Everything EOS EOS Node Operator Round Table, 3 May 2023

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Node Operator Roundtable - May 3rd, 2023

Description

Join us for an engaging discussion where participants share their thoughts on Antelope documentation, release notes, API references, and usability. We explore the challenges faced when extracting data from snapshot files and discuss potential improvements. Listen in as we discuss alternative solutions, their limitations, and future plans for snapshot utilities. Don't miss the closing remarks and learn about the next steps to enhance the Antelope ecosystem. Stay tuned for future meetings and more insightful conversations.

00:00 - Introduction
01:34 - Documentation & Release Notes
05:53 - API Reference
10:24 - Usability Study
12:57 - Extracting Data from Snapshot Files
17:39 - Snapshot File Use Case
22:03 - Alternative Solutions and Limitations
27:06 - Future Plans for Snapshot Utilities
31:50 - Summarizing Feedback and Next Steps
36:12 - Additional Suggestions and Discussion
41:28 - Closing Remarks and Future Meetings

The EOS Network Foundation coordinates financial and non-financial support to encourage the growth and development of the EOS Network. We’re harnessing the power of decentralization to chart a coordinated future for the EOS Network as a force for positive global change.

// EOS Network Foundation //
https://www.eosnetwork.com
https://twitter.com/EosNFoundation
https://www.linkedin.com/company/eos-network-foundation
https://t.me/EOSNetworkFoundation
https://discord.gg/eos-network
info@eosnetwork.com
#EOS #Web3

A

Welcome everybody., it's may 3rd, bs another note operator, roundtable.

A

Um, and as always,, we will start with a update on the software development side of things, and I think we got some updates on c d t from steven.

B

Yeah, that's right. um,, so we had been mentioning the last several weeks that we are anticipating a release of the first rc t4, zero zero.

B

Um, all signs are pointing towards likely getting that out the door today.

B

Um, and one thing that I'll mention kind of in compliment to that is that, uh,. I've mentioned this in the past as well,, but we have a.

B

Developer's round table,, which is a day after this every other week.

B

Um, we had originally intended to do a demo of kind of like the featured tool set,, which is antler prj in the previous meeting, uh,. Just due to some.

B

Logistics and, uh, personal things that came up for some of our teammates.

B

We're going to seek to have that happen, tomorrow.

B

so it'll be hopefully fresh on the heels of that release, where we'll be able to walk through, um, an example project.

B

Simple things like getting set: up, uh, dependency, management, stuff, like that.

B

Um,, so we anticipate that coming around.

B

again,. If anyone needs an invite,, you know who to reach out to.

B

that'll be me. um, outside of that, uh,. We also anticipate, uh,, probably a patch release this week for dune.

B

Um, and then probably a more material formal release, either the week after this one or the week to follow that.

B

So, um,, that's kind of where we're tracking at, uh, from the developer tool. Side of things.

A

Thank you, steven. any questions for steven?.

A

All right, well, that we can move on to, um, the round table portion.

A

I'll pass it over to brian.

A

all right, thanks daniel.

C

So, uh, before we get started, I've,, I've got two potential topics that we could discuss: today, but,, and, um,. Those are continuing our n types, discussion.

C

Um, we,, we got through the first couple: node types, and,, um,, sort of documented some.

C

Configuration best practices from folks on the call in previous calls.

D

Um,, we have a number of other note types that we could,. We could make progress through.

C

Um,, another potential topic is reviewing the, um,, the initiatives that we are targeting as potential items for the leap, 5.0 release.

C

Um, we wouldn't. go into depth on those items,, but more just kind of a real, quick high level statement. On, uh, on what it is.

C

Um, and, and basically the idea there would be to get, um, any feedback as to, you know, yep,. This is important to me, or it's not important to.

C

Me,, but I see why it's important, or why the heck are you doing this?.

C

Or,, you know, uh, any additional thoughts on like, um,. What you would consider, uh, su success criteria for those items.

C

So those are two that I have in mind,, but I also wanted to briefly open a sort of open floor to, to gather any other potential topics.

C

Um,, so if anybody, I'll, I'll just give about 60 seconds of silence, here.

C

Probably less than that. 20 seconds of silence, uh, here to see if anybody pipes up with any, uh, any items they'd like to discuss instead.

C

All right. it's looking like, no., so in that case, um,, let's do a quick.

C

Vote for folks., if you have the ability to do a,, uh, reaction, just do like a thumbs. Up.

C

So, uh,, if you would prefer to continue the note types discussion,, please do a thumbs up. Reaction.

C

Alright,, if you would prefer to do, uh, review of the potential 5.0 initiatives.

C

Most people are unop, opinionated., luckily,, uh,, matthew, darwin and daniel mckee have an opinion here.

C

Okay., so we will, um, we'll go ahead, and, and walk through those, um, at a high level.

C

Oops,, let me make sure.

A

E

C

D

C

So, um, you know,, this is a,. This is a mix of items that are being, uh,, developed, internally.

C

Items that are being developed, through, um, you know, rfp, grant initiatives, items that are sort of a combination of the two, you know, like, uh,.

C

Contributions being made, um, by the nf on grant initiatives, et cetera.

C

Uh,, but this is,- this is also not a list of things that are definitely gonna go in,. Nor is it intended to be an exhaustive list of everything that could go. In.

C

You know, there will likely be some, uh, some small items, that, that pop up and, um, and get added into this that,. You know, issues, buck fixes, um,, you know,, small,, very small,, um, items.,.

C

These are intended to be the, the really big, coarse, grained things, that,, uh,, that we are targeting as potential items for 5.0.

C

So the first one here, is, uh,, instant finality.

C

Um, this one,, if I can, it, might want to get a better gallery view so that I can see if somebody hops onto the screen.

C

There we go. okay.

C

so,, you know, again,, I'm going 10,000 foot view here.

C

Uh, instant finality is, uh, obvious,. I think it's obvious by the name,, but it's changes to, um, or protocol in order to allow for mere instant finality.

C

Um,, so in instead of the,, the current finality times, um,. I can't remember what the target here is, but it's,. It's like second or, or a couple seconds at most, uh, to reach finality.

C

This obviously has, um,, you know, far-reaching implications as far as like what, um,, the benefits that it can have mostly for, uh, developer applications that,. That would want to,.

C

You know, build on this because much quicker, um, uh, you know, ability to count on a, a.

C

Piece of transaction is actually being included in the blockchain and affecting state and the way that you expect it is or expect it to so much snappier experiences, et cetera.

C

Um, it does also, and we're not gonna go into a whole lot of details here, cuz it,, you know,. We would end up only talking about one item.

C

It does have some implications too, about like, um,, the roles of, um, network participants.

C

Uh, you know, it's. there's,, uh, concepts, of, um,, obviously still block production.

C

Then there's a, a concept: of, um, transaction validation as like a, um,, a separated concern.

C

Um, but, uh, yeah, so,, so there are some sort of like protocol level implications to this.

C

Um, but I guess, so my question here: is, first,. Are there any comments on this?.

C

And then, you know, second,, you know, what's,. What's the kind of level of interest in potentially diving deeper into this in, in a future topic.

C

With folks who have been,, you know, closely involved in this initiative?.

F

Well,, it's certainly interesting.

G

Um, you mean from like a technical,. I mean, I'll, I'll, be interested to understand, uh,, more of, uh,, the use case.

G

I mean,, I know, there's a use: case, right?, but, um,, how this would, uh, affect people that are building.

G

And, you know, other, other, uh,, you know, things that have been held back. With, uh,, you know,. What's three minutes, whatever,, just right.

G

Yeah. yeah., it's definitely a great topic.

C

Yeah., what's interesting about that one is I think, I think, uh, so,, so at,, obviously at a high level,. The use case is relatively obvious in.

C

That it's really just like a speeds and feeds kind of thing, right?

C

like it makes it, um,. It makes it effectively faster if, for transactions, where you, um, must be a hundred percent, confident that it has been included in the blockchain.

C

And you can't, don't want to do like, probabilistic type stuff,, you know?.

C

Um,, but I also think the, the,, you know,, you can get a little speculative about, you know, at a certain point, um,. You could potentially just pretend that the moment it's.

C

Submitted a transaction that it's definitely gonna work or something like that,, you know?.

C

Um, I don't, anyway, um,. I think that's stuff that we'd need to see in action.

C

How,, how developers actually change the way they build.

C

But how about level of interest of going deep into this, um, because,? You know, disclosure, this is,! This is a, um,, an r f, p project, and there has been obviously some, some internal contributions.

C

Uh,, some of our folks are working very closely in this., I'm not one of those people. so, uh,, you know, the technical details, I, I,. I could not really go deep. On.

C

Um,, but there are obviously folks, both inside and outside the e n f that could potentially, uh, guest, star here.

C

Okay., seeing one.

F

Yeah, I'm, I'm interested from a technical level as well, how it's actually accomplished.

G

It's more like an academic thing on my side.. It's not gonna affect my mode operator, um, day to day,, but I, I, they're, just me.

G

C

Yeah., it's just pure curiosity, really.

C

C

I, uh,. I relate to the, the academic interest cuz because I haven't myself been directly, uh, engaged in this one.

C

I, um, yeah,. I agree with you.

C

I have a academic interest in it. Too.

C

Okay. ram limitations. um, next, one here.

C

And I, you know,, I don't even know that I need to give a overview on this one.

C

I think,, I think, folks, know what I mean when I say ram limitations.

C

Uh,, if you don't,, let me know, and I can give an overview.

C

So, any thoughts on this one level of importance?.

C

Any questions?.

F

Um, well with luck more to know what the plan is for that.

C

Yeah. yeah. so I,, I'm glad you asked that. um, my best answer right: now, it's,! I don't know.

C

Um,, so we are,, we are,. We are working molan like part,. These are the initiatives we want to tackle.

C

Um,, the definition of exactly what we're going to do to address it.

C

Um,, it's still.

C

Open, right. um, being actively with, so,. So when you get to the point where you've got some proposal of what you plan to do,, then please share.

D

C

Yeah., so this sounds like one where you, you would, oops, note. Operators potentially would be interested in a proposal review on this one.

F

Well, this one really impacts node operators, so,, especially no operators on certain blockchains, which are very big or have a large number of accounts.

F

C

Alright. inter contract, communication.

C

um, another potential title for this one was composable contracts.

C

This one's more so targeting developers than it is targeting node operators.

C

Or obviously, end users, don't necessarily care about these things.

C

But, um,, this is just a, um,. A well-defined sort of first class protocol feature for allowing, uh, contracts to interact in a way. That, um, allows them to take advantage of.

C

The functionality of other contracts without introducing scaling concerns or security concerns or,, you know, those sorts of things that could potentially pretty bend with this.

C

So, like the goals are stay scalable and performance, and also allow contracts to communicate in ways that allow for,, um, uh,, neat things to happen with, with regards to smart contract, developments.

B

I guess to quickly dogpile on that, um,. There are, of course, other smart contract development patterns out in the wild that enable this and.

B

Are much more, uh, flexible so that you don't have to implement these.

B

Super monolithic, contracts., um,, ethereum, based contract development,, of course is,, is a prime example of that,, wherein we have a lot more limitations for developers on our side. Presently.

C

Yeah., thanks for jumping in,, stuart., all right, early start in broadcast for full blocks.

C

We've talked about this one before,, just to kind of give a, a quick recap.

C

Again, it is, you know, the ability, to, um,, if you,. If you have already reached the, like cpu budget, if you will, for a block, uh, don't,, don't wait around to broadcast it.

C

Do that immediately, which, um, we're, we're already doing that,, but the part we're not doing is getting started on the next block.

C

So, rather than waiting and waiting to start pulling,, uh, transactions off of the sort of pending queue, uh,, just immediately working on, um, pulling that in,. I say.

C

Immediately there, there's a little bit of nuance to the, the timing of things.

C

But, um,, the point is you,. You don't wait. For.

C

Uh,, the actual block slot,, you do it, you do it earlier.

C

Um,- and so this way it's uh, it,, it's another feature to help minimize the odds, of, um,, to help minimizing the odds of having a last block. That is not.

C

Full, while you still have pending transactions that you could have processed.

C

Does that make sense?

C

so, okay. any, yeah. answer questions on this?.

D

Well, good, chris,.

F

I, I,, I will just say on that one, that that feature will make it harder to diagnose where there's problems between block producer latency, because.

F

You don't know what, when that block should have actually arrived.

F

That's my only interesting comment with that.

F

I'm not saying it's really a problem,, but currently we have lots of charts that show,, you know, how fast the block got to the between different.

F

Block producers cuz, you know when it was supposed to be scheduled.

F

Now you don't know,, so you don't know how long it took to get there.

F

So maybe,, I don't know, there,. There needs to be something in the block propagation or the block header that says, well,. I actually.

F

Produced it that this time,, even though it was scheduled at this time.

F

And then you can compare that latency.

F

All right., otherwise it's all good.

C

Yes., potentially additional block header information would be helpful or something.

F

Right., it just removes one of the diagnostic tools that we have to troubleshoot peer-to-peer connection issued between blockers.

C

Yeah, that makes total sense.. I think, that's,, that's an interesting pattern, cuz in some ways this is intended to help.

C

Mitigate those issues,, but I do I see what you're saying.

F

Yeah, exactly. so, uh, may,. Maybe we don't have to talk about that anymore, because it'll all be great.

F

However, my, uh, network operations background tells me that that all be the case cuz there will only be some weird conditions that go on that you need to do. Troubleshooting.

F

And anyway, this just makes that harder., so we'll have to come up with some other way. Of.

F

Figuring out how well people are doing or whatever.

C

Yeah., it's like distributed, uh,, networked programming, 1 0, 1.

C

Don't assume it'll work out the way you thought. no.

F

Yeah. if,, if anything, it might make it worse, because now the blocks are gonna, be bigger,. They're gonna be unpredictable.

F

Maybe they don't even come in the right order. Anymore.

F

you know, it, it,. It could get a bit wonky., so anyway,, we'll see.

C

Yeah. um, kevin, I'll follow up, follow up with you on this just to, to see. If there's any,, um, well, I,, I saw the wheels going.

C

Yeah. I just brought this up.

F

C

But yeah, thanks for, uh,, giving some input there.

H

Yeah,, it's interesting.. I mean, block header's difficult because left consensus change. And, you know, you'd be reluctant to, to do that. Just for.

H

Um, yeah,, it would be like that repeat for that feature.

H

Now you could wrap the block in a envelope on a p dp level.

H

Uh, that'd be a new p d, p protocol,, but that's a much less, um, uh, scary thing than our,. You know, proto,, uh, consensus, change.

H

So, potentially something there,, although it might be,, we already are doing.

H

Heartbeats that measure latency.

H

um,, I mean that's,, so we could just pay closer attention to that, uh, feature: that's already,! That's already there.

H

Um, that doesn't really tell you about individual blocks,, but it would, uh, inform me about latency between nodes.

F

Oh, yeah., but it's only individual notes: that,, you know,. If it took three hops to get a block to get to you, that does not help.

H

Right. yeah. okay., that's a good point., all right. yeah,, so maybe an envelope for the block or something.

H

okay., all right.

C

Um,, I've somehow managed to.

C

Use that. there we go. okay, cool.. So next up is another,, um,, r, f, p item pi, node discovery, improvements.

C

Um, and again, this is, uh, um, an example of, of something that's being worked as sort of a,, a partnership, between, um,, external folks and e f folks.

C

G

Is there a question there?.

C

Uh, says no. so sorry, my,, my brain skipped tracks there.

C

Um, yeah, so,, so pure node discovery, improvements.

C

I think this has come up. in fact,. I think ross,. I think you, in the sort of genie lamp discussion, kind of had some.

C

Noon discovery, requests., right., so,, yeah., yeah., yeah.,.

F

So this is a good one that applies to all of us.

C

Okay. any, uh, thoughts or context or questions on this one?.

F

Yeah,, I have a oh, when you get to actually implementing it,, but I have a bunch of questions about how, uh,, you know,. This is going to prevent people from.

F

Or how do they appear to like private ip addresses?.

F

How does, uh, things get done with I p v, six blocks?.

F

So like, there's protection today, like if you only allow one peer-to-peer connection per ip address, or two. Instead, a limit.

F

Which it prevents you to think I'm getting overloaded., but if you've got an ipb six block,. Well now you have a whole slash, 64 worth of network addresses.

F

So maybe it needs to be,, have a subnet mask.

F

F

things like that. uh,, you know, cuz! If you look in the ethereum community or whatever,, there's lots of discussion about peer-to-peer connections, going to certain.

F

Band ip addresses,, which then gets you blocked on hener or whatever.

F

So, uh, there needs to be some thought into y, you know, mitigations, on which ip addresses are allowed and how that's controlled and things like that.

C

Okay. yeah, that makes sense. I think. uh,, so I've taken, uh, action item here, to, to follow up with,, um, steven and jeff. Since this is, uh, an rfp item on the best way to feed.

F

So I think the ethereum community has learned a lot in this area,. So, let's make sure we're not repeating the same mistake, like that's fine, kind.,.

A

C

We've talked about this one before, generalized ip port configuration for nods,, um,, points.

C

Um,, this one's kind of dual,, um,, dual origin, right?.

C

So the request has come up several times in sort of the history of the software, for the ability to have more configuration options and more.

C

Um, uh,: what's the word?.

C

Um,, more separated, uh, configuration options,, uh, flexible, um,, you know, ability to isolate these things.

C

We went through the proposal on this on a, on a recent call,, so I won't go into too much detail. Here.

C

The other kind of origin of this is that we built a single purpose: implementation of this for 4.0 for the, um,, prometheus plugin.

C

Um,- and this is also sort of a,- a technical, debt, uh, item for us to just, uh,, not have a, uh, sort of redundant implementation, of,, of,, um,.

C

Of what's,, what am I trying to say of the, of the server, um, solution.

C

Um,, which is,, which is how that. single purpose solution works.. So this is both addressing kind of a, a code based thing and also providing,. You know, requested functionality.

C

All right. re reproducible pinned builds.

C

actually,, I'm wondering if I could put, uh, steven diesel on the spot. To give a brief one on this.

E

I, it's, it's internal,, so I don't know that he needs to give a whole lot of detail. Here.

C

Uh,, what I you mean by internal is,. We are kind of the customers of this one for the most part,. I think.

C

But yeah, so I mean this,. This is basically just simplifying the ci process to where we're no longer like manually, generating assets that we attach to releases.

B

Um,, so that's the primary focus.

B

um, right now. It requires manual steps and, uh,, as we are kind of maturing as a team and the means by which we deliver products, um,. This.

B

Is something that should be much better? Automated and reproducible.

C

Yep. the intentions of items like this,, which you'll, you'll, see pop up in various releases, are, they're kind of meta features in that they don't really provide.

C

A feature to stakeholders directly,, but what they do is hopefully help us move much more quickly, um, by, by removing,, um,, redundant manual, steps.

C

C

Um., the illumina related to that last one, I'd be interested to hear some feedback,. If you don't mind, around, um, the,, the whole idea of pin builds and, and consistency and states.

H

So, you know,, the reproducible, pin, builds are.

H

You know,, when we do a release,, we,, we, uh,. We use just what we call a pen build,, which is a, a set configuration of compiler boost, version.

H

Um, and I guess os uh,- maybe that's a given.

H

So mainly boost and compiler version., um,, and, and arguments to to, to those so that you can restart a node simply by shutting it down and restarting it.

H

And the state file can be reduced. right., so that's, possible. for example, with the 4.08,. If you were to grant our pin bill, you know, the build that we provide,. We can shut down a 3.2 node and start up a 4.0 node without moving, you know, to start and stop.

H

Um, and so that's somewhat the,, the purpose of, of the pin build is to provide that functionality over time.

H

Right?, so I recently asserted, um, when talking internally about this topic,, that the importance of restart via state is not maybe what it used to be at, at what.

H

Point in times, in early days,, we took extra care to make sure that that was.

H

Doable as, as long and as, uh, you know, as, as much as we could,. We could to make that happen.

H

Uh,, some of that in the early days was because that, it also included the history plugin, uh, data,, you know,, because you didn't want to have to recreate all that,, which was awful in that safe file..

H

Of course, that's,, that's no longer the case, um, these days and,, and I believe that these days, at least what I'm asserting.

H

Is that people are, or more likely, and, and maybe more trained, um, to just go ahead and restart with, with, uh, snapshot as opposed to restarting with, with stigma.

H

Well, I'm wondering what are people's opinions on the, um,, you know, the, the.,.

H

All the care that's taken to produce, he's reproducible, restartable nose,, as opposed to just maybe a release. Saying, well,, you know, start with snack shop.

F

I think every,, what do you call it, feature release to start with the snapshot is reasonable.

F

If it's a patch release,, that's a little inconvenient.

F

So it was really nice. like I was expecting to go from 3.2 to 4.0 to need a snapshot, and then I, like the first time. I did, it's like, oh, it just upgrades.

F

Awesome, great., I saved this step. right?, so I think that's a reasonable expectation.

F

But yeah,, if I'm going to from 3.2 0.1 to 3.2 0.2 on,, I use a snapshot.

H

All right. so yeah, so certainly on, on hatch, release.

H

that,. That makes perfect sense.

H

and you also mentioned, you know, from like,, say a three point x to a four point: x, uh, expecting snapshot doesn't seem unreasonable.

H

About,, how about from like a 3.1 to a 3.2,? Would that be a reasonable.

H

Um, use a snapshot upgrade.

F

I don't know., it depends what your guys are doing with three point ones and three point: twos.

F

In the past, there seemed to be a more use of that.

F

Now it seems like you want to go from four to 5.0 to 6.0,, so I don't know what you're doing with that minus the middle digit anymore.

F

So you, you not, kevin, the toys,. It's always a little.

G

It's a,, it's always a little inconvenient.

F

Um, but let's flag it up with what, like the,, where the effort is on your team.

F

If it's gonna save you, uh, uh, some time to focus on something more important,, then hey,, you know,, we'll just deal with a snapshot.

F

I mean,, if you ask us what we want, we never, ever want to have to use a snapshot. Ever.

G

F

Yeah, yeah., but we know at the same point that that's not realistic..

F

There are gonna, be changes over the time that requiring now use a snapshot.

F

G

So I think matthew, I,, I agree with you. right. you know,. When there's a, a mock step up,, then the snapshot's no problem.

G

You know,, if I, I,, I often find, um,, I get caught out with, uh, having it to use a snapshot. If I compile myself or.

G

If I just use the deb, there's, always some trick, mismatch, um,, with whatever it was compiled with originally,.

G

But not happy to use the snapshot. That's required..

G

All right. appreciate the feedback.

C

Excellent. and now I just, you know, quick.

C

sort of statement, on, on,, you know, how we're, look,, how we're looking at prioritization of items, cuz. This is sort of a disclaimer of all. This could change, right?.

C

Um, so, you know, the, uh,. What makes, um, antelope, you know, sort of stand. Apart is really the,. You know, performance,, scalability, and reliability.

C

So, of course, um,, those sorts of features, that, uh,, enhance that and take it to the next step are always, um, you know, priority items.

C

The thing that trumps, those items are, uh, when we have opportunities, features, or opportunities that will help us drive,, um,, developer, adoption and ultimately, uh, end user adoption.

C

So,, where those opportunities present themselves.

C

Um, other things might,, you know,, might come in, and take, uh, precedence over some of these things, uh,, even causing some of these things to fall. Out.

C

Um, but yeah,, so that's kind of the high level framework is like,. You know, end user, adoption, develop,, then developer adoption, and then, uh, performance,, scalability,, reliability, features.

C

All right. um, yeah,, so this was,, you know,, obviously just a very high level overview.

C

I appreciate the, the feedback and commentary on, on those items.

C

If you have, uh, a additional,, you know, questions or, or anything that you want to go into,, I'm always available on telegram for just discussions, and, and also always welcoming of, um,. You know,.

C

Of, of scheduling, a, a call to do a one-on-one. If you want to really dive deep into something.

C

Uh, on telegram, I'm,, I'm at b hazard,, so pretty, pretty easy to find.

C

Okay., so I guess then, uh,, just staying on the, yeah.

C

Two zs. thanks steven at b, h, a z, z, a already.

C

Um, staying on.

C

5.0 a little bit longer.

C

you know, we,, we also,, I think our last call we did sort of the, or maybe it was two calls ago,. We did the quote: unquote: genie, uh, genie, wishes,, uh,, conversation.,.

C

Um, I'm curious about,, you know, people's,, you know,. We've talked about individual items, people's overall thoughts on the items that we've, uh, reviewed.

C

C

Are there things that you expected to see that you're not seeing?.

C

Are there things that you're, seeing here that you're scratching your head?.

C

Like why,, why is that the focus?

C

um, you know, any of that kind of feedback?.

G

I think it's what I,, what we originally discussed..

G

Obviously some of those lines, were, um, broad strokes- and we had like an hour sometimes talking about some of those things.

F

So, uh,, you know, when you just say,, uh,, optimizations or improvements, sorry, that obviously there's a whole lot of other layers,, but yeah,. That's what we spoke. About.

G

A

All good., this is probably less node operator relevant than an is developer relevant,. But while it's on my mind and on the topic, of.

A

Like antelope superpowers that we could enhance.

A

What sets antelope apart? um,, something that I think we're underutilizing that maybe there's some low hanging fruit.

A

I don't know exactly what it would be,, but the flexible account permissions of antelope is something that I haven't seen on any other blockchain, and they're very powerful. If you know how to use them.

A

The problem is, most people, don't know how to use them.

A

and don't,. You know, the potential for out,, you know, op opportunities to use them. Effectively.

A

Um, so dunno, if that's a documentation issue or an sdk.

A

You know, it's,, it's probably not a protocol level thing,, but tools, that'll, help people to configure accounts with more complex permission, sets, multisig,, all that kind of.

A

Fun stuff that right now you need to be an expert to be able to do that. All manually.

A

Um,, it would be great if there would, you know,, especially with gafi being with focus and things like that,. I think there's maybe some opportunities there.

G

No, aaron had, uh, some requests for some features,, like he wanted timem delay on missions.

C

Yeah, permissions.

F

C

Yep. yeah, permissions that timeout was, was his, uh, request.

C

I actually did discuss that, uh, internally and get some feedback.

C

And I think, uh, people, um, their,, some of the core sort of architects, uh, thoughts are already in that area.

C

Um, their,, their approach is a little different,, but to satisfy the same goal, like around being able to, um, being able to support.

C

Scoped, temporary, um,. What's the word, I'm looking for?.

C

Um, sort of granting of authority right.

C

To, to another, uh,, entity., so, wow.

G

Where's the, the documentation is that docs do?.

F

Yes. network.com. is that the, is that the,, the current documentation?.

G

I, I can see it's got the a p r, inferences and stuff, and, yeah.

F

Yeah, I, I, I'd love to know what the doc strategy is there.

F

Yeah,, that's an important one.. I mean,. We just called it no, I mean, danielle says,. Are we gonna,? Are we gonna teach everybody to use the permission? Structure?.

F

And I'm looking here,, I can't find it anyway,, but it used to be in the old docs, like the side, stuff.

F

Yep., I think that's important.

G

That really important.

F

There's a lot of stuff missing there.

C

Yeah. get, give me more there. like, uh, what,. What do you think are the most important things that, um,? We need that we don't have in that, in that area?.

C

Well, from a, from a user perspective, just pure user,, you know,, maybe we need some details on how to use, um, how to use, uh, kiosk.

G

And the permission, structures and that kind of stuff.

F

I just, I haven't been here in a while.. I usually come here to look for api stuff.

D

F

Um, and I'm just,, it doesn't seem to have,. It's called e, f c dm stuff.

F

D

E

Yeah. that, that is, they're missing like, uh, just jump in quick,, so you like, uh,. Last time I was looking for, just checking some parameters. For, you know, compute transaction e t I.

E

And it just doesn't exist like on the docks, like. As far as I can tell you, you know,, you go to like, okay, version 3.2 or latest,, but it actually.

E

Link links you to like some really old, uh, docks for a very old version.

E

It doesn't include any of these, um, new items.

E

And I think the only place is if you search for a compute in the docs, you'll, get linked to kind of a page that talks about dune..

E

And then link to the actual code of compute.

E

so, and also every time you search and click,, it opens their new window.

E

It's really not, honestly, this,. This whole doc thing needs to be thought: out,, uh,, properly.,.

E

Cuz currently it's,, it's like,, it's much easier to just use. Node er, slash help at this point.

E

Uh,, which most developers,, especially new developers, are now gonna do.

D

E

So how about that?.

G

Are you looking for wishlist for book five?

G

oh, before we even get there? needed the documentation to be like comprehensive,, um,, operate,, operate, operation of the protocol and all the programs that running in it.

D

D

Just wanna mention here that it would be really helpful for dev rail if we had, um,, basically a list of things that you might be interested in, um, or that would be helpful for you..

C

Yeah,, so that's an open call to make that a like really specific action, is, um,, if you, the next time you go searching for documentation, and, uh,. Definitely if you don't find.

C

What you're, looking for, or even if you find it,, but it's just hard for you to find.

C

Um, you can, you can., please let me know about it. uh, again, at the hazard or, um,, you know,, if you,, if you're, uh,, uh, already connected with, um, na uh,. Oh my gosh, nathan.

C

Yes. uh, then, um, please, you know, he,. He would also obviously be very interested in that, as he heads up debra.

C

So, uh,, which is developer relations.

C

if,. If we are to work.

C

So, yeah, next time, you're, you're ursing, us cuz. You can't find something, curse us to our faces.

C

E

I think just simply having like an api reference,, it should work like other, uh,, you know, applications documentations,, where by default you see the latest, you know,.

E

For example, in node, dios chain api,, you always have at the top left what version that is for as a dropdown that you can switch. So you can easily see the changes.

E

All those things are like kind of standard in the industry and here it's, they're, all linking to an old version.

E

Like worse than not having the documentation, is wrong, documentation or incomplete documentation.

E

So I think like just having that, as a start will just for least reference. It's necessity.

F

Yeah. related to that,, I made a comment: earlier, which.

F

Uh, on an earlier, a few weeks ago or a week ago, that, you know,, it'd, be great to see the documentation up to date as the, uh, release candidates come out.

F

But I was told, well,. We don't wanna link those two things cuz, those are separate: processes, but, well,, but if I'm testing rc four or rrc three,, then where am I finding the documentation for that?.

F

So, uh, yeah, think the way I captured your feedback, um, just to make sure it understood. Correctly, is like we,, you know, we're for sort of minor and patch releases.

C

Um, it, you know,, it might make sense. To, um, rely on like the built-in c l I documentation being up to date.

C

Um, but then for major releases, ideally, uh, documentation is launched in sync with that, right?.

F

Yeah. the, the other thing that I've seen,, some other projects, did it,. I don't know if this is the best practice or not, is that they have a ongoing release, notes mm-hmm.

F

Markdown file. so, as new features are added to the release,, the release notes are built incrementally.

F

So as part of the pr, getting a pr approved, well, you have to write the release. Note blog.

F

Right. so therefore, if I'm looking at the 5.0- and I want to know, what's new and how to use it, well,, the release notes are being built as it goes.

C

Yeah. so we, we actually, uh, plan to do exactly that or like internally.

C

Um, we hadn't really talked about specifically like going ahead and making that, uh, public,, but I think that's, that.

C

I don't see why we couldn't do that.

C

um,, I mean,. Obviously it's not. The final polished form, it'll get.

F

That's fine., you know, wordsmith thing when it gets to the end,, but at least you have the list of all the major features and then, when you go to do the.

F

Five, oh release,! Well you don't forget what was done six months, ago.

F

D

Because, hopefully it was added at the time.

D

C

Yeah. and we wanna do that, uh, internally, just as a process improvement because,, you know, the making the release notes at the end, takes, uh, like a whole day.

F

Yeah. I, I,, I already know the challenges. you know,. I used to run an r and d team with, you know, several dozens of people in it and building the release, notes at the release.

F

Uh,, that was challenge. yep.

F

and it was, was more than,. It took more than a day.. I tell you, in our team, also, ideally,. If the docs could have,, especially api reference, or mainly api reference,, could have some type of like swagger.

E

Like, uh, ability, where you can actually test it on the on, on the page, itself.

E

So you can, you know, test like get info or get block or whatever other a p I.

E

Especially for new developers,, they wanna see it's right there in the page. What to expect before,, you know, just trying themselves,, even if they have to input their own api endpoint.

F

Yeah, and I think, there's already a swagger thing somewhere in the background there anyway.

F

Or I have seen such a thing in the past.

F

So that would be cool. the other thing, you know, speaking of that swagger thing,. Maybe it should just ship as part of node d os or something so that the standard,, when you.

F

Go to a block producer's api page, there's already a swagger page, there.

F

Like, uh, I period does that, for example.

F

Every hyperion has a swagger page.

G

F

Um, I was just looking at the old one..

F

It's pretty,, it's a lot more comprehensive, even though it little two one than, or whatever.

G

I mean, it has an accounts and permissions.

F

Uh, well, the new one has accounts and permissions,, but I just found that it's under resources.

F

That was not obvious,, but I was clicking around since we talked about it.

F

I was like, oh,. There's a council question. resources.. What section is it under? oh, there you go. okay. oh, okay.? Well, that's not very, uh,, intuitive., no,, exactly.

F

uh, and it's the same,. It's the same. Pictures.

G

Yeah, yeah, yeah.

F

But it's, it is just there. so anyway. okay. alright., maybe that's off the bottom. funny stuff may,. Maybe it just needs some,, uh,, usability.,.

F

You know, take some users,, give them those challenges, find x and then see how long it takes them to do,. You know, those kind of tasks that you've been doing,. A usability, study, right?.

F

Because the goal is within,, you know,, whatever two clicks, you can find the, the thing you're looking for.

F

Yeah. so,, you know, clearly ross and I were not looking under resources for account management, stuff.,.

F

C

Makes sense. it's, it's, yeah.

E

And redo the whole circuit of the doc,, the es protocol up a layer or something.

C

C

Um, yeah,, so we, we've,, we've kind of talked about documentation.

C

Are there other things like this that, that.

C

um, strike you as obvious that,, you know, need, need to be, uh, need to get focused.

G

Sure. we'll think of things as we go,, but there's a wish list. For.

E

The item ipot before,, which is like to extract data from a snapshot file.

E

I played with, uh, leap, util ability to do that,, that, uh, kevin, uh, pointed me to kind of the issue with that is like, first,. It needs to load that.

E

Snapshot and to know dios,, it's not actually,. You know, uh, dealing with a file.

E

It's running like another chain, uh, in your temp folder.

E

So say you do this for wax, you're, you know,! First of all,. It needs to allocate the skit, seven gigs for the state. File, and, uh, then produce a jason.

E

That is also like gonna end up being like 60 plus.

E

Gigabytes and there's no filtration option there.

E

There's really nothing when you're running deep, your till.

E

um, and, in addition to that, the,, the output format is,. Even though, like you can programmatically read it,, you still need to programmatically almost stream it.

E

You can just read that into memory, uh, for most computers.

E

And the output is just like the structure is for me,, just not really usable.

E

Like, uh,, you know,, like you, have all the contacts one into one.

E

Kind of array,, like the whole structure of it, is not very usable.

E

I was wondering if there is any plan or effort to work on a leap, util, snapshot?.

E

Um, you know, extraction utility that does not need node eos itself ideally,, but if it does a way for you to pass on parameters, like, uh,, an array of contract names or like a csv of contract,.

E

Names, that way, it only extracts the data relevant to that,. That would be like, super useful.

E

Currently, I'm back choosing diffuse for that.

C

Can you give me, um, some clarity on, on the specific use case you're trying to solve with this?.

E

Oh, yeah. like, say I wanna, for example,. Do like air drop, to, uh, at a certain,, you know, block height.

E

So, uh,, currently I don't have any and I need to get like a bunch of different tables from a bunch of different, uh, accounts.

E

Now the only way is use either like the snapshot, file, right, with theil and output. Everything because I need to guarantee it's add that specific block.

E

Height and not, you know, like changed while I'm fetching the the table, roads.

E

It's either I use or I, you know, load from snapshot and then just don't, don't continue. Appearing.

E

Don't sync that node and just keep, you know, uh, like stop at a certain block and just, you know, get table rows, uh, uh, as much as you need.

E

Without syncing, that node,, which is obviously just not a good way to do, it.

E

Or currently, what I'm doing,, which is just diffuse.

E

diffuse, allows you to get the state at any block height and the scopes and everything else.

E

And on top of the lip output is like, you still need to.

E

Do some work on the values,? It's not coming to you and kind of like, you know, de serialized, clean daca.

E

So it might not be a priority thing, but, um, yes,, to get around,. I think the use case makes sense.

C

Um,, what I'm curious about now, so you,, you mentioned two, two potential solutions.

C

What is it about those two solutions that is, uh, undesirable?.

C

So you., so you can load everything into a node with no pierce or you can use defuse.

E

Defuse is great. so, uh,, you know,, like that's what I do. Now, like just wrote the small, whatever file, uh,, no gs that will just get you all. The scopes, uh, get the contract, get the a b. I.

E

Figure out all the tables, get all the table. Names and per table, get all the scopes and things of the sort,, and that way you can extract all the info and that's great.

E

But you know, running, like a full diffuse for wax, is not something that the average user can do.

E

It would be much nicer if you can just get the snapshot file.

E

and you don't need anything else,, which was what the diffuse and migrate tool used to do for older versions of node eos, you ba uh,. You basically just give it a snapshot.

E

File, and what it will do is it will just output this big hierarchy into like a,, a folder where it'll be like he, he,. He like,, like some type of like predictable.

E

Hierarchy and then you can kind of go down to exactly what you need and it just runs. Locally.

E

No, no need to load into no ds and output. Is,. You know, small cuz. It's divided in using that hierarchy.

E

So, each far you need, you can grab like right. Away.

C

Thank you for inspiring.

C

Excellence. we are about at time.

C

um,, so I appreciate everyone's feedback.

C

Um, and I'll hand it back to daniel.

A

Well,, thank you very much,, brian. um, and thank you. Everybody for participating and sharing your feedback.

A

We'll, uh, wrap it up for today and see you all in in a week.

A

Have a good day,, everybody.

G

Cheers everyone.