I know that my answer will overlap others, but enough people are waiting for me to comment on this process with an eye toward my case in particular that I'm going to post anyway.

The new removal process, while better than what was done to me, lacks important safeguards present in the pre-existing Moderator Action Review Process. Most importantly, it lacks the interactive phase where the accused can respond to comments raised by others. It also lacks transparency, being handled entirely by two CMs out of view of anybody else including the accused. It also does not allow the accused to challenge specific judges for substantiated concerns of bias; it relies on CMs to recuse themselves, but the moderator has no say or even knowledge of who acted. If there were more trust between the community and SE this might be surmountable, but in the current climate that's an awfully big assumption.

The new reinstatement process is even more flawed:

• There is no transparency or ability to audit. The moderator submits a petition into the void and eventually an answer comes out. If I were to submit an application, a CM or Director could simply veto it and then say "we went through a process so we must be right".

• The moderator cannot challenge judges. If a moderator has been removed, chances are that there were some interactions with some community managers and/or members of the community strategy team that would prejudice an appeal. A moderator should not be subject to summary judgement by the employee who ruled on the removal in the first place, and the moderator should have the opportunity to raise specific concerns about anybody. The moderator should know who the judges are.

• The process starts from a presumption of guilt. If a moderator went through the paired removal process and actually received the information about charges and warnings called for in it, then the moderator has a starting point for an appeal. But if that didn't happen, the moderator is forced to guess. "Guilty until proven innocent" is not a sound judicial process; this process should not apply in cases where no sound removal process was followed.

• There is no dialogue, no hearing, and no opportunity to present witnesses or evidence. The interactive aspects of the pre-existing removal process (MARP) are missing in both the new removal process and the reinstatement process. While synchronous chat for worldwide teams is a major hassle, there needs to be some way to include the moderator in a discussion of the situation.

Both processes involve annotations, which should be shared with the moderator and subject to challenge. I can only imagine what unsupported annotations a CM or Director may have added to my account, for example.

The new removal process, if properly followed, would have given me a fair chance, though the concerns about transparency and bias remain. That process would have required SE to share information about the specific complaint with me and, at worst, would have resulted in a warning. But that's not what happened, and against that backdrop, the reinstatement process would not be fair in my case. I shan't submit to it.

The reinstatement process is broken as written.

The process is rigged against the previous moderator (PM).

Community Managers (CMs) can recuse themselves if they feel they can't be impartial.

There needs to be a way to recuse CMs who can't be impartial but won't recuse themselves. CMs recusing other CMs? The PM recusing up to a small number of CMs? The moderator teams for the target sites?

The process is concluded here.

This is unacceptable. The PM never gets a say in this whole process other than applying. They have no way to argue for themselves, set preconditions of their own, or even say "Hang on, the reasons you gave me are factually incorrect/too vague to mean anything!".

If a precondition is set, CM1 will communicate this back to PM

Wait, that's done after talking to the mod team? That seems pointlessly opaque: what's wrong with telling PM "Is $thing okay with you? If so, we'll ask the mods if they want you back on condition you do $thing" first?

reinstatement is possible under the agreement that a reasonable and attainable precondition should and could be met

This is vulnerable to "Have you stopped beating your wife?". It may be okay if the conditions are strictly about the future, e.g. "PM will never call a user a doo-doo head". It's certainly not okay if the condition can be "PM will stop calling users doo-doo heads" (implying that they did so) or worse "PM will apologise for calling users doo-doo heads" (when PM denies having ever done so).

access is restored

This is the best possible outcome for PM. There is no outcome that involves an acknowledgement that the removal was in error (e.g. a miscalculation of inactivity duration, a mistaken belief that PM had broken a rule), let alone an apology or reparations if applicable.

subject to summary removal for non-compliance

Who evaluates compliance? This could easily be gamed.

Thank you for taking the time to put this process together. I know this was dozens and dozens of hours across 7-10 people, If I understand your internal structure correctly. I wouldn't be surprised if you've spent 500 people-hours on this issue already, and I understand the desire to 'ship it and iterate'.

Please take a moment to pause, and think: What do these processes optimize for? What does shipping these processes now optimize for?

The answer you have internally is different than what we're seeing.

Here's what we see:

Trust is at an all-time low. I wouldn't be surprised if for some people there was less than zero trust -- that the level of trust went negative, as if you need to improve trust just to get people to neither trust or distrust you.

Personally, I'm at zero trust right now. This is not an attack; and it's not you individually, it's you collectively. I know most of the SE staff involved in these affairs on a first name basis, and I respect and cherish our interactions. The ones I don't know on a first-name basis, I've still had net-positive interactions with. In short, I am pre-disposed through our interactions to trust you, but currently do not.

I tell you this not to raise your hackles, but to hopefully share my feelings with you and show you another perspective.

These processes are a bit like putting the cart before the horse. These are pretty good processes for a high-trust environment, but you don't have a high-trust environment, and putting these processes in place right now won't make the environment higher trust.

If anything, putting the removal process in place right now will result in lower trust, and without taking in the substantive feedback of your community and moderators, you will further erode any remaining positive trust you have with members of the community.

So, take a pause. Regroup. Have a meeting. And in this meeting, figure out what you want to do: Do you want to put this process in place, tout suite? Or do you want to restore trust with your community? If SE employees reading this trust me, please hear me on this: Putting this process in place will damage trust, not restore it.

I'd love to get on a call with you to listen to the issues you have and to help you solve this problem, and I believe there are concrete actions you can take now that will restore some trust and other actions you can do to further restore trust.

The two new processes are palliative at best. I would NEVER want to be a moderator with such procedures in place:

1. There is no community oversight of any community manager's action, and frankly, I do not trust the CMs to never make mistakes and do not see why the community should. The oversight could be performed by other mods under NDA, and be not binding. Both parties need an impartial voice guaranteeing that the process was followed and that the decisions were not too controversial.

2. CMs retain absolute power, including the ability to ban a moderator indefinitely without giving any reasonable reason or appeal. This is unnecessary.

3. Any annotation that says a diamond can't be restored prevents it, indefinitely and with no limits. There are no appeals or guarantees of reinstatement in case of mistakes. This should be corrected (in fact, I don't see why it should be allowed at all).

4. CMs are responsible for inevitable mistakes in judging moderators and are also responsible for deciding whether they made a mistake. This should not be the case, and community oversight is essential to align the incentives.

5. There is no oversight for emergency demodding and nothing to prevent its abuse.

6. Substantially, there are no guaranteed timings. You can delay responding to anyone as long as you want. This should also be corrected, especially in the case of emergency demodding.

7. Accused mods are merely informed that the process is in place once their guilt is "determined". This is "guilty until proven innocent". By talking to @Catija I know this is not as intended but it needs to be fixed: mods should be contacted before any determination and the accusation and the evidence supporting it needs to be shared too, so they can defend themselves.

While I appreciate the effort in writing this, it's clear to me that you have not demonstrated the intention of creating a fair process - this apparently codifies the previous "non-process" in a process that fails to solve the more substantial problems. Instead, it should make more effort to increase trust in some form, and make sure it does not codify that you do not trust us, instead.

You should extensively review it or scratch it and build a process that enshrines mutual trust.

The new process was created in response to the lack of process that led to the firing of Monica. So I think it is fair to ask in what ways the events around Monica's firing violated the tenets of this new process, and what SE is planning to do to correct their mistakes. I think SE should state publicly in which ways the rights this process grants to moderators were ignored by the previous events. We have read an apology by SE, but that was a bit fuzzy on exactly what SE thinks they did wrong, and it doesn't really leave me with the impression that SE necessarily understood just how much they screwed up. It would help to restore confidence if SE would actually state what went wrong, and how the new process addresses these mistakes.

One important principle of the new process is that there should be a warning, and that this warning should contain clear and actionable feedback. This is to me maybe one of the most important parts, and one I personally consider one of the aspect that were not followed at all in previous events, to my limited knowledge.

I think the process itself is fine, and I'm glad to see that SE incorporates the feedback it received from the moderators. But the process still requires us to trust SE, as it's entirely in the hands of SE employees. Which is probably the way it has to be, I don't really see any other viable options. But that trust is something SE has to earn again, and while the new process is a good step, it doesn't necessarily fix the fundamental issue behind the botched firing without process. Resolving the situation with Monica is a prerequisite for quite a few people to earn back that trust.

These processes miss the point.

All of the company's current big issues with the community boil down to one issue: The company lost a lot of goodwill and trust.

Without trust, review processes that work must rely on transparency and consistency. A written process theoretically creates some consistency, but unless it's accompanied by transparency or trust, in practice people won't know if the process is followed correctly, fairly, and impartially.

Keep in mind how all of this looks: The company fired a moderator, refused to let her know why, then the company posted some seemingly false and slanderous accusations against the moderator, missed 4 separate opportunities to apologize and own up to their mistakes, and now, weeks later, graciously allows the moderator to ask to be let back in, if she promises, so to speak, "to stop beating her wife".

This looks bad. If what actually happened is any better than the above summary, the company should be desperate to create more transparency.

There is no protection against directed findings

The only people who can take decisions on moderator removal and reinstatement are community managers (CM). All CM are part of the same corporate hierarchy. The same person, or a very small set of persons, decides whether all CM will keep earning their living tomorrow.

This means that Stack Overflow, Inc. management can direct findings, either overtly or implicitly. If a CM knows that their boss wants a moderator removed or reinstated, but their personal conviction goes in the other direction, they'll have to make a hard choice between going with their conscience and keeping their job.

A fair process must:

• avoid such conflicts of interest and loyalty;
• in such a way that everybody involved (especially employees) can be sure that there will be no retaliation;
• and in such a way that the community can trust that the people involved felt safe against retaliation.

This is a difficult governance problem. There is both a high need for confidentiality (especially so that if a moderator does something wrong, the victims know that they won't face retaliation) and high trust (which is hard to guarantee without openness). There is some inspiration to be had with whistleblower processes which face the same difficulty.

I don't see how this problem could be solved without involving non-employees in the process. There needs to be some independent oversight. In the first few years of Stack Overflow, Jeff was the benevolent dictator. It worked because Jeff was “one of us”. Those days are long over. SOI is a mostly faceless corporation that has already demonstrated that it does not have the best interest of the community in mind. This cannot be resolved by company-internal processes.

This does not address the core issue

This must be called out, again, and again, until an answer is recived.

Monica was suspended without a warning, and accused of violating the CoC. This has been brought up time and time again. No process, not matter how detailed, can work where the party is presumed guilty, and until SE, either reinstates Monica and apologizes for slandering her, or provides evidence to back their claims, or Monica declares SE has managed this privately and she is satisfied, no process that is created will have any validity, or indeed moral authority, as the issuing party has shown itself unable to engage the community with the necessary transparency to convince me that any process will be followed. Especially one like this, that relies on CMs, rather than the users.

Monica remains unsatisfied

In the restoration process, Discovery phase, step 2.1,

If CM1 finds any annotation that states reinstatement is deemed impossible, the request is denied....

Under what conditions would such an annotation be added to the former moderator's account? Neither the Moderator Conduct Review Process nor the Moderator Action Review Process seem to indicate who has the authority to make such an annotation or what circumstances must be found to exist for such an annotation to be made. Is such a clause intended only for the worst of the worst, the remorseless "serial killers" of the moderation world, or is it intended as a standard annotation to be used in most cases of moderator removal?

The last process we had was completely ignored. What happens the next time a CM decides there should be an exception?

Initiation

The process is started when the Community Management Team is made aware of an issue. This can be a complaint against a moderator, or the Team being made aware of a security issue (like a moderator's account appearing to be compromised).

Emergency Removal Phase

The CM (CM1) who "stumbles upon" the issue will take care of it. Moderator access should be removed immediately, to limit the potential for any further damage. Available info is documented.

This looks like what happened to Monica. I can understand why a security issue or compromised account would warranting immediate, unilateral action; however, I'm still bothered that "a complaint against a moderator" is grounds for immediate removal.

I suppose this explains why calls for Monica's immediate reinstatement have seemingly been ignored and gone unheard. SE still feels like they made the right move; now they've put a process in place that conveniently justifies and codifies their prior actions.

Complaints against moderators are part of the business of moderating. You can't please all of the people all of the time.

I'm very disappointed in the way this is heading. Had the procedure said that, in the case of complaints, the Emergency Removal Phase will be skipped (instead of can be skipped), my concerns would be allayed. However, as written, I don't like how skipping is merely an option.

What constitutes an "emergency" for removal purposes? While an emergency by its very name indicates that things must be expedited to prevent ongoing harm, I think that this needs to be defined better, so that the vagueness of one word in a policy cannot be used as a sole justification for a removal. I certainly hope that expressing disagreement is not an "emergency".

I don't know if a list of things that constitutes an emergency is desired, or a list of things that are not emergencies, or both. But this needs elaborating.

Funny how no one really thought about what constitutes an emergency before. But just thinking about why we need this: it's about something many others and I have pointed out plenty of times recently. It's a matter of trust.

This is just one case of needing to "get it in writing". The times when people need to get it in writing are when working with other parties/groups where the trust level isn't high enough given the importance: the mortgage company when getting a home loan, contract review, credit card companies, lawyers, etc.

With trust firmly in place, so many people wouldn't be questioning, nitpicking, and just generally being cynical about all Meta interactions.

In addition to getting the process right, do the right thing and regain the community's trust.

The real, essential problem that I see (that actually might not be able to be resolved, because of the underlying conflict of interest between company and community):

• Moderators get to their position via a democratic process, an election within the user community of a specific site on the stack exchange network. ( And note: "all" users are asked to vote. It is not like the moderator community decides for new padawans, invites them and later announces "they joined the force". )
• Yet, "firing" them is something that solely happens within Stack Exchange Inc.!

I understand that the company is running the servers, and I understand their need to able to act immediately to emergencies such as compromised user ids.

But for any other kind of depriving a moderator of his status, the user community should have a say. At a minimum by having all other elected moderators being able to vote on that. The idea that solely CMs working for Stack Exchange Inc. determine what is going to happen isn't convincing. Any more.

In the past, the moderator status was subject to decisions by Stack Exchange Inc. only. And when 99.99% of the users trust the company, why not?

But going forward from here: if Stack Exchange Inc. is serious about listening the community again, then a "moderator firing process" that doesn't include some form of community input is not going to regain the trust you lost.

I can see the company POV: the company paying for the servers specifies polices, such as the code of conduct. That affects their business, which in the end pays the servers. That drives the need to deal with conduct violations, especially when committed by moderators. But that doesn't dictate that Stack Exchange Inc. should be the sole party making such decisions in a unilateral way. And as said: we are at a point where we need real actions that make a real difference.

Thus my suggestion: step back and find ways to incorporate the community into this crucial policy/process!

I mentioned this on the Moderator Team site, but never had time to come back to it and follow-up:

For a re-instatement, will the information gathered during the Discovery phase be shared with the former moderator?

It seems to be mentioned in the removal process that the moderator will be informed of what's happening, but it's not completely clear there, and it's not mentioned in the re-instatement process that the former moderator will be told specifically why they were removed.

Regarding the reinstatement process.

Community Managers (CMs) can recuse themselves if they feel they can't be impartial

Is it the case that a CM should also generally recuse themselves if they were involved in that moderator's removal process?

tl;dr: Does being removed as a moderator result in an ineligibility to stand for election? If so, does the ineligibility apply only on the site(s) where the moderator was actually removed/de-modded, or does it apply to all sites?

How does the moderator restoration procedure relate to a former moderator getting elected or re-elected by a community? I asked this a while back and never got an official answer.

The moderator reinstatement process seems to indicate that a favorable result for the former moderator results in immediate restoration of their diamond, without any need for an election or even any community involvement at all. Is this intended to imply that the former moderator may alternately choose to seek restoration by being elected or re-elected by one of the Stack Exchange communities, or is it intended as the sole restoration process and/or a prerequisite for becoming an election candidate?

This answer asks specifically about a former moderator seeking election on the community that they formerly moderated, but there's also the possibility that such a person might seek election elsewhere. The Initiation section of the restoration process states:

Previous moderators (PM) may request reinstatement through the /contact form on the site that they wish to be reinstated....

Thus, it isn't clear if this reinstatement procedure even applies to a moderator seeking election in another community. What happens in that case? For example, suppose I was dismissed as a moderator for cause from Shoes and Ships Stack Exchange. I contribute quietly for a sufficiently long period of time (e.g. at least a year, to allow any possible general site suspension-related bars to expire), eventually deciding to leave my old haunt alone and instead seek moderatorship on Sealing Wax.SE, which I have never moderated before. What happens in that case?

• Do I have to seek reinstatement on Shoes and Ships.SE before I can stand for election on Sealing Wax.SE?
• May I just go ahead and run, letting the community determine if I am worthy?
• Do I have to apply for "reinstatement" on Sealing Wax.SE despite the fact that I have never actually been a moderator there?
• Do I have to begin a "generic" reinstatement process that does not give me back my original diamond, but restores my ability to stand for election?

I once heard of a consulting project in the IT sector to get certain changes made on a bunch of computers. One million dollars and one year later, there was a gigantic multi-page flow chart with an overwhelmingly complex administrative process for how the changes would be made, that no one ever read, and the changes never got made.

These processes read a lot like that. You have a pretty diagram, and someone obviously put in a lot of work to write a process that sounds good. They can show it to their boss, and obviously they are doing the assigned task of writing up a formal process.

Except that no moderators have been reinstated, and there are giant, obvious, glaring flaws in the processes presented. (There are 43 answers already, most of them detailing specific flaws, so I don't feel the need to point out all the specific flaws in this answer.)

What I think is that you are attempting to combine incompatible goals.

You want to create a formal, transparent process for removal and reinstatement of a moderator that will be fair and equitable.

And you want to run Stack Exchange as a for-profit company, with people assigned to jobs using normal corporate reasons for giving positions. Including the people assigned to approve or disapprove reinstatement proceedings.

These goals are not compatible. You need to recognize that what you are attempting to do amounts to creation of a new system for justice, that will in fact be just and equitable and fair, and benefit the group as well as not harming the individuals to whom it is applied.

That's a tall order.

Ultimately, all the formal processes in the world won't help unless you recognize who you are attempting to restrain. If the company is restraining itself, and the company has the power to remove that restraint "in case of emergency," then the system is fundamentally, irreparably broken.

That's not to say that you can't have justice. You've actually done quite well for years. Like with a benevolent monarch who rules well, there is no outcry for more formalized procedures because people have justice. Conversely, when you have injustice, it doesn't matter what procedures you have in place to legitimize it and mislabel it as justice—people will still not be happy.

You can write up all the procedures you want, but at the end of the day it amounts to slapping a veneer of formality onto the arbitrary decision of a few CMs and thereby elevating that arbitrary decision to the righteous plateau of unquestionability.

I would prefer NO formal process at all, than to have a formal process that can be held up as an excuse for irresponsible or damaging actions.

If you see the need to take action, then take responsibility for the actions you take, and don't shuffle off that responsibility onto "the process."

That requires a high level of responsibility in a CM. But really, that's what's required anyway, to make the whole show work. Moderators and CMs show this level of responsibility all the time. Except in the case of Monica.

There already was a process for moderator removal. It was not followed. So let's not fix what wasn't broken. No process in the world can work if it is not followed, and attempting to "fix" it will just overcomplicate everything.

How can we trust you to use this process going forward?

I presume that Monica's removal would not have happened under the emergency scenarios. That means the previous removal procedure was at minimum ignored. I personally believe that SE just completely forgot the old procedure existed. So how can we have faith that the next time you want to remove a mod you will actually use this process?

Moderator Conduct Review Process:

From time to time, a moderator on one of our sites may violate their moderator agreement, or the company standards for behavior as documented in our Code of Conduct. Such behavior may eventually necessitate their removal from office, however both community expectations and our Code prescribe that, in general and whenever possible, they first be warned and given an opportunity to correct their behavior.

1. Does “be warned and given an opportunity to correct their behavior” also refer to Monica's case?
If it does, when can we look forward to seeing her readmitted?

There is nothing written in the two processes that suggest it cannot be applied retrospectively.

The Reinstatement Process mentioned checking for annotations deeming reinstatement impossible. I expected to see the criteria for determining and applying such annotations in the Review policies but didn't. What is the criteria for such annotations?

It looks to me like the removal process is still driven by a single CM, with the slight speed bump that they have to convince one other CM of their case (which is pretty easy to do, given that they're the sole participant in the discovery phase). The only fundamental difference I can discern between what's written and what was done to Monica is that there has to be a warning annotation (and without any description of how annotations are created/edited/challenged/etc. there's not really hard evidence that there's a difference there either).

In my opinion, a fundamental tenet of justice that's missing from this process is the ability of the accused to know the charges leveled against them and to see and address the relevant evidence. As long as all the evidence and all the annotations that go into this decision process remain opaque to the accused, there's no way for anyone outside SE, Inc. to differentiate this process from the arbitrary process used in the recent past.

What will, if anything, change to the process of re-instating a PM when a senior company official (Think CEO, CFO, CTO, Director of [insert department here]) was involved in the original revoking of PM's moderation rights.

Will CM's be allowed to overturn this decision by senior company official's?

2. CM1 will look through account annotations for reasons for removal. Annotations weren't always a feature, so CM1 will also consult with the rest of the Community Management and Community Strategy (CST) teams for any extra context.

1. If CM1 finds any annotation that states reinstatement is deemed impossible, the request is denied and the reasons for it will be stated back to PM. The process is concluded here.

There needs to be an investigation into the correctness of the annotation. As it stands, it's possible to apply a permanent, unappealable veto to any reinstatement request for any reason or no reason at all.

In a perfect world, where the majority of us still retain some trust in Stack Exchange, this would work.

However, we're not presently in this world. That means that this will work out to be at best a paper tiger.

A real impartial process would include a company external arbitrator or panel of arbitrators and a public evidence phase akin to the arbitration committee that Wikipedia uses for these purposes.

If you're serious about earning back the trust you lost, the process needs to be completely decoupled from the interests of the company, and the company needs to agree to be bound by the decisions of the external arbitrator (s). Moderators serve their communities, and it should fall to members of these communities to assess the evidence and make a fair determination.

One question specifically about international sites.

2. CMs can recuse themselves if they feel they can't be impartial. CM1, CM2, and CM3 will be picked at random from remaining CM pool.
3. CM1 will review data and research concerns - look through account annotations, query the involved parties, etc. If there is a note to check with the Community Strategy Team (CST), they'll do so: the CST may have information relevant for next steps.

There is only one CM responsible for all international sites. So, even if CM recuse himself when he feels he can't be impartial, other CM's will learn about the situation from his point of view. For SOru it's especially important, because there is only one russian speaking CM in your team. Basically, the whole process in the case of disagreements between mod and this CM will be biased.

Any suggestions for this case?

Moderator Action Review Process

Proceedings

9. If the consensus was for removal, we will then revoke the privileges of the moderator to be removed, and also remove the moderator's name from the election and /users?tab=moderators pages.

Why?

I get the part about removing them from the moderators tab (because they are no longer a mod), but they were elected/appointed -- so why "rewrite history" by removing them from the election page?

This [reinstatement] process is available to all moderators, whether elected or appointed, regardless of whether they stepped down on their own or were removed from their position for any reason

Pro tempore moderators of beta sites currently lose their position when a site holds its first election after graduation. They would now be entitled to being reinstated at any time. Is this an oversight or a policy change?

What is even the point of writing down this process? If SO doesn't follow it they will just say "Yeah we didn't follow the process, but we don't re-litigate the past".

SO has demonstrated clearly that the current guiding principle is to stick with their decisions no matter what. If process was followed or not, if actions were right or wrong, if people providing value on their sites like it or not - it doesn't matter. Clearly "stick to your decisions" is the guiding management principle, overruling all other concerns. At most some empty words with vague promises for the future will be announced.

So the party in power is not interested in fixing mistakes or modifying past decisions that turned out bad. Something is done and then that's that. If some process wasn't followed - well, bad luck, maybe next time. There is no indication this will change anytime soon.

So, what's the point in writing down rules? They can be ignored whenever convenient.

I really don't get it.

If CM1 finds any annotation that states reinstatement is deemed impossible, the request is denied and the reasons for it will be stated back to PM. The process is concluded here.

Moderator was elected by a community. It looks strange that one CM decide mod's fate.

And even this:

Community Managers (CMs) can recuse themselves if they feel they can't be impartial. CM1, CM2, and CM3 will be picked at random from remaining CM pool

If there be any disagreements between mod and one CM, the rest, obviously, will be on the CM's side. How can we be sure of an impartial CM's attitude in this case?

Could we rename one or other of these policies? "Action Review" vs. "Conduct Review" doesn't make it particularly clear what the purpose of each policy is - the former is for mod teams to use in case of breakdowns, while the latter is for broader community concerns and for CoC/agreement violations.

Calling the former policy something along the lines of "Moderator Team Review Process" should serve to clarify that - it's not actions we're really reviewing, it's whether or not the team can continue to work together.