Jump to content

Talk:FreeBSD/GA1

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

GA Review

[edit]
GA toolbox
Reviewing

Article (edit | visual edit | history) · Article talk (edit | history) · Watch

Reviewer: Protonk (talk · contribs) 17:11, 10 September 2014 (UTC)[reply]


Overview

[edit]

In general the article is pretty good. It covers an evolving, technical topic in a relatively straightforward manner. The lede summarizes the article fairly well and there are no major problems with images, sources or claims in the content. As an aside, I'm now very happy I've discovered How does one patch KDE2 under FreeBSD?.

I have a few concerns which I'd like to see addressed before I promote this article.

First, the sourcing in certain sections (e.g. History and Security, but others as well) is relatively thin, either relying on a single source, using broad "about me" sort of pages as a source for specific claims which are probably true but otherwise uncited or otherwise leaving it unclear to the reader where statements are supported by sources. I've listed some specific issues with sourcing below, but that assessment isn't exhaustive.

Second, there are a number of copyediting errors throughout the article. This stuff is hard to catch (I didn't even list them all), but issues with subject/verb agreement, pluralization and some others all exist. Most of these should be easy to fix and I don't expect that all of them be fixed before the article is promoted, just that a proponderance get some attention.

Third (and this is relatively minor), a number of the sections and paragraphs are somewhat awkwardly worded, probably from editors working to insert or improve specific sentences within sections without re-evaluating how the section reads after the change. I didn't dig through the history to prove this, but that's usually the case with very old wikipedia articles (and this article predates our current version history! It's more than 13 years old).

I think if we fix the bulk of the above problems I can promote this article without too much trouble. I've detailed some specific comments below. Feel free to reply to them inline or strike them as you go along, whichever works best for your flow. I've also suggested (but will by no means require) reorganizing the article to better reflect the relative importance of certain facets of FreeBSD. If doing so helps you fix some of the other problems (I think it might) feel free to reorganize it as the review goes along. I'm happy to follow along. If you think it'll get in the way then just take it as a suggestion from an outside editor and either ignore it or implement it on your own time. Protonk (talk) 19:37, 10 September 2014 (UTC)[reply]

Caveat: I'll use the term BSD to refer to FreeBSD because it is less typing. :)
Hi Protonk. Thank you very much for this very useful and detailed review. Still working on the article, but until now, these changes were applied to the article. -- Bkouhi (talk) 04:30, 11 September 2014 (UTC)[reply]
@Bkouhi: I've added some follow-up comments including some suggested sources for places where I feel the article can still be improved before it is passed as a GA. I think once some of those comments have been addressed and I make another pass for copyediting I can conclude the review. Protonk (talk) 16:26, 12 September 2014 (UTC)[reply]

Style/layout

[edit]
  • This is a matter of opinion, but an article this big should have a less flat hierarchy for sections. The features section has 11 subsections. This is perhaps too many.
  • Uses should be folded into another section or expanded (more on this elsewhere)
  • Some terms are wikilinked too much, e.g. Walnut Creek CDROM (three times in the history section)
  • Images are tough for articles like this, but I suspect there's a better place for (at least) the KDE SC image than merely at the top of Features
  • Don't use inline links in the article (e.g. the link to Frenzy in the list of derivatives)
  • Might the Version history section be better suited as a child of the history section?
  • I agree that it is better to have only one History section but IMHO, putting that large Version history at top of the article is not a good idea, on the other hand, I think the "History" section should be placed before the "Features" section. -- Bkouhi (talk) 04:30, 11 September 2014 (UTC)[reply]
  • "Also, FreeBSD supports IPX and AppleTalk protocols, but they are considered old and it is planned to drop support of them as of FreeBSD 11.0." this is awkwardly worded
  • the paragraph beginning with "The project has also ported the NSA's FLASK/TE implementation from SELinux to FreeBSD." seems like it would be better placed after the paragraph on how TrustedBSD components are folded in to other OSs
  • There are two hatnote links to FreeBSD Ports in successive sections. Perhaps we only need one?
  • Sentences like "FreeBSD's development model is further described in an article by Niklas Saers." should be avoided. Imagine a reader looking at a PDF version of this article or reading it via Wikipedia Zero. They won't be able to read that article so the sentence itself will have little meaning to them.
  • Should Unauthorized access to FreeBSD's servers be moved to a subsection in History?
  • IMHO, it may be too soon to mention this in the History section. I think the reader must have a clear background of FreeBSD project first. Aside from that, mentioning this in the History may imply that this is a very important event in the History of FreeBSD, which IMHO it is not, however it has widely covered by press. -- Bkouhi (talk) 22:22, 12 September 2014 (UTC)[reply]
  • "...by Stealing SSH keys from FreeBSD's developers" aside from the capitalization problem, SSH links to SSH key. If this is intended you can just type [[SSH key]]s to get SSH keys.
  • Copyediting problems such as "The user can load and unload this modules at any time." appear throughout the article. Pluralization seems to be a recurring problem, see also "Among the most popular mailing list[s]" and "official documentation" vs. "official documentations".
  • "FreeBSD's official documentation consists of its handbooks, manual pages, mailing list archives, FAQs and a variety of articles, mainly maintained by The FreeBSD Documentation Project; they are available at FreeBSD's homepage." This is sort of a clunky sentence. Also are the mailing lists among the "official documentation"?
  • I removed the word official, but since I'm a language learner, not a native speaker, I may unknowingly make grammatical errors and particularly, this sentence looks OK to me, could you please provide an alternative? -- Bkouhi (talk) 22:22, 12 September 2014 (UTC)[reply]
  • "Moreover" is used twice within 3 sentences in the Logo section
  • "Only support 64-bit (V9) architecture" "support" isn't needed here
  • "It was written in C by Jordan Hubbard, is curses based and first appeared in FreeBSD 2.0." I'd move this line up (and perhaps reword some sentences around it) and move the following lines on the deprecation of sysinstall and the adoption of bsdinstall to a new paragraph.
  • "annually-held conferences" just "annual"
  • The Version history section contains a hatnote to History of FreeBSD but the History section does not.

Content

[edit]
  • The "Commercial" paragraph in Uses is a bit out of place with the remainder. Surely commercial uses of BSD would fall into servers, desktop or embedded systems. If we're looking at other use cases (e.g. using a particular component of BSD in a piece of software) then they should be noted and cited.
  • Nearly half of the lede is devoted to summarizing a single thought: that BSD contains the entire OS in a single repo.
  • The second paragraph in History is a bit hard to follow. Was 4.4BSD-Lite an unbootable operating system because it lacked AT&T code? Were the versions of FreeBSD between 1.1 and 2.0 unbootable because they relied on 4.4BSD-Lite?
  • We mention twice that BSD powered Walnut Creek's servers. Perhaps we're better off leaving that mention in the last paragraph and removing it from the list with Yahoo and Hotmail.
  • Surely Apple's adoption of BSD for OS X merits a mention in the history section. More generally, there's nothing in there post 2000.
  • "Much of this work was sponsored by DARPA under DARPA/SPAWAR contract N66001-01-C-8035 ("CBOSS")." Much of what work? TrustedBSD? OpenPAM?
  • Is there a reason why Tier 4 archs are listed on the portability tables? The cited source basically just says that anything not listed as tier 1-3 is tier 4 and unsopported. I see xbox here but I'm not convinced we need to include it in the table (I'm willing to be convinced, however!).
  • "This compatibility layer is not an emulation..." this sentence should be higher in the paragraph, probably before noting the caveats to the compatibility layer
  • What is an M:N threading model? Why is it the best in theory? How long was it in use in BSD before switching to another one?
  • I made a wikilink to the corresponding article and fixed the date. This is exactly what Michael Lucas said in his book, should we really mention why this model is the best? -- Bkouhi (talk) 04:30, 11 September 2014 (UTC)[reply]
  • It seems odd to say it is best in theory without providing a cite or a reason. I've dug around and found this conference paper on the switch and this very old paper on the scheduler in general. There are a few others but most of them which are specific to M:N -> 1:1 are student papers hosted on the web. Protonk (talk) 13:33, 11 September 2014 (UTC)[reply]
  • @Bkouhi: I've passed the article but I would like your input on the suitability of the above sources for the article. I think the threading issue is kinda interesting because BSD switched around the same time that other kernels switched and it seems like a lot of operating systems are moving toward the 1:1 model. Protonk (talk) 14:34, 13 September 2014 (UTC)[reply]
  • Hello and thank you very much for the review and passing the article. Interesting papers, but unfortunately there is little or no mention of FreeBSD. Scheduler.pdf does not talk about FreeBSD at all (although it is hosted by FreeBSD's website), while paper_79.pdf only says "FreeBSD uses 1:1 threading model". Wouldn't it be WP:OR to use these papers? I always thought that references must explicitly talk about the subject (FreeBSD), otherwise it would be OR to use such a reference. Is this right? -- Bkouhi (talk) 22:21, 14 September 2014 (UTC)[reply]
  • @Bkouhi: Not mentioning FreeBSD was a big stumbling block on most of the papers I found, unfortunately. :( If we're trying to support the claim that M:N threading works best in theory (which does need a source) then the theory can be independent of BSD. I'd prefer a solid paper on why BSD made the switch, but I don't think it is too far out of bounds when making a claim like that to cite a general source. There are only some cases where this works and sometimes it comes down to your judgment. If you're not comfortable citing it then I'd recommend simply noting that BSD used to have a M:N threading model and switch to a 1:1 threading model without making claims in the article as to why one is preferred to the other. Does that make sense? Protonk (talk) 22:38, 14 September 2014 (UTC)[reply]
  • "ClangBSD became self-hosting on 16 April 2010, an important landmark for further independent development." Why is it important? Who says?
  • Is there a reason we don't also embed the new logo in the logo section?
  • "FreeBSD's slogan is "The Power to serve" which is a registered trademark for The FreeBSD Foundation." how about "FreeBSD's slogan, "The Power to serve", is a registered trademark of The FreeBSD Foundation."?
  • The list of derivatives could probable be cut considerably (retaining the pointer to List of products based on FreeBSD), noting only the most significant or important derivatives.
  • "As of 2013, there are plans for supporting them in FreeBSD 10." It's 2014 and BSD 10 is out. What happened to those plans?
  • "For a long time, FreeBSD used the sysinstall program as its main installer." How long?

Sourcing & such

[edit]
  • The citation needed tags on FreeBSD contributions to other operating systems need to be cleared up.
  • Most of the History section is cited to a single source, the handbook. That's not necessarily hugely bad but it's something of a red flag for me. I'd like to see more sources included.
  • The security section is largely unsourced or dependent on a single source (basically the about page for the TrustedBSD project). At the very least I'm unclear where claims are cited to specific sources.
  • This is the hardest part, I added a source for ACLs, but I'm unable to find any reliable, third-party source for this section, however, I'm still working on this. Isn't that "about me" source enough? Please don't take this into account if I couldn't find any source. -- Bkouhi (talk) 04:30, 11 September 2014 (UTC)[reply]
  • It's probably enough for a shorter section, but I think relying on that source causes two problems. First, it makes it hard for us to contextualize claims without OR. How important is TrustedBSD to FreeBSD? How important are any of the individual elements to the trusted BSD project? It also makes it hard for us to choose between different claims on the page when editing. Is "Much of this work was sponsored by DARPA under DARPA/SPAWAR contract N66001-01-C-8035 ("CBOSS")." an important claim to have in the encyclopedia? I don't know (I suspect not) and it's hard to tell definitively without another source on the matter. Is there a conference paper on Trusted BSD or any of the components in it available? A talk at one of the bsdcons about implementing it? Etc. Protonk (talk) 13:43, 11 September 2014 (UTC)[reply]
  • So I've found this paper (citation info here), this talk (It's paywalled but I've asked around to see if anyone has a copy), also this talk from eurobsdcon, which doesn't appear to be published anywhere but Robert Watson (computer scientist) is probably an SPS on the subject. Protonk (talk) 15:11, 11 September 2014 (UTC)[reply]
  • So that usenix talk is here (html version). It was paywalled until 2004 but I misread the page. Protonk (talk) 17:14, 12 September 2014 (UTC)[reply]
  • "Through the years Beastie became both beloved and criticized as perhaps inappropriate for corporate and mass market exposure." Says who? I don't doubt that it's true, but we should find a source for that claim.
  • "The FreeBSD Project is unusual among open source projects in having developers who have worked with its source base for over 10 years before its release in 1993..." is there a source for this claim? The whole section is cited to the project admin page which appears (to me) to be just a directory of team members w/ roles noted.