Most of this was written 25 November 2015. I've started to provide nutshell summaries (one section so far) in a first step toward compression to a "PowerPoint" version, and have also updated some passages to reflect post-2015 happenings. Timestamp of my most recent substantive update to this review essay:  — SMcCandlish ☏ ¢ 😼  07:11, 21 November 2020 (UTC) Feel free to suggest updates, additions, corrections, etc. on the talk page.

The {{Nutshell}} atop ArbCom's 2013 discretionary sanctions review concludes "Discretionary sanctions resolve disruption and promote civil participation within the specified topics." Both of those claims are manifestly false as DS is presently implemented. The first is true occasionally in certain topics; but just as often the dispute is not resolved, and instead whoever is better at playing the WP:CIVILPOV game will "WP:WIN".

As for the second, fear is not peace, and reluctance to be involved is not collaborative participation, it's just another means to the end of long-con players at a PoV-based public relations campaign getting the version they want. The average editor is here to work on a cool, important project, and honestly has better things to do if they're put off by administrative muzzling, and has limits as to what NLP strategies they're willing to engineer, if any; most of us do not approach this as a real-time strategy game, but as a results-based meritocracy, like any other free content, open development project. (You can be a little irascible if you write good code/material; no one cares how sweet you are if what you submit pollutes the repository with crap that others have to weed out.)

By contrast, for someone on a mission to "right a great wrong" or just distort the truth for their own or their employer's/government's/religion's benefit, there is virtually nothing in the world more important than craftily steering public perception of their pet topic, even if it takes years and extreme discipline; civility, in this model, is not a good-faith behavior, but a tool of manipulation, a form of social engineering used to get the desired result. This mismatch, between being here to volunteer some useful work and mostly be nice, versus pursuing a focused, long-term campaign of manipulation, using fake congeniality as a disguise, essentially guarantees that flacks/zealots/propagandists can "herd" and provoke regular editors into civility breaches, here and there over time, that get them sanctioned away one by one. This is the classic war of attrition tactic. It's been ongoing here for a least a decade, and is a major cause of good editors abandoning the system (another leading one is administrative harshness toward them for standing up to the same PoV-mongers).

The 2013 review's main issue was that it seemed to have been launched with largely pre-determined outcomes in mind, and was all about wordsmithing how DS would be presented rather than adjusting how it worked or what its rationales and scope were in any way. I know that for myself and some others with related concerns, we felt that what were were saying on that review's talk page was falling on deaf ears, even if might ultimately have had some minor effect on the exact wording of the Ds/alert message and some other windowdressing. The predetermined-outcome effect may have been because it was being treated as an experiment: "Here's what we're going to try, and we'll come back and look at it again after a couple of years and see how it went." Well, that time has passed, and it went poorly (and many of us predicted that it would, for reasons that were already apparently in 2012; it didn't actually have to be done this way).

The entire 2013 review was structured as a section-by-section editorial clarification process, when much of what was needed was a discussion and reshaping of the entire DS rationale and "vision". It's not clear that the present DS system is working at all – not in how its setup creates various loopholes; not in how it frightens away drama-avoidant, productive editors; not in how it is insufficiently applied to restrain flame-war encyclopedia content battlegrounds; not in how it's misapplied to internal governance discussions; not in how it's extremely one-sidedly applied in the latter (with the effect, whatever the intent, of giving wikilawyering admins a supervote to silence, punish, and chill the future participation of those most involved in shaping WP:POLICY, in favor of those who object to it applying to their pet topics/projects/nitpicks); and not in how it's wrongly applied to derail ongoing legitimate dispute resolution processes at ANEW, AE, etc. That last is the most irrational result, leading to sanctions for aspersions/battlegrounding, in the very processes that exist for the sole purpose of resolving disputes about editorial behavior. It's like going to court to file a grievance that someone lit your lawn on fire, then having the officers of the court behead you for daring to raise such a complaint in their presence against someone they like better. It's a Dark Ages approach to conflict resolution and it has to stop.

Some form of and approach to DS is needed, but it's only needed for intractable "holy war" behavior in topics like ethnic/nationalist disputes, and whether an alleged therapy is fringe quackery or real science. And it needs to focus on the effect on the encyclopedia, not whether a productive editor was being not quite polite enough to someone advancing a snake-oil or racialist agenda. WP has way more important things to do than manners enforcement when it comes to topics like that, and it's disheartening to everyone who cares about this being an actual encyclopedia, not some giant social networking game, to have AE and ArbCom focus on punishing people for their temperamental imperfections instead of for programmatically warping content and process to favor their external agendas, and studiously thwarting consensus formation.

The last three sentences of the "Guidance for editors" section (of the current AC/DS page) badly need revision, to properly reflect that the primary purpose of the forums it mentions are for the airing and resolution of inter-editor grievances. The present wording and the variant before it have for several years been used as an excuse to permit all sorts of vile attacks on whomever is being pilloried as long as angry admins seeking to punish them don't care how they're being verbally abused, but turned into a particularly nasty form of thought-crime policing any time some admin wants to shut up someone who is criticizing someone they'd rather not see criticized (or simply because they don't like the editor doing the criticizing). As it's presently written, in a weasely way that on first reading seems reasonable, it defies the very nature of DR on WP, while missing the point entirely. DR largely consists of "casting aspersions" and an adversarial approach between editors. That's why it's a dispute and why we have boards for resolving it.

What actually matters in these disputes is a particular bright line of civility that should not be crossed. Like a highway divider, it has two stripes, detailed below.

This is the most common and most poisonous form of aspersion-casting that we actually need to address, clearly and without confusing it with other forms of allegation (legitimate or illegitimate).

This especially pertains to aspersions about another editor's motivation, belief, intent, or other mentality matters that no one could actually know except by mind-reading powers. This does not mean using "your edit was..." instead of "that edit was", and it does not mean vernacular constructions like "you just want this article to say..." that don't really imply psychic knowledge of another's thought processes, but are a shorthand encapsulation of observed editing patterns. It certainly does include "You're just a shill for Microsoft", and "the only reason you edit this page it to push a neo-Nazi agenda", or "your prurient obsession with this subject's gender identity ...".

However, what's is usually missed is that an accusation (including by an admin seeking to impose sanctions) that party X "is assuming bad faith" is also a violation of that bright do-not-cross line unless X has made an allegation that unambiguously questions another editor's faith and cannot back it up with evidence. If they have reasonable evidence, they are not assuming, but making a reasoned evaluation under WP:DUCK / WP:SPADE. Meanwhile, assuming and accusing someone of making an assumption of bad faith is itself casting an unprovable allegation of bad faith, in turn. An admin's (or a gaggle of ANI wanna-be admins') desire to swing a sanctioning hammer doesn't give them license to make what amounts to a veiled personal integrity attack.

Point A is also important because it is essentially impossible to address programmatic WP:NPOV and WP:NOR problems without sometimes addressing evidenced WP:NOTHERE behavior and what external viewpoints it appears to advance – albeit short of evidence-free accusations of shilling, obsession, or propagandizing. Productive editors cannot be made to fear Sword-of-Damocles punishment for it by hypocritical "enforcers" making accusations of their own on nothing but psychological projection. It's frequently necessary to observe that editing patterns appear to dovetail with off-WP political, religious, ideological, commercial, fandom, or hate agendas. Our task as mutual moderators of WP interaction is to channel this into noticeboard and ArbCom/AE requests that phrase this in observational terms about evidence patterns, instead of assumptive/projecting and ad hominem ways (I mean the latter in the strict sense of the logic fallacy).

Leaping on failures do this perfectly as things to treat punitively with discretionary sanctions does not work. All it does is teach the PoV-pushers how to WP:GAME the system against their opponents (i.e., those of use working to build an encyclopedia rather than a giant blog-thing), while ensuring that more and more of those who might have stood up to them will shrink away in fear of being sanctioned themselves. It is more important that the encyclopedia not be full of lies than that every discussion about these lies be conducted with utmost regard for the feelings of the liars. Cf. WP:COMMONSENSE.

Admins and noticeboard !voters frequently cross this line in another way, the allegation that an editor "cannot" change, is "incapable" of collaboration, or is mentally incompetent in some other way. WP:COMPETENCE and WP:ROPE (there's a move to replace the latter with WP:TIGER) are extremely problematic in this regard, because their wording, in places, their shortcuts, and their typical use involves the idea that it's perfectly fine to imply or even directly accuse another editor of emotional instability as long as you include a wikilink to an essay some of us like. While COMPETENCE has sections in it that are valid, the entire basis of ROPE is that you're probably a jackass and will probably hang yourself; the entire page is an AGF-violation engine. When there's strong evidence that someone has mental problems, e.g. an inability to reason, or a propensity to type strings of obscenities at people after being administratively warned to stop, it does not actually require writing out an accusation, in any form, that they have mental issues (or are "off their meds", or any other snide insinuation); just deal with it as disruption and move on without being WP:JERKs about it in public. I've started taking people to task for this at ANI, but it largely falls on deaf ears because people are so accustomed to this little quasi-sanctions excuse to be asshats to other editors.

This is the second most common and detrimental form of unjustifiable accusatory behavior that we actually need to address, clearly and without confusing it with other forms of allegation (legitimate or illegitimate).

Note that this does not mean "which were not diffed again for the umpteeth time, and every single time the accusation has ever been made". This cannot be stressed enough. If an allegation is made and it has been demonstrated before, or can be demonstrated easily enough, especially if it's based on a long-term pattern of similar behavior, it's natural and reasonable to bring that allegation to a venue for addressing editorial behavior, as long as one is prepared to prove it (again) within a reasonable amount of time, on request.

As with point A, this is another area where hypocritical enforcement happens constantly, with admins or noticeboard pile-ons, hot to sanction someone, making a long string of rather nasty accusations as the basis for the punishment they want to inflict but without actually proving any of it. Or an admin (who did not do any real background examination on a dispute) sanctioning on the spot which ever party they don't like, they disagree with, or they think just wasn't quiet enough, when what they allegedly did wrong was re-mentioning but not re-diffing something already well-proven in previous noticeboard disputes, etc. Such "secretly involved" admin approach very often blatantly allows the favored party in the dispute to vent and vent all kinds of unreasonable, unprovable nastiness.

I've seen someone do this at AE, while under a topic ban about the issue in the request, and not a party to the request in any way but just there out of vexatiousness, and while the entire focus of the request was on aspersion casting without evidence, yet they got away with it because their target was not popular with the admins there that day, who even criticized enforcement of the topic ban against the interloper. I kid you not. I'll be happy to provide the links on request via e-mail (the editor in question was eventually indeffed, by ANI, even as AE punished his target, and I'm not comfortable naming an indeffed editor in public who cannot speak in his own defense now). It happened again recently, where an admin actually disrupted an ongoing ANEW proceeding to one-sidedly sanction one party in favor of the other, despite the hammered party actually having proven their case with a large number of diffs and external sources, and the other simply engaging in civil-PoV denialism (same deal: I'll demonstrate that in e-mail if necessary, but that dispute has quieted and need not be rekindled).

DS being applied to a topic area must have a time limit (6 months, 1 year, whatever), at which point the DS can be renewed if an amendment request makes a good case to do so. The present rule is "Discretionary sanctions are authorised ... by committee motion. Once authorised, they may only be revoked by committee motion. ... When it becomes apparent that discretionary sanctions for an area of conflict are no longer necessary, the committee may be asked to rescind the authorisation of them". One should not have to file what amounts to a new legalistic case at ArbCom to prove that a major alteration of normal WP operation should continue; the burden of proof belongs on those who think that WP process is still continually failing at that locus of the project. A time limit system wouldn't even change that rule, it would just entail the authorisation including a time span on a case-by-case basis (I would suggest 6 months by default; that's often long enough to stir the editorial pot and get new people involved). This could also obviate the "expired-alert immunity" loophole. DS should not last forever-by-default, because intelligent editors who avoid drama – the kind most needed to make an article great – tend to steer clear of DS topics, making the shaky quality and poor atmosphere at these article a self-fulfilling prophecy. That said, if some topic areas are going to remain under DS indefinitely (i.e., are shoo-ins for renewal because ongoing strife continues), then deeply involved participants can no longer be allowed to escape DS because of notice-related bureaucracy; if they're involved they're involved and they know about the DS.

And it's not just a problem with individual admins imposing this or that individual sanction without an expiration. The problem is endemic to the entire DS system, and predates it, going back to its origins as "article probation". One only has to look at the WP:ARBANEG case. Several years ago, a page was placed under "article probation", and even then there was no actual cause to restrict editing in this area. There was a single individual abusing admin authority, to editwar and then lock pages to save his "correct" version, and misusing his admin weight and reputation, to engage in incivility and personal attacks for an extended period without check. This should have been an open-and-shut desysop-or-admonish case, with no implications for how the article may be approached by other editors. Instead, this novel "probation" remained in place for years, and rather than remove it, ArbCom converted it along the way into DS, which remains in force, despite the fact that only a single enforcement action relating to this case has been logged. Permanently turning entire categories of "broadly construed" topic areas into zones of menacing FUD is not what DS was enacted for, even if a few areas remain overheatedly controversial and may justify longer-term DS (a few self-evident examples are WP:ARBAA2, WP:ARBBLP, WP:ARBR&I, and WP:ARBIP).

Despite all the efforts to make them seem non-threatening and neutrally informative (including a "look and feel" overhaul a couple of years ago, primarily at my own prompting), the {{Ds/alert}} template still feels quite menacing to its recipients. In my experience leaving them on editors' talk pages, the reaction has been uniformly negative. Has anyone ever said "Oh, thanks, I didn't know about that DS stuff, and I'm glad you informed me of it?" No; the usual reaction is "why are you threatening me?", sometimes WP:REICHSTAG remonstrative reactions, but most often just silent withdrawal from the topic in question – the editorial chilling effect. See the thread (permalink) that started this discretionary sanctions analysis for what else is wrong with the alerts system, including numerous gameable loopholes, and a general sense (including on the part of admins) that it's a pointless, bureaucratic morass.

The bare fact of the matter is that more often than not, when a regular editor leaves a Ds/alert [why isn't that "DS/alert"?] they're doing it as a threat, a warning that if the other editor doesn't back off, they can expect a nasty pile of drama. But when an admin leaves one, it's also and even more so a threat, a very direct one that the editor can definitely expect a discretionary (i.e. out-of-process, judge-jury-executioner) sanction if they do not basically flee the topic area in terror.

This is the bed ArbCom has made, but we all have to lie in it, and many of us are not comfortable since it's made of nails, and the AE crowd love to hammer them into whomever is handy, if I may mix a few metaphors. [I wrote that sentence that way on purpose to illustrate the previous point that a vernacular turn of phrase about an observed pattern is not literally an imputation of emotional state. I'm sure that what even the most hard-assed Robocop admins actually love are their families and the other most important things in their real lives. They find strongly punitive administrative action here appropriate for reasons that vary from person to person.]

The section on what constitutes awareness in the 2013 DS review says in part that anyone is aware that DS applies to a topic if they "have at any time been sanctioned for their conduct for the area of conflict (and that sanction has not been successfully appealed)". This makes no sense at all (and ArbCom knew it didn't, having rejected a multi-editor appeal of a wrongfully-accusatory precursor of the {{Ds/alert}} template, on the basis that awareness couldn't be undone, before they understood that the appeal wasn't about the awareness but the false accusation. ArbCom can't take it back that they understand that awareness can't be taken back, and this too relates strongly to the prior thread [permalink]).

While the present AC/DS page does not contain this wording, it weirdly dropped the part about "been sanctioned [in] the area of conflict". While it retains "participated in any process about the area of conflict at arbitration requests or arbitration enforcement", the entire point of DS is that it doesn't require any such process, but can be applied any time at any editor's discretion. So, this is yet another loophole, I think the fourth that I've identified (see prior thread linked above). While the present AC/DS page no longer has "were mentioned by name in the Final Decision of the case", and does include participation in a relevant RfArb or AE discussion as sufficient, it didn't include ARCA discussions, nor does it include other discussions, e.g. at noticeboards or in user talk. In such cases it actually would be clear that the user was really aware of the DS about the topic. Loophole #5, then.

This is why my earlier post about this^* focuses on getting rid of the bureaucratic approach to "awareness"; it's a loophole factory, and awareness of DS should be a common sense matter like everything else is on WP. ANI and other noticeboards have no trouble sussing out whether an editor under examination was aware of a policy, of previous warnings, of the disruptiveness of their actions, etc., so this should be no different. "Ignorance is no exception" is part of virtually every legal and other rules system on the planet, and should apply on Wikipedia as well. The entire idea of being immune to repercussions because you didn't know what you were doing was bad is not practical in any system of governance.

While WP:Arbitration enforcement sees a number of well-rounded, encyclopedia-minded admins who do lots more than monger rules, this board has taken on too much of the character of a special "brotherhood" of judge-jury-executioner "I am the law!" admins. It only takes a few spots of mold before no one wants any part of the loaf. The WP community has shut down prior attempts (e.g. WP:Esperanza) to form "in-crowd" pseudo-organizations within Wikipedia, and it's wrong for ArbCom to defy this by erecting another one, with excessive powers.

There need to be limits on length of time and number of actions that any one admin can take under AE colors (6 months on 6 off, etc.).

We must also have an elevated standard of non-WP:INVOLVED action. There's a strong argument that an "I've sanctioned this editor several times in the past" record should force the recusal of an admin at AE, except as a requesting/commenting party, specifically because the DS are discretionary (i.e. based on an admin's opinion of the situation, not ArbCom's, nor a community noticeboard's, and that individual's opinion is already pretty well set if the admin has has run-ins with the same allegedly problematic user before); plus DS are virtually impossible to appeal in practice. In the real world, we would never let someone who, in a previous role, had been a cop who arrested someone 5 times get to be the judge at their 6th trial, or get to be their parole officer later down the road. But that's precisely how AE operates. It's as if 500 years of real-world experience in justice had never happened; certainly was not adequately learned from.

There also needs to be more community involvement in AE, so that it operates normally and sensibly, like our more familiar, other noticeboards, instead of being an special admin playground in which borderline hostility to non-admins' views is the modus operandi. WP:CONSENSUS can form anywhere, and AE is not magically off-limits to the editorship at large. The very fact that it's called "Arbitration enforcement" is problematic, and obviously will attract admins who want to focus on being "enforcers". WP does not need a class of "professional" Wiki-SWAT cops, or a standing army of Wiki-Marines. All admins should be first and foremost editors. Anyone here because they're excited to be sticking it to other people are, really, WP:NOTHERE to work on an encyclopedia, and we do not need a venue for indulging anyone's "how many chumps can I frag today" deathmatch gaming fantasy. It's utterly counterproductive to create a higher-stakes WP:BATTLEGROUND in the name of putting a stop to battlegrounding at article talk pages.

Any review of WP:AC/DS is necessarily going to have to be a review of WP:AE, since AE's sole function appears to be to bring down the DS hammer on other editors.

Here's why (from WP:AC/DS#Role of administrators): While discretionary sanctions give administrators necessary latitude, they must not: ... modify a sanction out of process ... Administrators who fail to meet these expectations may be subject to any remedy the committee consider appropriate, including desysopping. The 2013 DS review version has similar language, so this admin-frightening threat has been around a long time.

Admins cannot clarify or reduce, much less overturn, the DS actions of another admin (even one who is AWOL), no matter how egregious, unless a new drama-factory is fired up at AE, WP:AN, or WP:ARCA. This effectively makes DS unappealable except in the most outlandish of cases, because AE and AN are extremely reluctant to question the judgement of another admin, and ArbCom has made it clear that they generally will throw out such appeals on sight unless the administrative blunder is severe enough that it warrants a desysopping consideration.

ArbCom and AE generally seem to just not get it that a wrongful DS sanction, even a seemingly minor one, may be enough to 1) drive a productive editor into quitting, or 2) be used as leverage in every possible dispute by combative PoV pushers to chill and silence that editor and others like him/her, which is liable to lead right back to point 1. It's genuinely important that one admin be able to undo the actions of another if they have good reason to question them, short of WP:WHEELWAR. It's always been this way here. Arbitration enforcement actions are not immune to this fact just because a line-item of red tape at AC/DS would like them to be, for reasons that are not even articulated. It's important that ArbCom is able to make up procedural rules about its own operations, and it's empowered to hand down decision in cases brought before it.

At no point did the community ever ratify the idea that it could radically change WP's administrative operations in ways that a) empower "lone wolf" admins to impose effectively permanent and usually unappealable sanctions on a whim, and b) prevent anyone else in the community, even other admins, from doing anything about it, short of a WP:DRAMA process that is already strongly stacked in favor of the status quo of the questioned admin's unilateral action. This is, actually, precisely the kind of abusive "model of a police state" nonsense that opponents of the draft plans for ArbCom feared would happen. And it certainly was not within the remit for which the WikiMedia Foundation drafted the outline of an arbitration committee in the first place. This isn't arbitration in any sense, it's effectively unaccountable "Wild West" rule by wiki-gunslingers.

The community's perennial demands for some kind of additional appeal body, with non-admin, non-arb sitting members (e.g. on a rotating volunteer basis) arises directly from the bureaucratic morass that leads to the present appeals process being a sham[bles]. While no consensus has emerged on how exactly to address it, there quite clearly is a consensus that appeals are "broken". Once the DS hounds have been set loose there is basically no way to get their teeth out of your ankle, even if you were not trespassing, short of something pretty close to divine intervention. That's not really a process, it's a façade erected to disguise the fact that there is no effective process.

A secondary but serious effect of this problem is that sanctions that would have been temporary had the community arrived at them via normal noticeboard processes are effectively forever if applied under DS and the admin didn't happen to set a time limit, and doesn't like you enough to consider your request for one later. AE/AN generally won't do it, and ArbCom at ARCA usually won't either.

The obvious solution to this problem, short of creating some new Appeals Committee or whatever, is to make WP:ANI the primary path of appeal (after the sanctioning admin's own user talk page}}, or to create an appeals noticeboard if ANI is seen as too busy or too angsty for that purpose. The community can and should be able to overturn anything, and while WMF imposed the requirement for something like an ArbCom, it was left to the community to decide its operating parameters. Even as a simple WP:IAR matter, a community noticeboard has the capacity to undo unreasonable (or no longer reasonable) restrictions placed by a single admin making up their own mind; DS giving admins the delegated authority to act on ArbCom's behalf doesn't make them infallible popes. It simply does not work to have AE, essentially an exclusive admin club where they always have each other's backs, determine with any fairness whether a one-admin restriction imposed on some other editor is appropriate in every way.

Almost amusingly, the first statement in the "Appeals and modifications" section is misleading nonsense: "Appeals may be made only by the editor under sanction and only for a currently active sanction." I know this for a fact, because after two years of trying to get a bogus DS action vacated, and being ignored by WP:ARCA (twice), WP:AE, and WP:ARCA, a third party, non-admin filed a clarification on my and the other appellant's behalf, ArbCom finally listened, and the appeal was actually successful [despite extraneous prosecutorial activism against me in particular by one then-member of the Committee forgetting to be an impartial arbitrator, but that's another matter.]

Not only that, but the DS action in question was entirely moot as an "active sanction"; the appeal was about the retention of a false accusation in the DS log. So, not only is the "rule" not a rule, in at least one case the only way that justice was served was by ignoring both prongs of the "rule" simultaneously. See also the previous policy observation: WP:CONSENSUS can be formed anywhere at any time, including the special WP:LOCALCONSENSUS of ArbCom within its own sphere. ArbCom has said before that it will not be bound by its own precedent, and this is clearly the case. So why is there a pseudo-rule in here that attempts to bind ArbCom to a decision it kinda-sorta made several years ago and has since dispensed with when necessary?

This is unbelievable WP:LAWYER nonsense: "Nothing in this current version of the Discretionary Sanctions process constitutes grounds for appeal of a remedy or restriction imposed under prior versions of it" (also found in the 2013 version). This needs to be vacated, and nothing like it ever written again here. If the process was flawed, you don't fix it but tell people suffering under the flawed version that they're screwed. Again, this is a basic principle of post-Medieval justice. If you stop making homosexuality or publication of anti-establishment pamphlets a crime, you let the gays and pamphleteers out of prison. Duh.

Fortunately, this provision isn't just ethically wrong, it has not even been consistently obeyed anyway. Though it took two years, I was able to successfully appeal something, and it was ultimately the common sense of how DS works after the 2013 DS review that prevailed, not the confused "we haven't really starting thinking through how this may affect people" quasi-rationales that were in force when the sanction had been issued. A rule that is not followed is not a rule, it's just noise, so excise it. It's also interesting to note that when I first attempted to appeal that matter, I was repeatedly told that the then-forthcoming 2013 review would surely resolve my issue; it did not. Imagine my surprise when this "Nothing in this current version ..." wording appeared, which (had its own authors actually taken it seriously) would have completely precluded the appeal I was made to defer, and which eventually succeeded. I put this down to red tape and too many cooks, not disingenuousness, but it's illustrative of why DS is a mess that no one seems comfortable or happy with, other than some percentage (by no means all) of the admins who take on AE tasks. What should be modeled on a judicial process has instead become closely akin to a legislative one, complete with false campaign promises and "sausage-making" that undermines the intent of the "legislation" (rule-making) in the first place. This leads directly into the topic of separation of powers. In the real world it's called "judicial activism" or "legislating from the bench".

Proceeding with the "judicial activism" theme in the previous section, this is a fairly extensive essay unto itself, making an argument that is politico-ethical, socio-anthropological, philosophical, and pragmatic. The nutshell version:

• Our "judiciary", ArbCom, has no business trying to use our "executive branch" of admins to micromanage the internal deliberations of our internal "legislature" of policy formation processes, as it is doing in WP:ARBATC by wrongfully applying DS to that case. DS are a remedy that the community accepted (barely) because it was seen as probably necessary to restrain a completely different kind of dispute, namely intractable fighting over encyclopedia content, mostly ethno-political, or faith-vs.-science in nature. It's badly monkey-wrenching a significant aspect of WP's internal self-governance processes, and the use of admins to do it is a three-way corrosion of separation of powers, especially given the hostility that so many admins show toward the subject matter of ARBATC and toward editors involved in that policypage area.

An explication of this in detail follows.

There's an inescapable separation of powers problem inherent in cases like WP:ARBATC and the DS applied to it. ArbCom is essentially Wikipedia's judiciary branch, along with the [non-binding] WP:Mediation Committee. The admin corps are our executive branch, along with some other functionary roles, like the Bureaucrats, and WP:OFFICE. The bulk of WP's editors, "the people", collectively form consensus on things, and those of us genuinely interested in internal policy formation are WP's self-selecting legislative branch, helping form consensus on WP:POLICY matters. (There are other types of essentially parliamentary/congressional process on WP, operated on the same principles, such as all of the XfDs, and the organized engineering of our templates, modules, and other on-wiki infrastructure, when it is actually organized.)

The purpose of the legislative branch in any socio-political conglomeration of people, be it multi-nation confederation, or a small town, a world power, or a corporation and it policies, is to evenly regulate the "civic" behavior of its "citizens" for mutual contentment and productivity. It is not to dictate to the other branches how they go about their business within their own governance roles. A prime minister does not tell judges how to interpret the law. The courts do not make the procedural rules of the legislature. The legislative body does not determine presidential directives or law enforcement priorities. And so on. Aside from these roles, everyone is a citizen, subject to the same rules, enforcement, and adjudication.

DS is (even if nominally preventative not punitive) a judicial restraint – a reduction of freedom and autonomy – exercised against the subject "public" by executive officers. WP's community of "citizenry" has allowed the erection of this system of enforcement because it seems necessary to quell "civil unrest" in certain topic areas, like unto gang wars on city streets. The problem here is that it's been taken, wrongly, as blanket license to tell the legislature that it may not debate loudly, and that if anyone raises their voice they'll be banned from the capitol. Given that various members of the administration have evinced a desire to impose silence on the legislature, this is recipe for disaster.

In actual practice, the results have been very detrimental, leading to a resignations of long-term editors over unjust punitive treatment, effective derailing of proposals, emboldening of WP:CIVILPOV actors deeply involving in trying to change policy, emboldening of a particular stripe of insular wikiproject who see WP:OWNership over categories of articles they claim within their scope, one-sided enforcement against "regulars" at particular policypages, repeated re-opening of perennial anti-consensus proposals as if they're new, changes to policy pages that do not represent consensus at all but which go largely unopposed out of fear of sanctions admins who appear to support the change, "cult of personality" changes at ANI and elsewhere, increases in behavior barely distinguishable from trolling at various internal pages, tacit administrative approval of promotion of false information in internal documentation (which in at least one case has spilt out of WP namespace into articles not just on WP but also on another WMF project), and many more besides.

Internal policy debates in any organization can become heated. But they're qualitatively different from content debates at articles and their talk pages, just as the UK's notorious rancorous Parliament debates are not comparable to football-hooligan violence outbreaks at arenas (even if some of the innate hominid behavioral instincts underly them all).

DS needs to be removed from all non-encyclopedia-content cases, and never again applied outside them. It's not legitimately within ArbCom's authority to interfere with the Wikipedia community's ability to self-govern, even if some of the debates it gets into in the process cause some temporarily lost tempers. There are other mechanisms (particularly ANI and AN) for dealing with these matters if they get out of hand – community processes, not vigilante admins. And DS simply doesn't work in internal matters. All it does it give carte blanche to infrequent but agenda-pushing visitors to those projectpages, who mostly do not have {{Ds/alert}}s on their talk pages, to act out with near-impunity, while their victims have to remain obsequious or silent for fear that a predictable cluster of hammer-them-hard admins will descend in a pack to punish them for defending themselves and the project's stability.

That last point really matters: It's an order of magnitude more important that our policypages remain comparatively stable than that any particular line item in them be just so (perhaps with a few exceptions, most of which were imposed on us from on high by WP:OFFICE, e.g. some of the central points of WP:BLP). There's a reason that policypages (including actual policies, as well as guidelines, and key essays that have guideline-like community acceptance) are unusually resistant to willy-nilly changes; seemingly trivial changes have the potential to impact tens of thousands, even millions of articles, but most of the sought changes are essentially arbitrary and subjective.

Because of what's at stake, it's frequently necessary to show how irrational something is, or how biased (even if, yes, this would be done in a non-WP:JERK manner). But every editor who dares to do so more than a handful of times it at risk of disproportionate AE / DS enforcement, because it gets noticed, and someone can falsely claim its a "pattern of battlegrounding" or other bad-faith projection. An admin with a bone to pick, or who simply has not looked into the issue in any detail, is apt to side with the complainant, even if they're actual being disruptive (shopping for a new "parent" to ask, falsifying sources, relentlessly pushing a fringe viewpoint, refusing to accept a consensus decision because they weren't around to get their say, or whatever). The challengers to various central points of internal-process consensus tend to be random, from all corners.

The "shepherds", the institutional memory, of what that consensus is, when and where it was determined, and on what rationales, against what other rationales, are a limited and fairly slowly changing group. DS makes that group change more rapidly, but by nothing but shrinkage, with a concomitant loss of institutional memory and stability. It's getting to the point these days where on some of these pages the same tired old consensus matter is getting "re-legislated" in increasingly frequent repeated challenges (e.g. multiple mile-long flamewars on the Village Pump within a few months of each other, in which opponents of established consensus demonize its defenders in remarkably crappy terms with no repercussions of any kind, yet raising no new issues).

DS in ARBATC in particular has become a trojan horse for seemingly unending WP:FORUMSHOPPING and tendentious refusal to get the point and move on. It's been a not-so-fun experiment but a failure. Imagine what this would do to WP if applied to, e.g., WP:NPOV, WP:V / WP:RS, and WP:NOR policy formation? People's views are very, very strong on those matters (especially when it involves WP:MEDRS and WP:BLP matters). I think that ArbCom has realized, consciously or not, that WP would not tolerate the application of DS to internal deliberations over the WP:CORE content policies, and consequently has properly applied them instead to encyclopedic topic areas that involve NPOV/V/RS/NOR conflicts, not to the policy formation process itself. The community has not risen up in alarm about ARBATC (WP:AT and WP:MOS) DS for the role reason that most editors just don't care that much about titles and style, so they simply haven't noticed the negative effects. And this internal process area is being especially targeted by ArbCom – this is the second case in which ArbCom has tried to cordon off MOS/AT as subject to special restrictions out of left field, and zero other policy areas of Wikipedia have ever been approached this way, despite several of them being at least as fractious and subject to prolonged conflicts. The fact that grammar and naming debates aren't a popular thing on everyone's radar doesn't somehow make it okay for ArbCom to knock a huge hole in the separation of powers that enables WP to function as well as it does (to the extent that it still does, an extent eroded directly by ArbCom's misuse of DS outside of content disputes).

It's not even obvious to everyone what rules are actually in force and must be applied. While Wikipedia:Arbitration Committee/Discretionary sanctions/2013 review has a template atop it saying it's historical, not everyone gets to everything in that page from the top; it has sections which have been directly linked in various discussions. It says under "Housekeeping" The following are amended: ... the provisions in Wikipedia:Arbitration Committee/Discretionary sanctions are rescinded as they have been incorporated into the provisions above." Note that's a regular wikilink to the WP:AC/DS page, not to a diff of any particular version of the page. It's thus directly stated to anyone reading the review page that they should ignore what presently appears at AC/DS, as historical / void. It has not been amended to say that this instruction was superseded in 2014, and that WP:AC/DS has been revised and is now the controlling document again.

If someone goes and looks at WP:AC/DS, it says at the top, "This updated Discretionary sanctions procedure was authorised by motion on 3 May 2014 and superseded and replaced all prior discretionary sanction provisions with effect from that date.", reinforcing the confusion. It doesn't link to or name the review page, and the link it does provide is just to the ratification notice of some post-review modification motion that I'm skeptical many people knew about at all, not to that discussion itself which people could compare with the review and realize that they're different. It's a fair bet that many editors who remember that there was a review and where it was will think that this note atop AC/DS refers to that review, not a "re-review" that came later; they'd have to remember or notice the 2013/2014 difference, the only indication anywhere that these were two separate revision processes. Meanwhile, the 2013 review page never mentions this 2014 modification, and still asserts that it supersedes the entire ARCA page! This isn't really much of a version control challenge, yet here we are with conflicting DS pages, each claiming to supersede the other. We call this a WP:CONTENTFORK.

Given that the entire point of the 2013 review, and of the DS/alert system is stated to have been clarity, unambiguity, and reduction of bureaucracy, the opposite has been the result. This is strongly evidenced throughout my review, I just include this last bit as kind of the opposite of a cherry on top.

Overnight, DS turned adminship from "WP:NOBIGDEAL" – in its original conception – into a very big deal. There has been a decidedly negative sea change in community perception of adminship, with it becoming something that requires intense scrutiny and near-perfection on the part of candidates. WP:Requests for adminship (WP:RFA) has radically changed from a once-simple process anyone competent could pass, into a grueling ordeal few editors will ever submit themselves to.

This has had serious, cascading consequences:

• Our administrator pool is rapidly declining, and nearing the point where the system is no longer properly functional (many would argue it's already past that point, given the length of various administrative backlogs).
• The rate of RfAs has radically declined from several being ongoing at almost any time to often entire months with no candidates.
• RfA has become a hellhole, despite some attempts at reform. Failed candidates become so discouraged by the abuse they receive they fairly often quit the project, and few ever run again.
• Adminship is less appealing to genuinely competent editors, and more appealing to those seeking it as a position of power to abuse, of authority to wield, against an increasingly "peon" editorial pool – one chock full of admins at every turn, it is now only sporadically dotted with users who have the administrator bit.
• Fear of losing more admins makes the community afraid to deal with administrative abuse or to create processes for this; we presently have no means at our disposal other than an ArbCom case, and that process is so legalistic and trying for all involved that it is extremely rare of administrative abuse to ever be challenged.
• Fear of being desysopped, should anyone have what it takes to get through the ArbCom red tape, makes good-acting admins afraid to do anything controversial: if a case is brought against them, they're likely to lose the bit, and have a very difficult time getting it back because the "new RfA" is such a nightmare.

Over the long haul, this is a recipe for disaster – it is a non-sustainable system. The entire notion of empowering admins to issue – on the basis of just their own opinion – long-term blocks, topic bans, 1RR restrictions (a.k.a. pit traps for blocking good editors one after another for engaging in regular editing and not remembering an extremist rule on a particular page), and other remedies including ones admins can just make up out of thin air, has proven to be too much power. We must seriously consider just bringing DS to a close. No one can seem to identify a single case in which DS was the only way to deal with a disruption issue, especially since we have WP:ANI, WP:Community sanctions, WP:General sanctions, WP:ANEW and other noticeboards, WP:RFPP, WP:RFAB, and other processes with more oversight, and most of which (i.e., aside from RFARB) move quickly enough to be practical. We simply do not have an actual need for either WP:AC/DS or WP:AE.

• Wikipedia:Arbitration Committee/Discretionary sanctions – the locus of all this trouble
• Wikipedia:Arbitration Committee/Discretionary sanctions/2013 review – the first and only community review of discretionary sanctions since implementation.
  It was far more contentious than in it looks in this "whitewashed" form, because the large amount of dispute has been archived away:
  • Archive 1
  • Archive 2
  • Archive 3
• Wikipedia:Arbitration/Requests/Case/Arbitration enforcement – June–August 2015 case that could have had a reformative effect, but did not
• Wikipedia:Arbitration/Requests/Case/Arbitration enforcement 2 – October–December 2015 case that could have had a reformative effect, but did not
• User:RGloucester/Sanctions – a history of sanctions on Wikipedia