SIP 1 - Continuous Funding

[Rieki]

By @nik @julioholon and @Rieki

Summary

This SIP (Seeds Improvement Proposal) proposes to change the global proposal funding mechanism to one of liquid dynamic fund allocation by vote.

Seeds delivers perpetual funding to leaves, regardless of which organizations are included in the beginning, regardless of what direction any of the DHOs are taking, regardless of individuals who may come and go.

We are therefore building a growing living financial system which takes care of itself.

Problem

Currently individual proposals are made against the funding pools- It is very hard to assess whether or not the funding will be used successfully - it takes a lot of time to do due diligence both before a proposal is voted in, and during the proposal evaluation phase. All citizens have to do this due diligence, but, since it’s so much work, citizens generally vote on what they like / don’t like, and don’t evaluate. There is a lack of accountability.

The voting model will work better in a regional environment, where everyone knows each other, and at minimum 144 citizens vote on local projects

But in the global context, the current model cannot not scale to millions of users - either the funding will need to be kept restricted such that no more than about a dozen proposals can pass each round - artificially restricting funding - or the number of proposals will become unmanageable.

Solution: Continuous Funding

The Seeds harvest will continuously deliver newly minted coins to recipients that are being dynamically voted on by citizens. The % of vote determines the % of funding.

Continuous funding will ensure that the existing Seeds funds constantly stream towards regenerating the planet and towards growing the value of the Seeds network.

Focusing funding on organizations - which then can fund projects as they see fit, and install their own mechanisms for accountability - removes the burden on voters to do due diligence and follow-up due diligence on every project they are voting on.

Voting - Citizens Decide

Any citizen can at any time re-allocate their existing vote points from one project to another, or a percentage thereof. The length of time a vote remains with a project is counted. Citizens who don’t vote have no impact.

Calculating Inputs

The weight of a Citizens input is determined by their contribution score. Each Citizen gets “Funding tokens” = to their contribution score to distribute amongst their preferred organisations. E.g. A Citizen with an 80 CS would get 80 tokens and if they give 50% of the budget to 1 org that org would receive 40 tokens.

Conviction Voting

Staying longer in one vote is rewarded by the vote having more weight. This has many benefits - may consider this in the future or even right away. [TBD]

Slow release: when a citizen selects different %'s only N tokens move over each day causing it to take N time to fully reflect a vote change. (TBD, not as necessary with seasonal resets).

Who can apply?

Any Seeds organization can apply for this funding - we envision the Hypha DAO to apply for this funding, and any other DAO/DHO organizations can apply as well. Any number of DHOs can be funded. Having a Seeds organization account is the only hard requirement. [TBD] Perhaps - only reputable organizations?

Funding

Each project gets funded by the share of votes they’ve received, receiving an equal share of the harvest. E.g. 50% of the votes means the project gets funded with 50% of the harvest. Harvest gets paid out each new moon and full moon.

Minimum Funding Threshold

To prevent gaming the system, an organization or project must receive more than a 10% [N%] of total votes. Less than that, nothing is paid out. This is to discourage and not reward spamming or other forms of coordinated fraud while also doubling to ensure that organisations can recieve a meaningful revenue stream, rather than it being split so thin that it becomes meaningless to many, rather than useful to a fewer set (max 10), Citizens always retain the power to adjust this as our revenue stream fluctuates so we can meaningfully share revenue.

Maximum Threshold [TBD]

The goal of the system is to encourage projects to take as much as they need, but not more. Organizations can set a maximum threshold which makes it more likely they get funded. Leftover funds are … (TBD)

Reset

Every 3 moon cycles all votes are reset to prevent stale votes from being given out or stale projects to keep receive funding. This means citizens have to recast all their votes every season at minimum (following the quarterly sessions to inform voters.

Pro: Removes stale projects and stale voters

Con: Added friction for citizens

Can probably be solved very neatly with our vote delegation system

Pairs well to do sense-making with decision making so that voters participate in or watch the seasonal “Regenerative Roundtables” and then update their votes.

Further, this pattern of setting votes each season will give much more stable streams to organizations (as most votes will likely happen once after the Roundtable and stay until the next).

Flagging

Flagging a DHO/Org (using the standard Flag threshholds) makes it ineligible to receive funding (if N flags are given) for policing DHO’s that are no longer active or otherwise fraudulent.

Before Golive

At golive, the harvest is calculated from the QEV (qualifying economic volume).

Before that and until they run out, we can use the gift.seeds seeds that would normally go to proposals to be voted on, and instead give them to DHOs which then give it to proposals and projects to be voted on.

Funding Ceiling [TBD]

Some DHOs may not need more than a certain amount - they should be able to set a ceiling. Setting a lower ceiling should then create some rewards to encourage all DHOs to take what they need but not more.

Some DHOs may have minimum funding requirements - they should be able to set a minimum, below which their project isn’t worth pursuing.

This need to be thought out carefully, hence TBD

Appendix

Map

As a map and guideline, we can look at how block producers are rewarded in EOS and Telos - votes vote in the top 21 producers, which get to produce, awards are delivered based on number of votes. Instead of block producers we would have DHOs / DAOs which work on various elements of Seeds.

[ChuckH]

Question re: Conviction Voting

Is the proposed Conviction Voting algorithm described in more detail somewhere? My only reference so far is Conviction Voting: A Novel Continuous Decision Making Alternative to Governance | by Jeff Emmett | Giveth | Medium , which sounds a little bit different from what is described here. For example, the “Slow release” function described in SIP1 here may be already accomplished with the “leaky bucket” model.

I also have a new(?) idea to add: vote power decay. The algorithm is similar to the existing proposal vote decay (100% power for time duration T1, then decay to 0% over duration T2) but with a very different purpose. The intent here is to have a sort of continuous Reset going on, and it may remove the need for a discrete Reset described in SIP1. If you “set and forget” your preference, it will eventually be forgotten (T2 time). You have to remember to bump your conviction setting regularly (T1 time) to express that “yes, I am still supporting this project”. Maybe I should make a graph of this to communicate more clearly.

[Rieki]

For testing this all out, I’m very much in favour of Justin’s suggested approach (high trust and more centralised) as we figure out the right protocols to do this completely decentralised

e.g.

do we limit people from putting more than n% on a single org?

Do we mandate people put at least N orgs to avoid them only giving to their own?

etc?

This will all require some lived experience that I think we can use a spreadsheet to play with now as we work out these details.

[anahealingwords]

@Rieki

at minimum 144 citizens vote on local projects

at minimum 144 citizens can vote on local projects, right? Not required to…

@Rieki

removes the burden on voters to do due diligence and follow-up due diligence on every project they are voting on

It doesn’t remove it, it delegates it to “organizations”.
Due diligence shouldn’t be a “burden”, it could be a source of joy and celebration: to be aware of all the beautiful regenerative projects the SEEDS community is creating could be awesome imo🌈
Is it the case now? nop, i know…
I do think regionally/topically choosing projects & funding is in itself a great idea.

However, I have a big issue with delegating votes
I understand that most people in SEEDS do not see delegation as an issue yet so I’ve resigned to the transparency aspect that does seem to be held as a vital component of the ecosystem

So as a little “warning” of potential/probable harm I see down the line, I propose that it is in every org’s mandate (as in basic requirement to be elligle for funding) to consciously communicate to and with citizens what the projects are, how much of the funds is allocated to whom, etc etc… in EASY and SUPER ACCESSIBLE ways!

I see this as less about “trust” and more about “awareness” and the empowerement that comes with it:
One can trust a particular org yes, and also be a curious being that wishes to know and understand what the org is funding and why.

IMO if this isn’t a foundational aspect of every org we are simply walking straight into a “representative” democracy…We’ve been there, we’ve done that, we know how it goes…

@Rieki

Citizens who don’t vote have no impact.

We are talking about voting “yes” or “no”, right? Otherwise, this could be an issue…

[minos999]

I was exploring how it may work and have some questions:

• How do we compute different inputs % from citizens to come up with one value for each DHO? Is it average? Is it weighted average based on contribution score? Other ways?
• Do we consider any criteria for DHOs being eligible … maybe a min no. of contributors … or enything else?
• Is there a way we can connect the process with existing metrics in SEEDS ecosystem or criteria for organisations? I haven’t been much involved in those conversations, would it serve us to bring some of the conversations here?

I am exploring all these because of some concerns I have in relation to how the vote could go from what I am seeing not working on the Co-op.
From my observations it seems that the practice is for most citizens to vote on the proposals they already know (or the people behind them whom they trust) because this is easy to do, without any additional due diligence, which requires lots of time and different engagement. I do this often times.

There will also be the bias for people to add % for the group they are part of, which is natural in many ways, we are part of that group and want to support it. In order to know the contributions and impact of other groups we need to spend energy on it, which some may do, some may not do.

[julioholon]

Hey there,

So, my main concern with this proposal is that we shift voting on stuff that matters to voting on groups that you like. Then voting for budget based on personal preferences can lead to most popular people or groups getting more funds.

And since this can be changed anytime and is transparent on chain, this can lead to even stronger shadow politics games over this “consensus”.

The problem is that you obfuscate the intentions of the voters, by just capturing their budget percentages.

Here’s an example: When I say 70% to Hypha, there’s no way for me to explain why I choose that. So suppose I see a good reason as to why Hypha should get 90% of the budget (maybe because they want to build a great new project that will help SEEDS a lot), but since there’s no place for me to say that, and I can’t communicate to enough people, and most people vote Hypha should get 40% or less, so averaging it out Hypha gets 42% of the budget, not enough for them to deliver what they want to deliver.

Here’s an alternative, that also address the Problem or cognitive overload for proposal evaluation and due diligence:

Continuous Conviction Funding for Global Pool
(Very much inspired in Commons Stack Conviction Voting framework)

• Only Global Seeds Organizations are eligible to create Funding Proposals to the Global DAC pool (Citizens can still create Global Referendums)
• Citizens have voice equal to their CS, and can change the allocation of their tokens at any time to the active proposals, to indicate their support
• Tokens take time to move: say 1 token per hour, so it takes 4 days (96 hours) to change your whole pool of votes, if you have CS 96
• The number of needed votes to pass is proportional to the budget requested. If the proposal requests for 1% of the budget, it needs 1% of the votes (there can be a configurable minimum and maximum too)
• Proposals are approved as soon as the number of needed votes is reached, then that votes return instantly to the voters pool

Funding Categories

• Every Funding Proposal has to have at least one Category, from which the budget will be taken (so a Farming DHO will most likely create proposals to use the Farming category)
• The eligible Categories are configured at the DAC level, and can be changed by Referendum
• Any Citizen can post a Referendums to change the distribution from the Harvest in Categories (like 40% Farming, 20% Movement Building, 10% etc.), where they explain WHY they want to change the percentages, if they get enough votes, it passes.

Continuous Conviction Funding for Regional Pool

• Same as above, however Citizens and Regional Organizations can submit budget proposals to the Regional DAC pool
• Regions can have Categories and allocations different from the Global DAC

PS: If you don’t know about DHO and DAC, check here.

[Kath.PlanetHive]

Thank you @julioholon for this different view on the proposal. I like the idea of continuous conviction funding you bring to the conversation. I can see how voting and budging could flow. It makes sense to me.

I feel this is a very technical process oriented viewpoint. I am wondering how we might build scaffolding - people, conversation, culture - around it?

[minos999]

Some further thoughts:

1. The framing of the problem seems too narrow: lack of due diligence from citizens and inability to scale. Are these the current challenges?
   Some further aspects:

• SEEDS becoming too complex for only Hypha to contribute to it, Hypha currently a bottleneck and unique place for sustainable compensation
• SEEDS Ecosystem mature enough to have new orgs contributing to its growth
• Hypha willing to evolve Hypha milestone to SEEDS milestone (or landmarks)
• Need for mechanism to distribute common funds to different orgs (milestone funds and funding campaigns? before Go Live, harvest funds after Go Live) - are we looking for a universal solution? Why? Do we have same conditions?

2. Solution: lack of due diligence not addressed by solution, Julio’s point of vote for favorite group rather than stuff that matters pretty valid.
   Scaling still seems challenging with different orgs doing different work and hundreds of people need to make sense of that and vote coherently.

3. Who can apply? Any no. of DHOs can be funded - that’s a challenge, we see the limitation with campaign proposals now, it is difficult to do due diligence and share votes in a coherent way with many options. Further criteria: orgs with some history of working together, who issue their own token and have agreements of how they operate and a transparent way of showing how the funds are being spent, reputable yes! TBD.
   Can regional DHOs apply? (guess it was DCOs)

4. Minimum funding threshold. min 10% means that max 10 orgs could be funded, not any no. of DHOs as stated above. 10% good to start with, not sure how plays further on.

5. Before Golive: milest.seeds not mentioned … only gift.seeds? More details needed here: Funds from milest.seeds will be automatically transferred, all of them just like that? There is a complete disconnection of the readiness to grow in this case, when does that come into conversation? Or how far we are on the Go live path? Also, how the funding campaigns system connects here? Same pot? Different? Can orgs apply for both funds from milestone and from funding campaigns … TBD

6. Minimum Viable proposal - How the new proposal looks like? Update of the Token Allocation? How does a Good enough for now safe enough to try proposal looks like?

7. Challenge I see with % for orgs. Feels like a roulette.
   For eligible orgs: hard to come up with a proposed %, as there is no idea of what 10% means, or what 40% means. In terms of HUSD they can be the same depending on the amount of Seeds sold. Seems like an invite to cooperate and support the selling of Seeds together. Do we want to have this a goal of the ecosystem? Or a secondary goal, other than ready to grow and readiness for Go live? How do we know if we are not selling to much before actually getting at Go live? Some thoughts to consider.
   For citizens: Challenging to find a practical way in which a citizen can come up with a coherent proposal and meaningful %, have no idea how I would do it myself without getting into the biases and the patterns we currently see on the co-op voting, even with enough due diligence.

8. Getting input from citizen in a spreadsheet … could the gratitude system and the bots work to bring info from citizens in a data base, benefiting of the signing of transaction process? Exploring a way to do it easily before having a code for it…

9. Conversations between different orgs/DHOs participating feels crucial for coherence. That’s a huge challenge. Working very hard or not really up to now with only few players. Definitely need support on this side, maybe from an org or lab focused on this specific interrelation between DHOs in the SEEDS Ecosystem on the way to GoLive and further.

[apostol]

Guys, I want to point out (Very Important!) that I notice we’re mixing two different processes that can and should have two different governance structures to function properly.

Milestone Funds and SEEDS Sale - the purpose of this is to support the R&D of the whole ecosystem.

• I think the simplest way aknowledging the time constraints is @markflowfarm’s proposal for Milestone 3

Milestone 3 Proposal (from July 29 meeting)

Many beautiful conversations and tensions explored together on the zoom (expertly facilitated by Tyler!). Here is the simple proposal I posted on the chat. For continued discussion here. =-=-=-=-=- For me, this can be a focused decision for an Aug 8 proposal for Milestone 3. Simply, 39m Seeds will go into the global SEEDS sale (via SEEDS :: Buy Seeds), and the proceeds of that sale will be shared equally among a list of organizations. The set of organizations would be listed specifically i…

• For Milestone 4 we can come up with a system that is similar to what @julioholon describes with some form of conviction voting

SIP 1 - Continuous Funding

Hey there, So, my main concern with this proposal is that we shift voting on stuff that matters to voting on groups that you like. Then voting for budget based on personal preferences can lead to most popular people or groups getting more funds. And since this can be changed anytime and is transparent on chain, this can lead to even stronger shadow politics games over this “consensus”. The problem is that you obfuscate the intentions of the voters, by just capturing their budget percentages. …

Co-op Proposal process - due diligence & follow-up & imrovement of voting.

These funds in my opinion should focus mainly on regenerative and expansive efforts - funding projects that create circular economies(demand for Seeds) and also do regenerative work on the ground - like planting trees or supporting projects like Joe Brewer’s Earth Regenerators.

I think it’s a good idea to set up a system that enables the due diligence to be done by an expert team before the Co-op proposal is put up for voting (Setting up something like Regenerative Score, Distribution Score, Circular Economy Score) so the citizens don’t need to do so much work upfront by themselves. //cc @DaYa. Of course, there is a problem with how objective that due diligence is, but still I think it may do more good than harm at the current stage of the ecosystem’s development.

[andythegreenie]

I completely agree that due diligence on proposals is not a burden but a source of joy and is very educational as it teaches me a lot about SEEDS generally and makes me look hard at a lot of very inspiring organisations and projects and helps me with formulating my own regenerative projects. But the due diligence is sometimes hampered, for me anyway, by a lack of information about some of the proposals and the organisations making the proposals, and the lack of metrics about the bioregional contexts in which those proposals are meant to be implemented. I don’t know how best to solve this, but I notice a dearth of feedback from other Citizens on some of the proposals before they come up for voting, which makes it hard to understand sometimes why some proposals end up with such a resistance to them during voting, and the people making the proposals may also not know why the resistance has emerged so suddenly. Perhaps a snap poll of each proposal amongst the Citizen electorate before the formal voting process begins might give proposers some valuable feedback, with this polling facility integrated into the initial proposal post on the Passport Forum and/or the Co-op Proposals discord channel?

[DaYa]

For clarity:

You propose to replace the current coop proposal governance and distribution with a completely different model?

@Rieki @nik @julioholon

[MaxVBohn]

I have been writing my response here a day after is was posted and playing it through a couple of times and explored it in my imaginary world
My response was so in depth that was just too much.

I agree with the essence of what Julio and Irina bring. It doesn’t provide more clarity or structure to actually solve the problems we want to solve.
It doesn’t reduce the time needed for being a diligent Citizen or makes the process easier.

@apostol I find it unnecessary to separate funds for “regenerative and expansive efforts” and go back to limit Milestone Funds to “support R&D” as if it is separate from the first where we could see in the Milestone discussions that it is not.

Ideally all the campaigns should provide a lot of things important for the theoretical Milestones set by Hypha. From feedback to growing the movement, measuring the regeneration and exploring new ways for it.

Due diligence is needed for both Milestone funds as well as Co-op funds and it is basically looking at the same thing: “Are you doing what you said you would?”
And there must be a way to say “Yes you do, here are the funds” or “No you don’t, we won’t give you the funds (anymore)”.
These two buckets of funds can definitely be managed with the same framework.

Again I don’t know if it is necessary to separate these funds. It would all depend on the mechanism in place to manage it.
But I agree how this proposal is functioning it wouldn’t work for both.

The other thing @Rieki already is implementing according to the last compensation video is making Campaigns into Invite Campaigns where the funds are in escrow until they are send out via invite and can be taken back at any point by the Citizens.

So I also don’t know how these two proposals really match.

[DaYa]

@Fabi
@Rui
@Magickdragon

[derjogi]

I only recently (well, now a while ago, but similar to Max it took some time to phrase a reply) discovered this thread and took a while to digest it… Super interesting. (And I am awed by how fast some of other people’s feedbacks came in, I always have to take a lot of time to digest things like this… )
Unfortunately I’m not that well-versed in the seeds metaverse yet, so much of what I’m saying might be out of place, feel free to ignore or just take the parts that seem valuable in this context.

First, let me give some general feedback, and further down some more specific feedback to individual points / questions.

General feedback:

The main issue as laid out in the problem statement is that there’s too much to vote on to do proper due diligence. (Irina had some more aspects, but to think about these is definitely out of my scope due to lack of understanding)

I agree with @anahealingwords and @andythegreenie that mandatory good communication in an easy & accessible way with good metrics would be a great (and some time in the future the ultimately necessary) step to improving that. (We’ll need to become good at sensemaking, and clear communication is the most important part of that!)

One thing I’d add to that is that this communication has to go both ways. I believe we need really good and easy ways for interested parties to leave feedback and possibly even adjust the proposal themselves (or ‘fork’ it for an alternative proposal). Voting is a good way to check the status quo and to decide whether to go ahead or not, but the most important work is to get feedback, so that we can iterate on the proposal until as many as possible (everyone?) and everything (environment, society, economy, …) is satisfied.

I also believe however that fine-grained delegation can be a really good thing (@anahealingwords curious to hear what downsides you see about delegation).

Lastly, and possibly most impactful, I believe it can be a good thing to get away from requiring as many votes as possible from as large a group as possible, and instead I’d propose that only those that are affected by a proposal have to be encouraged to vote (but still making it open to all else who want to).
With that, we wouldn’t need to measure approval as a percentage of the whole any more, but could either do it as a percentage of whatever subgroup it is in (which I believe SIP-1 is kind of promoting?), or we could even have some absolutes instead of relatives, e.g. requiring a minimum of 10 approvals with one additional required approval for every 100 seeds (or whatever), and and could only have a maximum of 5% negative votes; impact metrics could be used to set some of these values, etc… (With tech nowadays there are sooo many different possibilities how this could be tweaked and expanded).

This would encourage the promotion of issues to exactly those that are affected, imho an important way to get the tasks of doing due diligence into the right hands.

Related to that there’s one big change though: Under SIP-1 the funds would be distributed proportionally from the available bucket. This means, that every single proposal affects the others, because the more Proposal A] gets, the less B] will get.
I would love to see that decoupled, but I am not sure whether that is possible with the way how seeds is set up?
I.e. is it possible (or already a reality?) to ‘mint’ seeds based on approved proposals?

E.g. If in one moon cycle there are 5 approved proposals asking for 10’000 seeds each, we’ll create 50’000 seeds, if there are 200 approved proposals asking for 40k each, we’d create 8’000’000 seeds. (?)

I didn’t dive enough into hot the seeds economy is working at the moment, and concede that this might be a really big change and would possibly require some other adjustments, e.g. some mechanisms where unused seeds could disappear over time or so. Probably too big of another topic to talk more about here, so won’t go into more here, just wanted to state that these things are all connected, and changing one might well mean we’ll have to change these others as well.

There are of course alternative approaches, some of which are probably much better in the short term (e.g. expert teams for due diligence)

Ok, now on to some more specific feedback:

1. Resets: Do we need them? Could we also have the orgs/companies/proposals set a desired duration for their project (“Planting 500 trees” in October, we’ll need to have everything prepared by end of September → Deadline: 30.09.; if not enough votes in favour by then the project would be reset and need to be re-submitted to stay active), or possibly an amount of desired seeds (need 300 seeds per tree → 150’000), and as soon as that time is reached / amount is received (cumulative) it will automatically stop/be archived and they’ll have to create a new one / re-list if they haven’t finished yet?

More about Resets

(Incomplete list of) Advantages:

• Less overhead for voters (I don’t want to have to revote that often if avoidable, especially if there are potentially hundreds of different proposals that I’m interested in!)
• More suitable for different kinds of projects, some may be one-off projects that might need a lot upfront and won’t necessarily last 3 moon cycles (or whatever the Reset rate is)
• Forces projects to think thoroughly how much they might need / how long they’re going to run, voters can see these stats and consider them when voting

(Incomplete list of) Disadvantages:

• ‘Stale’ Projects that predicted a longer runtime than what they actually need or more seeds than what they need might ‘unfairly’ receive more → incentive to over-predict. But I think that’s actually a big problem, and possibly better than the alternative of giving false promises (we’ll be cheap and fast, but throughout project the budget/timeline explodes)
• Financial cycles and flows might be a bit less predictable? (But more steady?)

Some projects might be long-running low-budget (e.g. a 5 year project to take care of a new forest will require a long time commitment but not a high budget); others might be expensive but short (buying an expensive machine for my village), both of these might work better in a system where the onus of setting the right amounts & timeframes are with the proposer (and voter) instead of forcing a 3 moon cycle for each project onto them.

2. Re @julioholon 's Continuous Conviction Funding:

   1. Voting on Groups vs Proposals
      I think it will have already come through in what I wrote above, to me it also makes more sense to vote for proposals rather than groups. Why don’t we just have both, the ability to vote for either, the proposal that’s been brought up within a group or the group itself? In that case the group basically acts as a delegate, and we’re at Liquid Democracy, which imho is a great thing. (again, @anahealingwords please share your wisdom why possibly not )
      If you have insight into a specific proposal within a group then vote directly, otherwise give your vote to the group. Or do both.

   2. Returning votes

      Proposals are approved as soon as the number of needed votes is reached, then that votes return instantly to the voters pool

      Questions to that:

      • Do we need to have a cap on how many votes a user has? What’s the ups/downs of capping these votes in different frameworks? (I do have my own thoughts, but not particularly well formed out)
      • Do all votes return to the voters? If votes are being ‘returned’, would that mean the proposal will then ‘lose’ these votes? I think this goes against the benefits of continuous voting (funding), so probably the proposal will keep them, but the voter gets that vote (assuming only one) credited onto his account available to vote elsewhere (but not again on the same issue?).
      • If people can give multiple votes I’m wondering whether this might promote ‘gaming’ the system a bit more: If I see that a proposal is shortly before passing, I’d increase my votes to that proposal to bring it over the threshold, and would immediately regain all my votes and can distribute it to others, potentially repeating the same thing a few times.