCP/IP v4 stack and internet command line tools from BSD
Microsoft Windows incorporate TCP/IP v4 stack and all internet command line tools from BSD

@User:DagErlingSm%C3%B8rgrav you say that my sentence is "unreferenced and trivially demonstrated to be untrue", but the reference is the Wikipedia crosslink BSD I put, and the stack v4 is really from BSD. This is trivially demonstrated by the same command line options and syntax, the same packets network behavior, the same old ARP bug, and strings that you can find in binaries. All is well explained in BSD descendants and BSD Technology--Efa (talk) 11:48, 27 November 2013 (UTC)
 * Sorry, but your edit went way beyond what can be demonstrated and documented. Firstly, I don't think there is any BSD network code in Vista or newer.  Secondly, even in XP / 2000 Server, only a very small subset of BSD command-line tools was available; I believe nslookup was the only tool that was available in near-original shape, and even that probably came from ISC BIND, not directly from any BSD.  I'm not sure where they got rsh, but it was not included in the base system; it was part of Services for Unix or its predecessor.  In any case, none of this belongs in an article about FreeBSD unless you can document that Microsoft took code specifically from FreeBSD and not from 4.4BSD or one of its other descendants.  DES (talk) 13:33, 27 November 2013 (UTC)


 * A Wikipedia link is not a valid reference for other Wikipedia page, as WP:WPNOTRS clearly states, so "the reference is the Wikipedia crosslink BSD" does not suffice. "The same packets network behavior, the same old ARP bug" does not ipso facto demonstrate that the networking stack is BSD-derived; in this Kuro5hin article, somebody who claims that "[they] worked at Microsoft for ten years, most of it on the core Windows NT/2000 (hereafter referred to as NT) networking code" says that the original NT TCP/IP stack was from Spider Systems, who had a BSD-based stack implemented atop STREAMS, and that it was replaced by a new stack, rewritten from scratch, in NT 3.5. Guy Harris (talk) 20:41, 27 November 2013 (UTC)

I know, was Win9x, NT3.5, NT4 and Win2000 that included the code from BSD. From WinXP and ongoing they gradually replaced the code. Command line commands like 'arp', 'route', 'nslookup', 'ping' was exactly the same. Also Services for Unix is build in the same manner. The link kuro5hin.org is right, to me seems a valid reference, I cannot understand why you do not added the version of Win instead or removing the complete sentence.--Efa (talk) 07:34, 28 November 2013 (UTC)


 * The Kuro5hin post says that the rewrite was done for 3.5, not XP (5.1). It also says that the rewritten stack was used in Windows 95, hence probably in all Win9x releases.  They might have done more work to it between 3.5 and XP, and there was a "next generation" stack (whatever that means) in Vista.


 * Some of the commands might be BSD-based, but there's a huge difference between some BSD commands having been ported to Winsock and the in-kernel stack up to the transport layer being BSD-based.


 * And, even if the NT 3.1 stack was ultimately BSD-based, unless the code in question came from FreeBSD rather than from other another BSD source (unlikely, given that NT 3.1 came out in 1993, the same year in which FreeBSD came out), a claim to that effect isn't particularly relevant to FreeBSD, and is better placed on the Berkeley Software Distribution page. (I also need to go take a look at the XNU source to see which of it is specifically FreeBSD-based; the VFS layer, for example, is somewhat changed from that of FreeBSD.) Guy Harris (talk) 08:39, 28 November 2013 (UTC)

I'm sure the 9x and NT4/2000 stack was not rewritten (maybe only partially), contrary to what said the former employer, but I admit I have no valid (external) documentation link to point for this, was my personal experience with wireshark and protocol packets confirmed in BSD article. Most of the rewriting happened between XP and Vista, so NT5.1, anyway post 2000 (XP sometimes still show an old ARP bug). About command line tools, why you remove the sentence and not modified accordingly? Anyway I understood your point about BSD vs FreeBSD--Efa (talk) 21:02, 28 November 2013 (UTC)


 * "I'm sure the 9x and NT4/2000 stack was not rewritten (maybe only partially), contrary to what said the former employer, but I admit I have no valid (external) documentation link to point for this, was my personal experience with wireshark and protocol packets" That's insufficient to make me sure that the stack was not rewritten; it's not impossible for two separate network stacks to implement the same behavior, especially if the people writing one stack had seen the code for the other stack and used similar algorithms (which is not the same as using the code).  It's also original research in any case.


 * As for the command-line tools, I shall leave that to Dag-Erling Smørgrav to address, as he's the one who removed the sentence. As per his comment "In any case, none of this belongs in an article about FreeBSD unless you can document that Microsoft took code specifically from FreeBSD and not from 4.4BSD or one of its other descendants.", at least part of the reason is probably that "XXX came from some BSD" doesn't make XXX specifically relevant to FreeBSD.  It's also worth noting that the ipconfig command-line tool doesn't look particularly like any BSD tool, and is arguably a command-line tool for the Internet protocol stack, and thus might be a counterexample to a claim that "Microsoft Windows incorporate [ ... all internet command line tools from BSD." Guy Harris (talk) 21:21, 28 November 2013 (UTC)

"it's not impossible for two separate network stacks to implement the same behavior" simply not, lan timing and bugs are clear fingerprints. Instead your point about all command line tools counterexample is valid, but the remaining of tools sentence removal to me is a bad service--Efa (talk) 07:04, 2 December 2013 (UTC)


 * " simply not, lan timing and bugs are clear fingerprints". Or so you claim.  I think your claim is false. Give me one good reason to believe it is not false. Guy Harris (talk) 07:27, 2 December 2013 (UTC)

The Windows XP FTP utility still includes a BSD copyright notice:

$ strings ftp.exe | grep Regents @(#) Copyright (c) 1983 The Regents of the University of California.


 * Yes, Microsoft apparently believe that if it's not broken, you shouldn't fix it, and continue to offer a BSD-derived FTP command.


 * That doesn't say anything about the IP or TCP code being BSD-derived, however. Guy Harris (talk) 17:50, 31 January 2014 (UTC)

Neutrality
I feel that overall the article has a pro-FreeBSD tone about it. For example, "Characterised in 2005 as "the unknown giant among free operating systems"" - although it states it was in 2005, a single reference from one article 10 years ago with such hyperbolic language is not really suited to the introduction. "FreeBSD has several unique features related to storage." Unique? How? This isn't backed up. "No noticeable performance penalty over native FreeBSD programs has been noted when running Linux binaries, and, in some cases, these may even perform more smoothly than on Linux." Again, this relies on one (unverified) reference from nearly 10 years ago, which has no direct relevance to performance or compatibility since then. I think that it reads like an article to promote FreeBSD, which has been toned down here and there with additions. Teppic74 (talk) 15:26, 23 June 2014 (UTC)
 * Hi Teppic74, of course the article can be improved and you are welcome to do so. AadaamS (talk) 06:51, 25 June 2014 (UTC)

Hacked FreeBSD servers
Hi, is it worth to mention about hacked FreeBSD servers in 2012? ,,. It has gained enough media attention, I think -- Bkouhi (talk) 22:24, 5 August 2014 (UTC)

nomination for GA
Hi, is this still a "C-class" article? Or does it meet the good article criteria? What're the problems and drawbacks? Can someone please tell me what parts of the article needs more work? -- Bkouhi (talk) 00:23, 10 August 2014 (UTC)


 * Sorry, just saw this message. I think the biggest drawback is that the article is not consistent. E.g. some minor aspects are written in all broad technical details while other important things are missing. Over the last 10 years from time to time some feature has been added as the new hot thing, but later it was forgotten to adapt the paragraph when the feature became standard or even obsolete. For someone without BSD-Background it is quite difficult to read. Arved (talk) 08:02, 9 September 2014 (UTC)

Does this article really need a "Version history" section?
I think the "Version history" section is more suitable for the History of FreeBSD than this article. I suggest removing this section from this article and merging it into the History of FreeBSD article. I think only a table with brief description (like this) is enough for this article. What do you think? -- Bkouhi (talk) 14:09, 11 August 2014 (UTC)
 * I think a merge over to the History of FreeBSD is a good idea. AadaamS (talk) 15:55, 11 August 2014 (UTC)


 * I think this section should be a prose summary of the history of FreeBSD. Specifically, it should mention the first version, important releases and branches, the latest supported releases, and expected major releases. It should also mention important technology introductions. I think the version summary similar to that of the Dragonfly BSD article is excessive for this article. Mind  matrix  16:04, 11 August 2014 (UTC)
 * BTW: I agree that the bulk of this section should be merged to History of FreeBSD. Mind  matrix  16:06, 11 August 2014 (UTC)


 * How about a table like this? It is not a "prose summary", but I think a brief table like this with all major changes to the system is good enough for this article. This table is only an example and it may not contain all significant changes, but, it's a good starting point.

-- Bkouhi (talk) 19:32, 25 August 2014 (UTC)


 * The table is nice but has serious holes. e.g. from the current supported releases according to FreeBSD.org 8.4 and 9.3 are missing. Also mayor releases like 6.0 and 8.0 are missing Arved (talk) 07:52, 9 September 2014 (UTC)
 * What exactly constitutes the difference between a "Current version" (green background) and an "Older version, still supported" (yellow background)? It makes sense to me that release 10.1 is definitely a "current version".  Release 10.0 is currently yellow (older reelase still supported) but the end-of-life date suggests that it isn't actually supported anymore.  The status of releases 9.3 and 8.4 are much less obvious to me.  Clearly they are still supported, but the FreeBSD project classifies them (according to http://www.freebsd.org/releases/ ) as "Legacy".24.222.2.222 (talk) 11:36, 30 March 2015 (UTC)


 * I've updated the table to match the table at History of FreeBSD. Mind  matrix  13:17, 30 March 2015 (UTC)

Security - irrelevant information and HardenedBSD promotion
Under FreeBSD, there is a section filled with HardenedBSD ASLR promotion with only primary references. Furthermore, the ASLR upstreaming to FreeBSD by HardenedBSD has been abandoned by it's developers. https://reviews.freebsd.org/D3565#97806

There is a new ASLR project in place, sponsored by the FreeBSD foundation, which should be upstreamed to FreeBSD. However, until it is complete, it's really not relevant nor ASLR needs to be mentioned on the Security section of the page. https://reviews.freebsd.org/D5603

There is simply no need for mentioning HardenedBSD or ASLR on FreeBSD page. I am intentionally starting a talk about it, because I believe consensus is necessary prior to removing this section. Specifically, consensus is needed because I suspect edit war will start otherwise. For example, under Address_space_layout_randomization you can see unverified and primary source explanation that HardenedBSD ASLR project was not upstreamed to FreeBSD "because of unwillingness of FreeBSD developers to include the proposed ASLR implementation" which is complete falsehood if you see the link above (about project being abandoned by HardenedBSD) Mr.hmm (talk) 01:22, 19 June 2016 (UTC)


 * I agree. I've rewritten the ASLR:FreeBSD section to be more neutral (before I saw this Talk section). KMeyer (talk) 19:31, 6 October 2016 (UTC)

Derivatives — Embedded device operating systems list
Not sure this list of commercial products is noteworthy. I think instead perhaps we should just link to List of products based on FreeBSD. KMeyer (talk) 18:10, 12 December 2016 (UTC)

Reliability
This operating system looks much more reliable than competitors. Typical uptime of production machine should be listed. — Preceding unsigned comment added by 178.41.41.56 (talk) 16:41, 2 March 2017 (UTC) version 5 : 2222 days — Preceding unsigned comment added by 193.87.32.16 (talk) 11:51, 4 March 2017 (UTC)

Prominent patriostistic flags should be avoided world-wide
The screenshot with the big flag is inappropriately out of topic. I do assume that it was put there *on purpose* or the guy who did the screenshot has been living for so long in an exaggerated patriotic community that he doesn't *think* anymore. — Preceding unsigned comment added by 82.236.187.110 (talk) 11:32, 23 December 2017 (UTC)

Really? The screenshot includes a photograph of one of the contributors of FreeBSD and a FreeBSD foundation board member on a FreeBSD desktop and it took me 5 minutes to even figure out what flag you were even talking about. 50.77.48.113 (talk) 06:26, 15 February 2018 (UTC)

The “patriotistic” flag, whatever that is supposed to mean, is actually an Xfce tray widget that indicates the current keyboard layout. I know because I created the port in 2004. It looks huge because nobody in the Xfce project has a single clue about graphic design. DES (talk) 18:48, 17 February 2018 (UTC)

Violation of the Unix trademark?
The text of the article currently states:
 * FreeBSD is a direct descendant of BSD, which was historically called "BSD Unix" or "Berkeley Unix" (in violation of the UNIX trademark).

This sounds wrong to me. BSD Unix was Unix. It shared the code base. It wasn't until the AT&T stuff was excised in the early 1990s that BSD lost the right to be called Unix. Does anybody disagree? I plan to remove this claim of violation. Groogle (talk) 07:13, 15 January 2019 (UTC)


 * As the saying went, "Unix is a trademark of Bell Laboratories", later "...a trademark of AT&T Bell Laboratories" and, still later, one of the "registered trademarks of AT&T" listed.


 * So Bell Labs, later AT&T, got to decide to whom to license the trademark. There was a time when they only allowed the trademark to be used for systems with few differences from the code they sold; "shared the code base" wasn't sufficient.  (When I was at Sun, in the late 1980's, we changed the name printed in the OS banner from "Sun UNIX 4.2BSD" to "SunOS", at AT&T's request, when we started working with AT&T on what ultimately became SVR4, due to the trademark issue.)


 * So BSD UNIX "was Unix" in the sense of being derived from the AT&T code base, but that didn't grant the University of California, Berkeley, or any vendor who based their OS on BSD UNIX, a right to use the Bell Labs/AT&T trademark.


 * However, AT&T then sold the code, and the trademark rights, to Novell; Novell later sold the code to SCO and the trademark rights to X/Open. Later X/Open merged with The Open Group; The Open Group now own the trademark. They license it for use with any system that passes their validation suite for the Single Unix Specification, regardless of how much AT&T code, if any, was used in that system.  (NetBSD explicitly say they're not UNIX - not because they have little AT&T code left, but because they haven't been licensed for the trademark.)


 * So if somebody were to take FreeBSD (or NetBSD, or OpenBSD, or DragonFly BSD, or pick-your-favorite-Linux-distribution, or...) try to run it through the Single Unix Specification validation suite, fix the problems that appeared, try to run it again, lather, rinse, repeat..., they could eventually produce something for which the "UNIX" trademark would be licensed, making it a "UNIX(R)", regardless of how much AT&T code remained. In fact, an organization did take an OS with a kernel based on a combination of Mach and BSD kernel code, and a lot of BSD userland code, and go through that very cycle, and produced something that can have the UNIX trademark applied to it (and have continue to do so for all releases since the first one for which that was done).


 * And AT&T did, in fact, sue the Regents of the University of California and Berkeley Software Design, Inc. for, among other things, trademark issues. At least as I read the court's opinion (usual disclaimer: I Am Not A Lawyer), they didn't completely dismiss the trademark complaints of AT&T, so I'm not sure whether AT&T's claim to a right to decide what's a "UNIX(R)" was rejected. Guy Harris (talk) 09:12, 15 January 2019 (UTC)

Which version was the first, free from AT&T code?
I found it confusing in the history section of the article. The 1991 (Net-2), 1992 or 1994 after the lawsuit was the first release without AT&T code? — Preceding unsigned comment added by Siavoshkc (talk • contribs) 14:45, 24 May 2019 (UTC)
 * Net-2. Its predecessor Net-1 was hacked to remove AT&T code, but six files remained that contained at least some AT&T code. These were excised, and the resulting code was released as Net-2. Later, 386BSD was released containing (among other changes) replacement copies for those six excised files. Mind  matrix  16:19, 11 November 2019 (UTC)

Trimming "Version history"
Is anyone opposed to trimming the version history section? All this information is contained in FreeBSD version history, and having so many older versions listed clutters the article, in my opinion. In a previous discussion, I stated that I preferred having a brief prose section mentioning "the first version, important releases and branches, the latest supported releases, and expected major releases". I'd like to trim off all versions earlier than 10.0 from the table, and add a short blurb of text as a lead for the section. Mind matrix  14:44, 22 June 2018 (UTC)
 * I just did so per this suggestion and the one in the GA reassessment. I'm sure this could be whittled down much more as I simply merged the major versions together and could remove some of the more minor bullet points. I could remove the ones prior to 10.0 as you suggest, but should probably at least give a brief mention to the ones before. However, even the change so far resulted in the size of the page going down more than 5 KB. Tonystewart14 (talk) 18:21, 28 April 2020 (UTC)