Wikipedia:Articles for deletion/Xmonad

The following discussion is an archived debate of the proposed deletion of the article below. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review). No further edits should be made to this page.

The result was Discounting the SPAs and "this is useful" arguments, whats left is No Consensus. —Ocatecir ^Talk 03:21, 28 May 2007 (UTC)

[edit] Xmonad

Xmonad (edit|talk|history|links|watch|logs) – (View log)

Non-notable window manager written about a month ago. Was spammed all over the internet, but now that all the Ooh's and Ahh's are over, it seems it doesn't belong on WP after all. Estimated userbase: <25 and unlikely to gain much more. Catofax 10:01, 15 May 2007 (UTC)

• Delete Vaguely reliable source here, but I'd say it falls under the category of "trivial", consisting as it does of less than 3 lines of text. JulesH 15:45, 15 May 2007 (UTC)
• Comments

I know I can't vote, but the number of lines don't shows that something are trivial, specially in functional programming, when less is more. I am not sure if this work is really notable, but it's a great piece of software. I also think the user base will increase with the time, so wp-en should not be so greedy to delete. --189.12.138.73 05:38, 17 May 2007 (UTC)

• Comments The window manager is particularly interesting due to its use of a purely functional programming language, Haskell, and semi-formal methods to ensure correctness (notably, QuickCheck, and an unusual 'zipper' data structure internally) -- a first for a window manager. Due to these interesting implementation techniques, it can be radically distinguished from other window managers.
• Articles and coverage since the 0.1 release
• A recent article on its design
• The window manager has been the subject of a mechanical proof, by an external researcher, for correctness of the window manager's internal logic, a significant reference.

It is also unclear how catofax reached the user base conclusion of "<25 and unlikely to gain much more.", as the irc channel alone has grown to 30 in the last week, and the mailing list to over 60 users. Without stronger justification, and given the application being referenced as a subject of research, growing discussion and analysis of it by external parties, and the technical innovation of the application itself, deletion seems particularly unwarranted. Finally, similar, yet less technically interesting window managers, such as Dwm happily have entries on WP. 220.233.48.34 14:53, 17 May 2007 (UTC)

The applications of QuickCheck and Catch are not interesting. The window manager is trivial, and a fully developed human being with moderate programming experience should be able to verify correctness without resorting to convulved automatic checking. The application of the Zipper structure is not interesting either - since the amount of data is small the updates could be O(x^x) and there would be no performance hit. XMonad cannot be radically distinguished from other window managers - it is essentially dwm rewritten in Haskell but with even less features.

The #xmonad population consists largely of lurkers, including dwn, dmenu, dzen, ion developers, most of which do not use XMonad. Moreover, blogs are not press.

Finally, I note that writing a window manager is sort of a rite-of-passage in unix circles, and that many of the window managers with pages on WP do not in fact belong on WP, and I intend to prune the list as time allows.

Catofax 09:35, 18 May 2007 (UTC)

Disagree with your assertions. Most channels have a lot of lurkers. The use of QuickCheck has caught numerous bugs; the user of zippers has simplified Xmonad considerably - before, the devs had bumped the LoC limit up to 550 to fit everything in, but with the zipper changes, it's back down to ~500 (so it's the first real use of zippers that I know of and the use had nontrivial effects for the codebase). And if correctness is so very easy for any decent programmer - why do so many window managers keep having bugs? For such a simple task, as you seem to think of it, it has tripped up a surprising number of programmers. --Gwern (contribs) 16:59 18 May 2007 (GMT)

Disagree. The Zipper is not used for performance reasons -- but to embed window manager focus behaviour directly in the data structure, leading to simpler code, and code that is easier to automatically test -- a main development goal of the project. The automated testing and proofs are novel and critical to the development project, as they test a side-effect heavy program like a window manager, and continue to catch bugs as the code is developed, ensuring the unstable branch is more stable than in similar projects. The Catch proof alone is significant: there are simply no other similar projects with automated proofs of their behaviour, and such levels of assurance. Xmonad is entirely unique in this respect. 220.233.48.34 00:10, 19 May 2007 (UTC)

While not relevant to the question of the merit of the page under discussion, Catofax's unsourced assertion that the IRC channel is mostly lurkers disagrees with the only available evidence, the logs, which indicate of the 31 irc users, 21 have written more than 30 lines of text in the last 4 weeks, and 51 unique users have contributed text. The second assertion, about the number of users of the application, is entirely unverifiable: it is simply impossible to know. Unsourced assertions by Catofax regarding user base, and glaring misunderstandings of key contributions documented in the article, such as the the zipper or correctness proofs, only emphasise that the nomination was made in bad faith, based entirely on incitement from the 4chan forum, ["Haskell bullshit on Wikipedia"]. 220.233.48.34 04:36, 19 May 2007 (UTC)

• I find this entry useful and informative. It is humorous to imply a lack of reliable sources when everything stated can be directly verified from the source code itself.

TomMD 15:14, 17 May 2007 (UTC)

• Comment, as the author of the article, I've got to agree with the anons. I wrote on Xmonad because it's one of the more popular pieces of Haskell software (especially given how young it is), that it was intrinsically interesting. I've edited the article and hopefully more fully brought out the subject's points of interest and novelty. I don't think it lacks notability any more than do dwm, larswm, StumpWM, etc. and I'm disappointed to see a more than usually interesting one singled out for AfD.

I particularly commend to the closing admin a reading of this 4chan thread, "Haskell bullshit on Wikipedia", and a look at Catofax's contributions. In short, this looks to me like a bad-faith nomination by a single-purpose throwaway account. --Gwern (contribs) 22:48 17 May 2007 (GMT)

I have been using XMonad from the onset on my laptop and think it is rather neat. I also think it is not notable enough to be included in WP, and take issue in the way it was spammed everyhwere. I think this AFD soft of demonstrates the non-notablity with the lack of votes: nobody seems to care either way!

While this is indeed a new account, because I do not wish to be identified in the Haskell community (otherwise nice people) it is not necessarily a throw-away account (like always, that depends on my level of disgust when this finished.) I did try to do this accountless, but WP would not let me. Catofax 09:35, 18 May 2007 (UTC)

Most AfDs don't get many votes. If it gets as many as ten, it's doing very well. Also, you can't create an AfD as an anon because page creation for anons was disabled back during the Seigenthaler incident. --Gwern (contribs) 16:59 18 May 2007 (GMT)

• Keep I think that functional programming is an important research domain in computer science, and that systems-level functional programs are often important demonstrations of what can be done. The number of regular users is not the point - the research is the point.Bhimaji 23:02, 17 May 2007 (UTC)

• No other window manager on the site is up for deletion, including crap like XWEM. What's this really about? 209.149.58.156

Relisted to generate a more thorough discussion so that consensus may be reached
Please add new comments below this notice. Thanks, W.marsh 14:23, 22 May 2007 (UTC)

• Keep This article presents useful information about a window manager that is doing a lot of noatable things (zippers, functional programming, automatic testing, formal verification) - at the same time as being used by people for real tasks. Responding to Catofax: the suggestion that "automatic checking" can be "convoluted" (high complex to who? its automatic!), and that a "fully developed human being" can "verify correctness" shows a lack of understanding as to the current state of computer science. --NeilMitchell 16:04, 22 May 2007 (UTC)

• Keep (or just possibly merge). Xmonad's launch was mentioned in a short article on OSNews (see http://www.osnews.com/story.php/17744/Xmonad-a-Tiling-Window-Manager-Written-in-Haskell) which suggests that it is notable. The content is informative and probably deserves its own article. However, if another editor comes up with a good suggestion on merging, then I would consider it. Greenshed 22:07, 22 May 2007 (UTC)
• Delete or rewrite As it is now, the article contains no information as to why this is somehow a notable program. If someone can explain in the article why this is notable, then I'll vote Keep. Elrith 00:13, 23 May 2007 (UTC)

• Comment: "This emphasis makes Xmonad unique in a number of ways; besides being the first window manager written in Haskell, it is also the first to use the zipper data structure for managing focus, and its core has been proven to be safe with respect to pattern matches...." --Gwern (contribs) 00:41 23 May 2007 (GMT)

Yes, and are these things somehow notable? If they are, this needs to be somehow explained in the article. Elrith 00:51, 23 May 2007 (UTC)

The use of Haskell is notable as it is one of the first non-academic projects with a focus on practicality to make use of Haskell. The pattern-match checking means that XMonad has a formal proof that its core will never crash - something which makes it very notable indeed. Perhaps a section on the formal proofs, testing and derivation strategies would improve the article and bring out the unique aspects.--NeilMitchell 01:29, 23 May 2007 (UTC)

That's rubbish, and you know it. There are plenty of such projects. Permit me to name a few: Conjure, FRAG, darcs, hmp3, lambdabot, HAppS, TagSoup, Cabal... These are the ones I've used, and there are a lot more [1]. Futhermore, application of Catch (and I note that you are the author) may or may not (who verifies the verifier?) catch inexaustive pattern match errors - XMonad can crash for any number of other reasons that have nothing to do with this (and in fact I've had it crash once, which is one more crash than other window manager as far as I can remember).Catofax 09:10, 23 May 2007 (UTC)

Darcs meets the criteria I mentioned, as does HAppS. Lambdabot might, but I don't think any of the others do - and I wrote one of the things on that list :-). I am indeed the author of Catch, and the question of verifying the checker is a difficult one - the intention is to generate input for a theorem prover, along with a semantics of Haskell and a formal proof that Catch is correct. If you want to know more about Catch either email me or go on the Haskell IRC channel (I'm ndm), as this isn't really the place to go into detail on it.--NeilMitchell 20:14, 23 May 2007 (UTC)

• Keep The article goes far deeper in technical detail than any similar project's article (see ion, dwm, wmii), provides many useful references, and does a good job of highlighting the novel aspects (formal methods, automated testing, unusual data structures, extensibility via Haskell). 220.233.48.34 01:55, 23 May 2007 (UTC)
• Delete if we have any principles besides ILIKEIT. There are zero third party sources, reliable or otherwise. It's being kept not even on the basis of OR--on the basis of personal belief based on personal experiences that it is an important program.

Neil, reference one in the article is from your own blog.

Reference two is from the program manual, and that is all the references there are.

The first group of external links is the program's own site, divided up into homepage and press releases

The second external link is from the programming language site, the third is a blog with a long thread of personal experiences with the product

The fourth and fifth are descriptions from the guy who wrote the program, and the sixth is a newsletter for those who use the programming language.

This is the epitome of COI: the subject is important because we who are working with it say so, and we've said so before in the blogosphere, and if that's not enough we'll explain right here at AfD just why it's important--

As the guideline on self-written autobiographies says, if it's important, someone else will say so. When I've argued before that blogs were acceptable sources for internet things, I didn't mean that one's own blog was acceptable as a source for one's own program. If such are the standards, all the Arbuthnots are notable because kitty has a web site on them, and because some members of the family have compiled books about each other. All the Louisiana politicians are notable because Billy wrote a thesis about them which he can now cite, and has recorded where their tombstones are.

Ironically, just as perhaps some of the Arbuthnots may actually have done important things, but KB hasn't done the research to find out, this may be important and there may be discussions from those who are not themselves working on or with the program--but the article makes no attempts to find them. A nice looking windowing program like this might well have been noticed in the trade press. DGG 03:54, 23 May 2007 (UTC)

Under what criteria are any of the window manager pages on WP suitable? Barely any provide external references, other than their own web sites. The Xmonad article would seem to be the most well referenced, citing the OSnews article, the HWN article, and Neil's work on pattern match verification (who is neither a user nor developer of xmonad) -- so it is odd that it is singled out, trying as it does to provide a useful, referenced article. See for comparison pages on dwm, wmii, Ratpoison, TrsWM, XWEM -- all cite merely the project's own web page. Why delete the most thorough article in its class? Should the technical/blog references be reduced? 220.233.48.34 04:29, 23 May 2007 (UTC)

I've rewritten some paragraphs to refer to the external sources directly. I hope this eases concerns. 220.233.48.34 05:08, 23 May 2007 (UTC)

The external sources are not independent, and so this does not erase the concerns. The people who work with the program telling each other about it. An article on other subjects with sources of this nature would be rapidly deleted without much argument. WP seems to be asked to make an exception on this subject, because of the acknowledged technical competence of the various editors here.

It could perhaps reasonably be argued that these are the only available sources, that knowledge of such programs is diffused in this manner, and that the importance is shown by the impressiveness of the work itself. I'd be willing to accept such a completere-orientation of the notability rules if we accepted this for all phenomena and projects that have similarly blog-based and self-publishing sources. I am open to the argument that anything adequately documented on its own terms should be included--I might even support it. What I do not accept is that it should apply only to this subject area DGG 02:09, 24 May 2007 (UTC)

Yes, I quite like this analysis. The article in question makes a good-faith attempt to refer to the available material on its subject, and for web-based subjects like this, that often means internal documentation, blogs, and maybe trade articles based on blogs. I'd imagine many software projects already on WP are in fact, in this class? 220.233.48.34 03:09, 24 May 2007 (UTC)

How exactly is this any different than any other software product particularly one that is prerelease. The Windows Vista article was started 19 September 2003. What in your argument wouldn't have applied then? jbolden1517^Talk 09:49, 25 May 2007 (UTC)

Keep -- This one is clear cut. I admit it probably got listed too soon. However I'm getting 44600 google hits on it. And I when I check they are genuine articles. Reviews, people excited about future versions, a tutorial on X window manager user xmonad as an example. And this is for a 0.1 version! I think we keep and revisit in 2-3 years. We have no idea of this thing pans out but at this point its notable enough and obviously becoming more so quickly. jbolden1517^Talk 16:51, 24 May 2007 (UTC)

I admit that it was spammed surprisingly well, but it's still trivial and non-notable, and does not (yet) belong on WP by a large margin. Catofax 06:17, 25 May 2007 (UTC)

• Delete In case anyone's counting. Catofax 06:17, 25 May 2007 (UTC)
  • Comment. I think as the nominator it's safely assumed you voted delete. :) --Gwern (contribs) 14:15 25 May 2007 (GMT)
• Comment The article's changed quite a lot since I originally !voted 'delete'. The program in question now appears to be an interesting academic excercise, even if it isn't particularly notable as a window manager. However, I'm still tending towards deletion. My primary reason for this is that all of the sources other than the OSNews article are self-published, and aren't reliable sources. Specifically, the interesting aspects of this project can only be verified from (1) the project's own web page or (2) the blogs of two students, one of whom is one of the project's authors and is therefore not independent. My suggestion: delete for now, bring the article back if and when something more substantial (e.g. a conference paper) is published on the subject. JulesH 13:54, 25 May 2007 (UTC)

It will definitely appear in my thesis, in some detail, by Christmas. I suspect it will also appear in a conference paper before then.--NeilMitchell 22:45, 25 May 2007 (UTC)

Keep -- original submitter of deletion request is exercising some agenda based with no basis

I think you will find my basis to be quite orthonormal and sturdy :) Catofax 11:15, 27 May 2007 (UTC)

It is interesting that you as nominator have moved from "it doesn't belong on WP" to "does not (yet) belong on WP". The 4chan thread where you tell the mob you have AfD the article, along with the fact no other wm pages are up for deletion, makes your motivation highly suspect.

Stop grasping at straws. Here's what's what: (1) I'm not clairvoyant. XMonad may belong on WP in the future. (2) It's just me--I've never told anyone about this AfD. See these people voting keep? They're mostly associated with XMonad somehow. See the people voting delete? They're random Wikipedians. (3) XMonad is probably the least notable WM on here. If it stays there's no reason to AfD the others. If it goes, I'll start adding the others too. Hope that clears things up! Catofax 20:56, 27 May 2007 (UTC)

I can safely attest I have no idea how to write a window manager. jbolden1517^Talk 00:29, 28 May 2007 (UTC)

I voted 'keep' and I don't know anybody associated with the project. In fact, I've never written anything in Haskell before. You're not going to AfD the others because, in your opinion, they're more notable but still don't belong? Strange reasoning. If they don't belong, AfD them and clean up as much as possible. Maybe you'll find that enough people disagree with you on the other WMs that everything stays. Bhimaji 22:23, 27 May 2007 (UTC)

"probably the least notable WM on here". Have you even looked at the full list? For example, the array of window managers with single paragraph entries, pointing at their own website? It would appear xmonad is the best referenced, and one of the few that attempt to seriously justify notability. If notability was really your concern, you'd have nominated one or all of the obviously trivial articles such as Wm2, EvilPoison, HaZe or JWM. But you didn't. The xmonad article appears quite reasonable, and interesting!, by comparison. 220.233.48.34 00:46, 28 May 2007 (UTC)

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review). No further edits should be made to this page.