Internet Engineering Task Force 102, 19 Jul 2018

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: IETF102-TRANS-20180719-1100

Description

TRANS meeting session at IETF102
2018/07/19 1100

https://datatracker.ietf.org/meeting/102/proceedings/

A

Hello welcome to trance.

A

hmm I'd hoped a more verbose link to the note well, but um if you've seen the note well pay now and just assume, this is the same. One.

A

So on our agenda, there's not much as to 69-62 piece if we've talked about threat analysis and nothing much has happened. So those are our item schedules. Does anyone have any other items?

A

Ok, so then, let's continue so the 696 to do base. We have six items left. You can see them at the bottom of the screen that we would like to resolve so that we can get this document out. So the first item is the document status has been some description on the list changes from proposed Stannah to experimental I'm. It seems that everybody is okay with extra mental, but there was some pushback. So if there's anyone who wants to say something about this.

B

Yeah rich Sol's, Akamai, yeah I, was I. Think the major push back against it was like. It was more about a process question and there was question. If there was support to do another round of experimentation, it seems that there is so as long as there's consensus on finally.

C

Yeah I would agree with that who.

A

Are you exactly.

C

Malik did assert it's unfortunate. The nature of public certificates are that we have to do experiments on the public Internet and that's some what's undesirable at times, but it's unavoidable in order to figure out how these things work at scale. There's no there's no test environment. You can use for the public, Internet and so yeah.

D

To clarify the the we're doing an experiment on this protocol, it's at the protocol itself because it hasn't been implemented and tested is experimental state. So you know we. We believe that there's a commitment to implement over the next few years, but just it's you know it's transitioning from the current format to the next format. It's gonna take time. It's gonna, take effort, so yeah.

C

I completely agree with that comment. You just this, it's the only thing we can do and yes, the yes, the new improvements, we're making. The new improvements were making our very important and it's important that we get them out. There do.

D

You guys run a CT vlog yeah. We.

C

D

Can you say anything about transition transition plans? Is there anything that you could we.

C

We support the draft we intend to we're working with Google. We were at their CT days. We work very closely with them, having the only other implementation other than theirs, and so yeah we're very supportive about this. We have internal desires to do even more exciting stuff, with CT that I hope I'll be able to talk about in the future. I.

E

I'm, a Devin O'brien from the Google Chrome team, so kind of piggybacking off of what Tim said. Oh, we actually do have plans to begin experimentation. I mean right now our priorities have been bringing CT enforcement, and so that has taken our you know: limited engineering resources and sort of funneled it towards there. But within the next, probably six to nine months, we we're laying out our plans to do experimentation to sort of get like hardened data so that we can back up some of these like aspects of 696, you list.

E

So it's on our roadmap. We don't know exactly have that roadmap ready to share yet, but as soon as you do share it on list, um but I just wanted to like voice that this is absolutely something we stand behind and we're gonna be working on it just right now, not just not yet.

A

Okay, so the next item- and this came from a Curry's- unfortunately not here- but he suggested to remove the preventing track and Klein's claim because it was a bit too strong and it would cause problems with RSA PSS. So we've got a full request for that and the question is: does anyone object to this claim to be removed.

A

Okay, so everybody seems okay with it. We see here no objections. We got a remote on so let's see.

A

Go ahead, Andrew, hey.

F

Hi I'm Angie. Oh yes,.

A

F

That I did right onto the list to explain why I object and the reason why I think there's been a misunderstanding, which is that using a deterministic signature scheme isn't the only way to achieve this.

F

It's also possible to just sign a message like an sth just once, and then you know just store that signature and just return that signature for future requests for the sth and I think this is important because we don't know yet exactly how gossip is going to work and it might not be possible to deploy gossip experiments if there's a concern that that's going to lead to tracking. So this is the sort of thing which we could always. You know remove this requirement in the future.

F

But if we don't have that requirement in now, then it could lead to trouble getting gossip experiments going and I'm, not the only one who has expressed an objection to removing this section. Tom, Ritter and Linus Nordberg also chimed in on lists, saying they object to this.

D

So obviously this means really discussion on the mailing list before.

A

So then, the next item also came from a cursory few. Is the no security implications claim georgg hackers not here.

G

B

A

Okay, so some me not so let skip this for now. Mew I'll go to the other one. So there's another proposal to clarify what we mean by the current ntp time. There was there was some some discussion but I think, since these are new items, I think we really need to be be a little strict and say if someone doesn't give us text- and we have nothing to talk about that- so will not continue with these items. So if anyone really strange feel strongly about this, they should actually volunteer to also provide the text.

A

So does anybody here feel strongly about this.

A

Any concerns from this, like do people, don't care like.

H

A

If it's not this important, then we're pretty late in the process and we're just not going to do anything with this. So so we'll see if we can get focus to propose some text and then maybe we'll see if it goes anywhere, but it's just not not a strong participation, we'll just come to drop it.

A

So next item is the using JSON format for error. Reporting. I would hope. We get a little more input on that item because you have implementers here, I.

E

Koribo know Trustwave, obviously I'm the one who proposed this. The idea here is that in section 5, the actual response body that defines any additional error data is that as it should, there's no actual requirement for logs actually returned any data just besides the HTTP status code.

E

But my thought here was that, if we, if we use the JSON problem, reporting format logs can send back or metadata so that consumers of the log can basically automatically respond to any errors. One thing that came out of us CT days is the potential.

E

Some thoughts around rate limiting based on roots or intermediate CAS, if, if, if submitters to those logs, could get more information back from the lot of operators, say, like you know, you're being later limited because you're intermediate here and in change being currently what rate limited try back in two thousand seconds, or something like that, that information can be provided through this format and what's nice here, is that the Jason response format?

E

It doesn't mandate that you support any extension metadata fields so at a later date, if we decide that some CT log endpoint wants to go and return additional metadata, that can be done at a later date, maybe in like a further draft. This is kind of just laying the framework such that we have this error format to fine now and we can leverage it later.

E

Okay, so are you willing to write up some text for this? Yes, Corey again, um yeah sure I'll, give it a shot. Okay, great thanks, yeah.

C

Tim, how big I'm, assuming that Corey I'm, assuming that you're modifying a should here right? So your are you thinking that this is going to be a proposed format that we may change later or not sure I'd want to make it mandatory?

C

Well, it's still kind of half-baked, but I, really like the idea of having a format that we can start experimenting with and returning error data, because one of the feedback that we do get from lots of the people who consume our logs is in fact that they get error response codes back and they have no idea what's going on and what they should do. In response.

E

Quote from Trustwave again um so the the JSON format it's actually defined in RFC I, think it's 78, oh seven, and it defines in there there's a couple standard fields in that Jason problem document, but it also allows for extensibility in the future. It basically says that if a if a client consuming that problem document sees a field that it doesn't understand, it must ignore it. So if that's not treated as an error case, if you know it'll later at a later point, we want to add some additional metadata.

C

All right, Tim LD, so your proposal would be you. You should report errors and if you do, they must be in this RFC format. I.

E

Inquiry again for justice, us so yes, well, I, I, think it'd be great. If we could say must, like you returned so I believe from every endpoint to find in 69-62 beast. There are error, codes and error messages to find. It would be nice that you know that they all come back in this problem document. The document format.

C

Yeah, just because of the fact that it's kind of late- and we haven't really played with this I- think I personally, would prefer a should and should.

E

According from trust with um so yeah I mean I, guess if it is late in the game, you know and there's no provision to change this and I guess you know maybe just remains it should right.

D

If we're talking about publication as an experimental standard at some point in the future, you know if everything goes well, we're going to transition to a different standard status, and at that point that's a new document, and if this turns out to work well and it's and people like it, we can turn it into a mosque. But you know it should for now.

A

Okay, now is the last item. I, don't know where this, where this came from it just was. It appeared on a list of of items that were sent to the list, but I don't know who originated this and I can't find any discussion of it, but the option for City log so signal when rejections were due to rate limiting and again, we've got the same thing. If this is important and we would need a volunteer to write text.

C

Tim hobby I would think that Cory's proposal covers that adequately.

A

That's actually true: okay, okay, any other 69-62 biz items, not so we're just sort of waiting on a curve.

A

Okay, so, okay, so so we'll skip those items now. So let's go back to the D Trans threat analysis, so this has been mostly a work of steve, canned and and and then the last about three IETF meetings. I think we see you know small updates every four months and two people mostly arguing about what it should be and most other people not caring anymore. What happens to this document? So we really don't want to keep the working group open for just this document.

A

So we're sort of pondering what to do here and it doesn't seem to be that much interest talking it's pretty far finished as it's also a ways of it if it doesn't get finished. But but we really don't want to have more more infighting between two people and just waiting forever on this.

D

You know I I, think the one way to frame the question is: is the value of the document greater than the cost of getting it closed? And at this point right and at this point we're just not seeing a lot of interest in the document right? It's it's I, don't and it's hard to know if it's that people have lost their taste for discussion of it and but and that they really do care or that they just don't care. I should say you just don't care so.

B

Rich Sol's Akamai. Well, you know it's got like the same level of interest that the rest of trans and gossip has these days. I would say: yeah, maybe less interest, because it's been worn down by the infighting flip a coin call for consensus and move it forward. I think it's a worthwhile document. It's a shame that it's going to get wrapped around the axle over this one discussion of what is or is not an attack or how to describe that attack, cut the whole thing out.

B

Then I, don't I don't care, but yes, I'd, like to see it published. I can leave it at this point. You know it's only those who have egos involved in this fight.

D

We'll take it back, Paul and I'll discuss it after the meeting and then take it out back out to the mailing list.

A

Okay, then, just a summary on on some of the items that haven't really seen much much progress as it's the gossip city, gossip decline, behavior end up, binaries, I, think I think it would be good if that, if we see more work going that way that that's just a new working groups will be started up or above started up and did that work there I don't think we need to remain open for it, because the the progress is really really slow. At this point,.

D

We also seen this with the NTIA sponsoring a workshop soon on software module transparency, and it seems, like you know, if there's an interest to end them, you know specifying something it may cover. You may cover these and.

A

Similar for the DNS SEC transparency, I did a draft PowerPoint, which I've presented at Athena's option. You can have a look at it there, but I also. Don't think we need to keep the working group open for this specifically also because it will probably not reuse much of the certificate transparency format, so it would be something fairly new anyway. So I think it can also find it's a new way into the into the IDF.

A

Okay, so then, who wants to fill up ten minutes with interesting discussion before I crush yourself?.

I

Ben Schwartz I'm very interested in in the possible interaction between redaction and IP certificates. I wonder if they're not.

A

Considered considered that specific case, so yeah you're right actually I skipped one item. It's the redaction. The redaction has come back and forth to the working group like three or four times now and and again, I think we should. We should just if there's an interest in there we should do you have those people that are driving it to start up off and maybe working group, because they've come here and vanished on a number of occasions. So I think that should just be split off from this work. Amateur from Linda.

D

There's folks that I mean this is a browser vendor you're kind of imposing their view of the world. No offense, but chrome has proposed a mechanism that sort of addresses kind of indirectly redaction through on, through the use of basically enterprise in an enterprise, namespace defined in chrome, it's to permit stuff that won't get logged at all rather than stuff. That has a label edited and I that doesn't really address your juice case in the slightest. So maybe the chrome folks could say something about that. Tim.

C

How big the lack of interest is actually entirely due to lack of interest from the browsers that a lot of people who remain interested in redaction and I think the the fair summary is that none of the current methods work particularly well so, but there also is until we get there.

C

The chrome folks were very early on this and everybody else has been just starting to come up with their own CT policies, and so I would encourage people who think redaction is important to you continue having discussions with the root programs about their root policies. If we can get a root program interested in discussing a concrete proposal, then we can start discussing it again, but yeah I, don't think this working group has to hang around just for people to come back and do that.

A

We have 8 more minutes.

D

C

D

Say more about your use case? You know the question of redacting IP addresses.

I

Sure so this is this, isn't exactly my area, but it seems to me that the IP certificates have very different security properties from the perspective of a client who's establishing a connection. In particular, there are fewer intermediaries.

I

In some sense, the client validates the location of the validates, the name by its route, ability it, depending on your perspective, so I think I I'm a little concerned personally about essentially creating a system where any IP end up with an enumeration of all of the IP addresses that have ever requested an IP certificate that that seems unfortunate to me.

I

I guess, I, you know, I I agree that this doesn't need to hold a river, but I mostly wanted to come and see if anybody else is interested in talking about these things, maybe outside the group, I also would like to see the binary transparency work received in it, and my my group is very interested in binary transparency in something like a JavaScript context, with maybe a browser extension. So again. If anybody wants to talk about that outside the group, please let me.

C

So on the questions of IPS IP certificates are weird and actually quite rare. A nomination of the IPS that have IP certificates is a little bit less dangerous because they have to be public IPS anyway, and the information that you're leaking is that a certain ipv4 address has have a HTTPS certificate at one particular point in time. That's perhaps not the most sensitive piece of information in the universe.

C

The IP certificates have a lot of interesting problems that are being actively discussed. Just because of the fact that, as you mentioned, the security properties are very different and the lifetimes of static IPs are often very different and much less stable than DNS names and the, and in particular, they get farmed out to various entities on a regular basis, sometimes very quickly.

C

So there's a there's, a reasonable position that perhaps they shouldn't exist and the validation rules for them is kind of being actively discussed as something we will either clean up, or we will just eliminate this I think we might be one of two companies in the entire world that actually issued IP certificates on a regular basis.

A

But just to be clear, there's there's no additional requirements on my 69-62 bits for any of this right.

C

Yeah, as far as I'm aware, they look just exactly the same for the verses of Trance yeah.

D

C

Well, that's an interesting question. Actually, in many cases, actually IP certificates are more commonly used in non browser use cases, even if they're issued off for publicly dressed recruits, just because of the fact that they're used in server to server in IOT use cases where the trust store generally is the same as the trust or that was stolen from some root program. So they have to be publicly trusted. But there probably are people using I teach their IP certificates with browsers I. Think captive portals would be the most likely use case.

E

This is a Devon from Google Chrome, again trying to wrap my head around the point that you brought up about the numerating IP. You said you knew bringing IP is concerning I'm, not sure I understand why it's concerning is that is it possible if we still have I think we thought sometime.

I

Sure so I don't want to take up lots of working group time if there's if this is outside the group, but this is Ben Schwartz. The my use cases generally are about services operated by by clients that they want to be able to use securely from a browser, but the whose existence they want to be confidential, because otherwise those services could be subject to various kinds of attacks such as, for example, a DDoS attack.

I

So if the, if the certificates are logged publicly, then it makes it potentially much easier for an adversary to discover the existence of the service and mount an attack against it. In ipv4 I agree that you can essentially enumerate the ipv4 space you can zmapp for these. The existence of these certificates, but n ipv6, that's at least not so clear that that's possible. So personally, for my use cases, it would be nice if, if browsers, would take ipv6 ipv6 IP certificates and not require them to be fully enumerated in the in the certificate transparency box.

C

Well, and that that's complicated too right, because one of the goals of certificate transparency is, we wanted to know who is issuing what so we pay attention to naughty practices and the dark little corners of validation is exactly where the naughty practices tend to be so I personally find a lot of value in logging.

C

These I think you're right that there's ipv6 is a spot where you could hi, if you wanted to, but I'm, not sure how what the, how that sort of weigh the benefits and I think I think the world is moving to a place where you know you eat a lot of the redactions stuff. People thought they were doing a better job of hiding than they then they were, and they had semantics, for example, had an option of please don't log my ICT certificate, and it was rather popular.

C

Most of them ended up getting logged by Google anyway, so whether they were actually getting, what they wanted is a little bit questionable, so I think I would sue. I. Think I would suggest to people that if you have a public ipv6 address and your security is based on the fact that nobody knows your public ipv6 address, you should perhaps revisit your threat assumptions.

E

Yeah I think one of the things that we saw, a sort of anecdotally was looking through logs and we- and this is in the context of finding certificates with likes- maybe malformed se T's in our a so we're doing a big analysis trying to help CAS get up to speed with that and one of the most common responses we got was like sheer surprise that these will publicly know like these things were that they were like these are behind lockdown intranet they're like no.

E

This was crawled and put in here, and it's not just by Google crawler right, there's tons of white hat security researchers to do these things at scale, and yes, currently, mathematically ipv6 is so much harder right. Just as an aside for that, but echoing Tim I think any model that relies on those things not being discoverable for long. It may not be like a mall. You want to build a lot of foundation on sort of not addressing your concern but sort of the rationale for your concern.

A

Thanks for coming so we briefly had the your items from the adv fee, which you can see the bottom just like, and so we're just wondering about your your updated views on those items so.

A

There seems to be no no potential by changing the document status to experimental so that that one's resolved, the remove tracking, preventing tracking clients claim there's still some disagreement but where it should be removed completely or revised. So what is your a few unsure I mean that's a working.

J

Of decision so on the my position, I think, is that, on like they good I strike, I was trying to get out is like that. What the maybe you could pull a PR from my understanding is when I read it was that when you say well, new terminal signatures like that that there's a sit, the thing that says don't do to, or ministers integers and yet simultaneously the ITF is pushing people use RSA towards PSS, and so there seem to be in conflict and so um and I understood and so I thought.

J

Okay, no one was trafficked to be um dad. There was some question about how to manage that in particular on like ever, who was where we said that you could.

J

You could deal with this by like storing, might exploring the city or somewhere and then then the race that wasn't practical and like um I guess like I, think exists, easy walking decision you how to proceed, and you could proceed like we've deliberately decided that we think PSS is stupid and we don't do PSS and because he has one one five or it could be we're gonna like remove this text.

J

It could be we're gonna, we're gonna, like you, we're gonna, say what you have to suppose integers, but it's like, like I, just think these things are just some things intentionally. You guys need if we have to resolve. Oh so I'm like not pressing for any particular resolution, um but I do think that something's got to give in terms that triangle um so I, don't know if that helps. um I didn't really understand, like oh ye. My way to the no list was a frankly pretty inconclusive, but any any of our solution.

J

The chose are willing to stand behind I'm willing to stamp on okay.

A

And then, the next time the most have removed no security implications. Can you.

J

Pull it up, I read it myself their day, but I have to admit I'm a little rusty on.

I

J

This is about huh maybe that.

D

J

Is finally me on the road gestures, Ennis I'm, fine with that, but I guess I'm reading. So, okay, hang in a second moved: yeah, okay, I.

J

Try what was my complaint about this and I remember now, is that we able to find that.

A

Not quickly I think you actually wanted to remove it and there was some pushback on not removing it, but I think.

J

It does seem like what most likely happened here, another wise. What am I complaining about.

A

Well, I think he objected to removing and I think echo suggested, removing it and then now.

J

Confuse cuz: this is issue you should remove it.

J

It in order or did or did you put that in respond response? Did you put that in response to my comments and now he objected and I guess to have to pick it I got again.

F

Alright sigh so Andrew here so um I think that there was some confusion about why this. This was why this was in here in this from its Roque I'm, trying to think now uh right. So it all goes back to why a log has to validate the chain to a trust, anchor and I think that originally there was a misunderstanding, which was that the only reason the log had to do this was for anti-spam.

F

You know to prevent it from being spammed with certificates from you know, someone's you know own trust anchor, so that I think then led to the suggestion that we say there's. No, that I think that led to your suggestion that we should say that there is no security implications to requiring the check to the trust anchor, because it's just anti-spam but as I pointed out, it's not just anti-spam. It's ensuring that we can attribute every certificate in the log to some certificate authorities, so that browsers can take action against the certificate authority.

F

So because of that- and we do not want to put in that- there is no security implication. There is a security implication. It's just not anti-spam something else so.

J

That make sense, paging back into memory and I agree with you, I'm happier with this text. I think it might be productive for somebody to add what I what you just said somehow, because, apparently at the fact that apparently several were confused about this suggests that maybe a bigger- and it is a subtle point about CSS, but have you documented but I'm? Fine with a planet is PR yeah.

F

Well, actually, the I suggested some text in my email and then Rob actually included that in his PR, so the PR will add in the explanatory text for what is required. Okay, what's that, if.

J

You're, our new sex yeah.

F

If you go back to the PR yeah, if you go over to the drink.

C

F

Files changed right right right. There we go.

J

Looks good to me yep this one's fine.

F

Awesome thanks.

J

A

D

J

Minds are all stiff need that from you, while Teresa and then I can ad review and make them shipped off. Yes, I mean I think make sure that they got yeah double check and then we can go right. Thank you. So.

A

So, since you may be elisa, knows a bit more about what happens to adopted working group work when the working group closes.

G

Elisa Cooper it's so it's kind of up to you. It wasn't really clear to me when you were going through if it's just that you that the working group doesn't want to work on these anymore and there's individuals who would like to have things ad sponsored. It didn't really seem like that, but that's one possibility, but the other possibility is the drafts expire and that's it.

A

It seems to mostly that the works been really slow and if people haven't haven't had these cycles to work on them,.

D

We've got a gossip draft that there's a there's continuing interest in it, but there hasn't been a lot of progress. It was intended for publication as an experimental standard. Is it can that be ad, sponsored sure.

J

It could be I think, honestly on the one holding that up, so maybe it will be better would be. If we try to resolve the remaining issues and then just and and and and then um you know, III guess: I grown I like what were you hoping to Colonus, actually ask you that I'm sorry were you hoping to close? Let me ask you that, like now, um I am.

D

Yeah I mean, unless you know, we've been discussing this with several meetings. Now, there's been no serious objection and by the way, my can you aid the sponsor. This was a speculative question, rather.

J

Than a request just leave it so technically, yes, I could but um I guess his ended, which these are all working with documents it'd be better to and unless I may feel differently and like I, this I think guides. The charity fight we're here, but um it might be better to just like keys, are working rely for the month or two of law. Hard lanolin! Is that what do you think.

G

Yeah well, I would only keep the wooden group open until the Biss document is through the IHG, because you never know there could be something that comes back where you need a working up. Consensus call. So, if you're going to do that anyway and there's energy to just process the other document relatively quickly, then just leave it open until those two are done. I think.

J

I can get through that pretty quickly when we another okay.

D

Any other business.

E

Hi again, sorry, this is more of a procedural question due to lack of familiarity, so bear with me. um So, as I mentioned, we plan on doing some experiments to get like that at the back like we mean, because basically, there are like material improvements and in bits that aren't in sequencing right, and there are things that I think are going to be very useful in the long run, but maybe not be in their final state if the WG closes and we come up with these results half a year later.

E

What sort of procedurally happens then? And again it's just because I don't know.

D

We could potentially spin up a new working group or because this would be what we need basically dropping about, as fans assume, um ultimately transitioning, something from experimental to standard. um Potentially them I. Don't know, I may be another ad sponsorship, but we could spin up another working group pretty quickly on this I. Think, okay,.

E

Yeah and just the reason why I asked this concern is just that, like these things, move as glacially as changes in the PK ecosystem do, and that seems to be a pretty big mismatch to the speed at which, like IETF WGS, go it just hello Maisie. It seems that extenders laughing right next to me so yeah.

D

Yeah IETF working groups don't sort of work in an area and cover top. You know covered topics. We have specific pieces of work to complete so.

J

Precision, music, the concept of process which movie star, which means are slowly that IETF process because there to be a pacing in a rapid pace, uh-huh I think I. Think like um it's craft or aggression. When we do, you think you have possible white feedback from this process.

J

E

Hoping to have initial feedback by 104 right.

J

Yeah so I think I guess what I would say is there are really two possible like we have the procedure. The coupon they're two possible like outcomes here right. One is that, like we think expenses, this is essentially correct and in which case like we could simply read, we could simply you know rebadge. It is standard as proposed standard, or you know, issue like a Navy sponsored small things as well. Some small Corrections. The second is, we think, is not substantially correct and like more what needs to be done, in which case we would.

J

We would have that one if our working group, in order to do the work given the time frame, you're telling me I, think I'd prefer to close this working group and then spit up a new one rather than because we're gonna ice for like a year or two so I think what we do is we just basically spin up.

J

You know eight commands to and maybe even call it trans elastic and tell me I, don't know if there's like these things get like you know, trans OH or something, um but um you know to do that, but that'd be the process but think yeah again it really depends on what the feedback is. The feedback is it's basically right then, like just promoting is easy. The feedback thing you don't want to work then, like you have to do that. Work. Someone.

C

Yeah, this is Tim, I. Think I agree with that I think. Actually the document is I, don't know about. You, guys are what you guys are thinking about proposing. We probably will actually have proposed improvements before you do.

C

If you're not gonna do anything until 104, but most of the things that we're thinking are basically augmenting trends with new ideas and different features, and things like that and I think those are better handled as new drafts as opposed to a business, and so those can just go through the normal IETF process and hey we've got a new draft. It's got great new trans stuff, there's no, where it can currently go. Please tell us where you want to send it.

C

No so I think the existing process will work fine, even if this close group closes.

G

There, a bunch of other groups where you close the group, but you keep the list, and this seems like a good candidate. If you follow this process.

D

Okay, has everybody sign the blue sheets.

D

And other than that, I guess we're done see you on the mailing list.

D

Ending is even short when I'll go through the recording.

D

Okay, by the way.

H