EPF6 Office Hours | Sally

Okay, awesome. Okay, so yeah, it's a it's a bit smaller uh call today, but uh nonetheless, thank you so much for joining everyone. Thank you so much for being here, Sally, today. Um so we have a another special office hours today. Uh in a special time today because Sully joining all the way from Australia.

uh so a special guest but we have actually a bunch of fellows who are from East Asia and uh and these regions where they normally cannot really attend our regular office hours. It's uh hard to coordinate all the world in one time. So I'm glad we could uh schedule a special one for uh for some more people to join. Um so from the east uh so thank you so much for being here. Um uh Sally here works in consensus besu team uh has been contributing to the best execution client for years and um yeah we would like to get to know you uh learn a bit more about uh what is it like uh being a core defe what uh what your story is and let's get started with your uh with your current work or what do what is your role in the besu what kind of uh uh issues you deal with regularly what uh what sort of work you do in uh in the client Yeah.

So, um I've worked for Consensus and worked on Basu for over seven years now. So, um yeah, a a lot has changed in that time. I started off um more in enterprise feature side of things. Um and since I've transitioned over to mainet basu is mainly focused on mainet now. Um uh about 12 months ago I sort of um became involved in the performance team.

So, we've focused um quite heavily on performance with the push for um pushing the gas limit and that sort of thing. Um we have dedicated performance engineers. I'm not a performance engineer, but I'm working with the team to um help with that work. Um a little a little bit sort of project manag prioritization. Um but helping out with with tasks uh development tasks as well with that.

but also as well still doing core development on you know the main basic code. Um so yeah I kind of I'm a bit of a generalist. I'm a bit across a lot of things. So I deal with RPCs um hive tests various bits and pieces anything everything.

Okay that's really cool. Um yeah uh so generalist yeah focused on so many things and I've seen that so contributing for seven years that's that's incredible I mean huge respect for that uh we got bunch of people here for them it's more like seven weeks or seven months so it's uh it's a very different stage and I'm wondering uh if you can tell us more about how uh how you got started I mean what brought you to uh to consensus to Bessu how did you get started with the core development uh yeah what was your journey like started

yeah So, Basu um obviously is a client written in Java. Uh I've been a Java developer most of my um career and a former colleague sort of reached out to me um about seven years ago and said, you know, um I work for this like crazy blockchain thing. Um you know, and we're looking for Java developers. Um would you be interested? And that's kind of how I got started and blockchain.

So that was 2018. Blockchain was pretty new. Ethereum was pretty new. I'd sort of heard a bit about stuff and listened to some podcasts, but I definitely didn't know very much about the technical side of of blockchain. So that was all new for me when I started.

Um, and it's sort of a distant memory now. like it that is a long time ago, but I definitely, you know, can relate to the struggle of like it's all new and it's all weird and different and um yeah, there's a lot and it changes so quickly. So, every every week, every month, there's a new, you know, new EIP, new gas limits, new everything. Yep.

Yeah. I mean so that's that would be an interesting question like how do you keep up with all of this like uh for seven years also I understand you were coming from more like a traditional development role like more like corporate sort of job right like how do you how do you deal with this like crazy and open source and fast-paced sort of um evolving ecosystem

yeah it's a challenge um I think it's a challenge for everyone um I think you have to pick and choose like I don't think you can be across everything there is just too much. So I think like I know I said I was a generalist. Um we do have a lot of specialists in the team. So I I don't you know dive deep into every sort of um like conversation thread that's happening about every EIP. Um I I can't do that.

Um and we do try within our team to sort of you know someone will follow a particular um conversation or thread or or channel or whatever and share um things within the team that sort of come up. I think you have to you have to pick and choose but if if you know like I guess this group it's more individuals rather than like working as a team you're more working as individuals but I think like whatever your project is you have to be really focused on that and you know keep a breast of other things as they're relevant but you you can't digest everything it's impossible. Yeah, that's uh it can be pretty overwhelming like it's uh um it's a lot

at the same time. Yay,

we're using that more and more. Yeah.

Yeah. Okay. Yeah, that's that's that's interesting. Cool. Cool.

Um Yeah. So, yeah, seven years. It's been it's been a while. So I'm wondering if you maybe can give us some highlight of uh of the uh of your Ethereum based career like uh um some uh interesting challenge or problem that you solved something that uh you are specifically proud of or you remember uh um that that that you really enjoyed working on? Um yeah, anything comes to mind?

Um well recently going to the Berlin interrupt, that was my first interup um event. So that that was pretty cool. Um the merge was definitely a highlight. I was only peripher periphally involved with the merge. Um that was sort of around the time that I was switched across to mainet development a bit more so.

Um but like that the merge is a huge achievement across like so many teams and I think every every hard fork it becomes more and more obvious how what a coordination feat it is that um you know that it happens every time.

Yeah. It's uh each each one of them is like our little moon mission, right? Like coordinating on all of this and uh and yeah, I mean uh it's really cool also your first interrupt like I just met you there in Berlin. I uh uh and after seven years like first time meeting all of these people across the teams like it also must be interesting feeling like yeah um just just being online and then uh then finally meeting like just 100 people there. Uh right.

Uh yeah, that's uh that's quite a change. Um cool. Uh folks, if you have any questions for Sally, feel free to raise your hand. Uh feel free to write in the chat or or uh whatever you prefer. Uh we will take questions from the audience.

I have uh I have some of them myself. I want to give you guys opportunity. Don't hesitate. Don't be don't be shy. Um we have I think we have at least one person working on Besso here.

But yeah, if you have any kind of question about uh technical or uh or uh uh some cord of advice, feel free to ask. Um uh yeah. Meanwhile, um I'm uh I'm wondering like whether you have some uh some advice or some uh um like I I wonder about the Java language itself because like me learning learning the programming like I was doing C and then try to get into some other language looking to go rest Java and uh I was always told to avoid Java. I was I looked into it like it's a it's a tough uh it's a tough uh syntax. It's hard to optimize.

It's a it's uh it was it was challenging when I looked into it and I I still like have a huge respect for people who are able to deal with it on such a complex scale as as Bessu as the whole uh client code base. So um yeah, I'm wondering what's what's your take on this like Java compared to maybe other popular languages like what uh uh where do you see the the benefits where or how do you deal with uh maybe some uh some of these uh uh struggles like JVM optimizations like being able to uh get the performance out of the client. Um yeah um how do you how do you feel about the language?

Um yeah good question. So, Java, I've I've programmed most of my career in Java. I I did um do a bit of C at uni, but that's a long time ago. Um I've looked a bit at, you know, GE uh G code, rest code. Um but yeah, it's Java object oriented is obviously very different to um those languages.

Um, as far as optim, so I find Java quite intuitive, but then that's like it's it's I'm pretty used to it. I'm pretty used to it by now. Um, it's quite similar in syntax to C, but like the memory management stuff is a bit different. Um, as far as optimize, how do we squeeze the performance out of it? That's obviously the focus um, as I said, of my team.

Um, and yeah, I guess it's it's different to, um, you know, something like Rust. it's you approach it quite differently but um you know that's the specialization that um that my team is sort of is is focusing on and I guess the concerns are are different but the benchmarking um and all the testing suites that the the Panda Ops team have set up like th those are really really helpful um and you know being able to compare across clients and um get results really rapidly that's that's Very cool. Um, but yeah, JVM warm up is really important in Java. So, you have to warm up the JVM. So, you have run, you know, 15,000 iterations of a specific thing before you test whether it's f, you know, performance compares versus like um, ref and guess they don't need that kind of warm-up.

Nethermind does, but I think it's more of a factor in Java than than .NET, but yeah, there's different concerns to worry about.

Yeah, it's interesting. Thank you so much. Yeah. Uh yeah, I mean huge respect like I was um yeah uh like um I think Java yeah I mean uh it's a it's a powerful language just like uh yeah uh hard to hard to deal with for me when I when I ever looked into it like yeah respect like it's uh um it's it's great to see like uh having such a diverse set of clients with with PESU as well and uh um I'm wondering now um another thing that come to mind is like what uh what is some current or upcoming work that you were excited about like some a lot of things happening as we said like uh many many changes coming to Ethereum so what are some EIPs or major features that uh you are going to work on or you're working on something that you were excited about um something that you look forward to.

Uh so block access list is one that um we're kind of excited about. Fossil is another one. Um Justin's really excited about encrypted memp pools. Um I haven't looked into that one too much. Um me um I've been look focused on Fusaka stuff and helping out with getting Fusaka over the line.

So I haven't looked too much at um what's coming in Glance for for my personal what I'm going to be working on. So, um, uh, account ab, uh, native account abstraction. That was one that I was going to maybe look at, but I think someone else is going to implement that now, but I'm sort of on call for, um, mentoring, reviewing, so forth.

Got it. Thank you. Yeah. Uh, yeah, a lot of new exciting stuff. Yeah.

Awesome. Okay. Thank you, Sally. And yeah, before I exhaust my questions, we have a question from the audience. developer Uchi um want to ask something?

Yeah, sure. Hey Sally, so I've got two questions.

Uh one is kind of ka opinated and the other one is um tricky. So let me just straight on you. So you've been in the space as a core engineer for seven years plus now. So I would like to know what have you seen that have changed over the course of seven years and um for someone who is coming in from a big career shift to be a core engineer. What are your advices from your years of experience?

What should we stay off? What should we focus on? And if you to do it again, how are you going to do it? And the second question um what are the plans for BSO for supporting stateless block execution for things like ZKVMs guest programs um client like rats get they have some sort of implementations where you could provide state weaknesses and um and um some other metadatas with the transactions and you can validate that yeah this set of transactions are properly executed without having to hold a full node or like an asset node or something. So what are the plans for that for for B cuz it's it's important.

So you could have um um this as one of the client also that you could come down to maybe Rick 5 or something intermediate to Rick 5 to to um still enjoy the client diversity when um doing things ziki props on like um Zik V like SP1.

Yeah. So I might answer the second question first. So with with regard to risk 5 ZK and um real-time proving and that sort of thing that is a research area that we're looking into um right now how um Java can um compile or port over to to a risk 5 um environment. Um so yeah that's definitely something that we're that we're looking into. Um also how that relates to linear.

So linear is a zkm that uses basu um underneath um we work closely with the linear team. So that's also um an area of research. So that those sort of um the the nexus of those things. Um that's not something I'm personally looking into, but it's definitely something that the team is looking into. Um, and then the your other question about like what's changed in seven years, like a lot like just about everything has changed in seven years.

Um, and but what how would I do it differently? I I don't know. That's that's a really good question. like we we talked about, you know, the the fire hose of information and how do you keep across everything? Um and and you just can't.

But I mean, you know, there might be different choices that I would make along the way, but I don't know, looking back. Um yeah, it's a good question. I probably would have gone to a interop event sooner, I think. Um I think I was a bit scared actually and it's a long way from Australia. um and you know a long time away from um family and stuff but I think I would have probably gone done that earlier.

All right.

It's um it's always a long way for you guys like we need revenge of Australians and do the next interup somewhere like

New Zealand or coffee day.

Yeah. Yeah. Yeah. Yeah. It was it was actually an idea and I think it would be so fair to you guys to to do something like that.

Yeah. Um Ellie has a question.

Uh hi. Uh my question is that uh why in the uh codebase of BU I'm actually working on um BU as my project and Oh,

oh yeah, I'm sorry. I think her connection could cut out. Yeah,

I think Irene had her hand up as well.

Oh, Ellie.

Oh, okay. Yeah, it's back. Hey, Ellie, are you with us now?

Sorry. Yeah. Yeah, I get disconnected. Uh, why um Hey. uh why the um some of the services inside the codebase has been defined as plugin API.

Yeah. So the the plug-in API um that's a way that you can um have code that's outside of core basu code but you can have a plugin that will implement a specific service. So you can basically so an example is um like permissioning of transactions. So you might um write a plugin that will allow um certain types of transactions and um filter out other types of transactions, but you you don't we don't want that code in Basu, but we want to give people the freedom to write that other um their own implementation of what whatever services. So there's various services that the plug-in API will provide.

Um and permissioning is just one example selection. Um there's we have a Rox DB plugin. So you could have a different database plugin um um that sort of thing. So Linear I spoke about Linear. Linear makes extensive use of um plugins for doing various things and we've sort of been expanding on that plug-in API in in concert with Linear, but there's other um other people and projects using plugins for doing different things.

Um, but it is it is a work in progress. So if there's something inconsistent that you've come across, um, it could be the case that it needs needs work still.

Um, yeah, thanks. No worries.

Awesome. Thank you. Glad to have like a technical specific uh question. That's awesome. Yeah.

Uh, any more questions, folks? Don't be uh don't hesitate, don't uh don't be shy for you to use the the space. Irene,

hi Sally. Amazing to see other women in the space. Thank God. So my question would be more about like your personal experience. So since you said that you came you come more from a traditional tech background and now it's like seven years that you're working on Ethereum.

So like um after understanding what Ethereum was all about or the concept of decentralization etc etc what made you decided to stay rather than going back to something more web more traditional besides the technology also because at the end of the day so for example my project is is um together with load stars I'm like coding with Typescript but you could also code in Typescript in web two companies in general.

Yeah. Yeah. Um I I just kind of fell in love with the the technology and the like it's such a rich um landscape of problems to solve. Like there's just so many um things you can dive in like different areas you can dive into different um yeah just there's so many problems to solve and that the answers aren't like nobody's figured out all the answers yet. So it's just really interesting to me that um you know like web two like not all the problems in web two have been solved but there's a lot of um you know work where you're sort of implementing some where you already know how it's going to turn out whereas um Ethereum web 3 like you know like anything could happen.

Yes. Basically have more or less the same motivation. Yes. Exactly. and and the like it's like it's a rich um tapestry of problems to solve, but the people are are so interesting and motivated and excited and um supportive and and interesting like that's really valuable as well for me.

You mean people in terms of like uh personalities or like uh different interests or like people in terms of um I guess just all the different um client you know like my basu team the tech team we work closely with Lineia like um but also all the other client teams like it's just I'm I'm learning something from every person that I meet like there's you know there's there's no boring people in Ethereum. Thank you.

Oh, that's awesome to hear. Really nice question. Thank you. Thank you guys. Awesome.

Um yeah, uh feel free to keep asking guys. Uh we still have uh another half an hour to go. And Katya has a question. Nice.

Hi Sally. Uh I wonder about your early years of development career. How did you choose development um as a field as your career? How did you start? Right.

Um, so way back at high school, I had no idea what I wanted to do. Um, I didn't I didn't really I really didn't know. I did I actually did a year of medicine um cuz I thought medical research might be interesting. But um I yeah I did a year and then at the start of second year I um decided that wasn't what I wanted to do. and I went to a computer science lecture cuz I had a friend who was studying computer science and um thought yeah that that's really cool um so yeah switched over and um the rest is history

so it's your um education right so computer science and then you started work as a developer

yeah so I did a bachelor of information technology and um I got my first job um while I was still at uni. So um that I remember that year very well. That was a bit of a struggle. Um you know being a bit overcommitted and trying to do honors and work um like halftime or something at at the same time and sort of feeling like I wasn't I wasn't doing anything well enough. But um got through that and uh yeah, ever since then I've I've been a software engineer.

Um, yeah. Thank you. Always always really liked it.

That's really cool. Yeah. Uh, love it. We have so much female energy here today. Uh,

and uh, yeah. Um, it's um, it's inspiring. Um, nice. Um, folks, yeah, feel free to keep asking. I uh I had some some thoughts but uh want to give you guys opportunity to to meet Sully to learn from her and uh I'm not sure if I see all the messages in the chat like uh in case I miss your miss your question let me know.

Um yeah. Okay. One question I was wondering about or one thing I was curious about is how um in team or as it best like how did you balance this like uh what you mentioned before uh sort of commercial private blockchain uh sort of business and uh and the um the mainet. you mentioned that you moved more from uh from one side to another and uh uh but you also mentioned you work with Felina and so on. So I'm wondering what is the difference there like uh how different it is to work on mainet and like these uh more private use cases and uh yeah um uh yeah if you have any thoughts on that.

Um

yeah. So I guess the um the the concerns are different between mainet and um private or permissioned uh networks. Um the security concerns are different. The transaction volume is different. Um but with with BESU, we made a decision um a few years back to focus more to prioritize mainet over the the private um and private network features.

Um and actually in the last sort of 12 months we've actually uh removed and deprecated gone through a process of sunsetting those features and um deprecating and removing um uh some of those so that we can sort of remove some of that tech debt from the code base in order to be able to speed up and simplify the mainet side of things. um and the the plug-in API is actually part of that. So if um if people still want to you know use Basu in that way then some of those features um we've sort of said well here's an here's the API you can write a plugin to do that if that's what you want to do and some of the code can be moved out of Basu into a plugin. Um because it is actually really hard to focus on all of those things and for basic to sort of be a Swiss Army knife that's trying to do everything and you know like me trying to um study and work at the same time. You feel like you're not doing a good job of of anything that's important.

Um so yeah we we sort of recognize that in Basu and uh I mean these things take time but um yeah some of some of those some of those features have have now been uh removed from Basu. Um but we do still support uh permissioned networks um in addition to mainet um and we work with um so Basu a lot of the contributors to Basu work for consensus but um it's also open source so we do have um some folks from Kido who help out um with BESU as maintainers and they are more focused on the private and permissioned um side of things. So, um things like QBF, the QBFT consensus algorithm and um other things to do with um private transactions and um that sort of thing. So, we we sort of work in conjunction with them and um but we at Consensus are more focused on the mainet side of BESU.

Got it. Thank you. Yeah. Um yeah, it's interesting to see how you balance these different use cases basically as Ethereum client because also like you want to have some sustainability and business case. So I understand it's been it's been a focus on of uh of the client before but also the ecosystem or the the market change like I remember like back in the the earlier days bunch of companies just excited about having their own private blockchains which uh uh now rarely happens I guess.

So, um yeah, that's that's interesting development. Yeah, thanks so much. So, yeah, nice. Um yeah, we got uh we got couple people leaving, but we still have some audience, folks. So, um uh folks, if you have uh some more questions, feel free to ask.

Katya has another one, so go ahead. Um uh I would like to ask about work life balance as a core developer because um sometimes it seems like they are really overloaded with the work and uh especially um to hear from from a woman it's interesting how you balance it.

Yeah, good question. Good question. Um, I think as well with working from home, like it's it's pretty easy to just spend way too much time sitting at your desk cuz it's always there. Um, so I think it it is something to be mindful of for sure. Um, how how do we balance?

So within it, so the basic team, we have we're spread across uh several different time zones. So we have people in the US, people in Europe, and people in Australia. Um, so I guess we're pretty deliberate with not expecting people to be online all the time, even though someone is online all the time. Um, so we do try and sort of hand off things between the time zones and, you know, um, different So in Australia, I don't I don't go to the allore devs, um, any of those calls cuz they're always on at midnight. I I just I I just can't do that.

Um so we do um try and make sure that we share the notes. Someone someone who goes to the call shares the notes. Um and we we try and have a bit of a study grip before those calls as well so that even if we can't attend the calls, we can sort of be conversant in the issues and the EIPs that are being discussed and any decisions that are likely to be made and that sort of thing. So that's that's one thing that we do try and do within the team. Um we do try and have crossover meetings like with the different time zones so that um you know Australian people uh talk to Europeans at a couple times a week or um and we talk to the US folks uh you know on a different day and um that sort of thing.

My team is spread across Europe and Australia. So we sort of have calls at this this time of the evening. Um, but if I do, so for me, my work life balance, I often take a few hours out during the day. Um, if I know I'm going to have evening calls, um, yeah, I I don't know what I do. Um, exercise is important, you know, all the usual things, but it it goes in waves, too.

So sometimes I, you know, have a really busy week and I feel a bit smashed and then um, you know, I'll take an afternoon off or Yeah, we we we like to let people set their own hours as well. But um, yeah, I don't know. It's a continual struggle to find the balance, right?

Yeah. Thank you.

I don't Maybe that was a terrible answer.

No, it's all right. It's all right. It's it's a good answer because

Yeah. Yeah. Um, yeah. I think we are all dealing with a similar problem like working remotely as well like managing your own

uh time management in the in the day and everything. Yeah.

And and sometimes there is like a really critical important issue that you do need to stay at your desk and and finish off or hand over to someone else or whatever but sometimes some days there isn't and you can you know take it a bit easier. So yeah, every day is not the same.

Yeah. Yeah. Thank you so much. It's interesting. Awesome.

Awesome. Hi folks. Okay. Do we have more questions? Like um I like uh we can we can Okay, Irene.

Yeah, go ahead.

Yes. Now more of a technical question. So uh Vitalik recently tweeted about ZK privacy. What's your take as a core dev on the privacy and anonymity side? Um I I have not read Vitalik's tweet so I'm not sure of the the context um specifically.

I I think um you know like privacy for some things is important. Um yeah I I don't know. Uh what what's your take on that Irene? Um it's more of like my like values more than the tech. So I'm still like checking ZK but not not as a like uh something technical but more as a as a concept.

I know that like many layer shoes have been develop develop developed now. Let's see in the future. I don't know if it's just like um um I don't know like just the hype of the moment or there's going to be like some kind of future or some good ideas. For sure there is but like you know the economic incentives is always strong because I know that like as a code later are also built on the top of what you are building. So like you really can see people using what you're building at the core level.

Yeah, I think I think it probably makes sense for that sort of privacy to be in the layer two more than layer one maybe.

Yes, that for sure. Also probably because it's super complicated to be like fully to go fully private on on the layer one like basement. Yes. So do you see any values in like ZKbased projects? Uh not values but but like something that might be useful in the future.

Um yeah for sure I think ZK can be used for loads of use cases. Um then there's a lot of development happening in that space. I

think so.

Yeah. Amazing. Perfect.

Okay, thank you. And uh Oh, we have a couple hands up. I think Charlie was first. Finger first. I think it was very close.

Hi, Sally. Sorry I was late. Um just a quick question around uh like when you come to Ethereum to start with, it just seems like so unbelievably overwhelming with where you can start and the entire landscape that you can get into. Do you want to just uh talk a bit about your experience with that and where you decided to start and uh Oh, sorry that's my daughter. Uh on on sort of like you know first taking the first few steps into the ecosystem and contributing.

Yeah. Um I remember reading the yellow paper. Um that was like I was told to start start with the yellow paper. Um that was kind of intimidating to be honest at the time. Um I think starting with like a good first issue, you know, starting with like an actual coding task that you can achieve and complete that doesn't require a lot of prior knowledge.

Like I think especially if you're new to open source, if you're new to Ethereum, if you're new to any of that stuff, um you know, getting some small wins is um empowering and let you get all your tools tools set up and all that sort of stuff. Um and then then tackling something a bit more complex Thanks.

Yeah, thank you. Um, yeah, it's it's interesting. I mean, the yellow paper it's it's intimidating. It's true. And uh because

Yeah. Um, yeah. I mean for the execution and because like last week we had Julio here from Aragon and he's like the sort of a the sort of a contrast here because like 22 years old and you know started started uh uh much later than you and and he he also got started with the yellow paper although like it's it's uh it's a tough one but uh that's what that's what uh he got in. So okay it's still it's still strong. I I I respect that.

Awesome. Uh so Ellie has a question. Um, what are the aspect of a good PR to you?

Aspects of a good PR? Oh, good question. Um, so I like a PR that's not too big. uh that um you know like 20 files or less 100 files is like no I'm I'm asking that person to break it down. Um yeah so not too big um a really good crisp explanation in the PR description um and the title.

Um, I find if you can look at your PR with the eyes of a reviewer and um, you know, does it make sense? What questions would I have if I was reviewing this? Where should someone start? So, if something's like a oneline change, then it's it's really obvious when you're reviewing it what part you should look at and how you would struct, you know, structure your review. But if if it's like 20 files and like 300 lines have changed, well like you reading it from top to bottom isn't always the best approach.

So um you know sort of a start here or like these are the most important changes that you should um be looking at. Um test coverage is really important. Um uh what else would I um if uh it should compile like it there's some there's some checks in Basu. We've got a bit of a checklist on the PR template that says like here are some things that you can run locally before you push your PR to like you know ensure correctness but not only to ensure correctness but to ensure that your PR can be reviewed in a in a timely manner. Um so yeah com you know compile we have a llinter that you can run spotless um and some some tests that you can run locally obviously those tests also run on on on the PR once it's pushed to the environment but um if you can run them locally uh that's gets you a review faster.

Um so yeah if your code doesn't compile um it it can't really be reviewed for example. Um so if you can fix those things before asking someone to review then that's that's excellent.

Uh actually some of the tests on the basic code base uh doesn't pass and it is um yeah it might not be related to the PR.

Yeah. Uh, do you wanna can you post a message in our Discord for that? Um,

yeah, sure.

Yeah, if that's if that's true, that's something we should fix. If it's something specific to an environment or whatever, then that's something we should be aware of. But yeah, if there if there are tests that um are not passing for you, yeah, shoot us a message in Discord.

Yeah, sure. It sounds like those tests are um kind of old. I mean,

yeah. Okay. Yeah. Yeah, let us know. It's true.

Cheers.

Very well. Awesome. That was that was a great question, by the way. Uh maybe we have some some uh more more of those um some more specific uh sort of advice you guys would like to learn. Feel free to ask.

We covered many things already technical and uh sort of personal. Um yeah. Okay. I have a I would ask a question then uh if K are giving me the space. Um, I'm wondering uh to phrase this uh what is some uh what is some you did?

What is some what is some some uh some sort of uh issue you caused? Some problems you had, you know, something that you had to uh or maybe something that you think still back in your mind like um um something didn't go uh as you planned and uh how did you how did you fix it? How did you how did you reconcile? there is something that comes to mind that uh uh that you have to deal with. I just want to show the guys that you know nobody's

perfect. No, it's a great question and I have a really I have a really good example of from recent memory. So, um, uh, I pushed a PR to the basic codebase that, um, fixed something. Uh, but it also broke like a thousand hive tests and I didn't check the hive tests on my PR. uh and some someone approved my PR and I and I merged it and I didn't check those hive tests um for until I went to do do a release which I'm still in the middle of doing.

Um uh so yeah, that is definitely something that I uh feel badly about and I'm still still, you know, doing a personal retro of like how that happened, but also as you know, that's something we need to address as a team, but also me as an individual. Um but yeah, there's there's automated tests. We have a lot of automated tests. Um obviously this is something that's not tested in the Basu automated tests, but it is in the hive test. So, um, when there's automated tests, you should definitely take advantage of the automated tests, but also look at the results of the automated tests.

So, there you go. That's a that's an example. So, we had to revert that PR and um we'll have to refix refix that issue because um it was an issue that someone reported of a discrepancy between Basu and and other clients. Um, but yeah, my fix was not good. I see.

I see. Thank you so much. Yeah. Um yeah, I appreciate you sharing that. Um yeah.

Um yeah, I mean that's that's my goal here as well. Like you know coming to Ethereum like all these cordfs they seem like super experienced. They know everything. They're perfect the geniuses working on on these clients. So like we are all people and uh we're all learning.

So just wanted to wanted to show that. Um thank you so much S. Yeah. Awesome. But yeah, there's no shame in reverting a PR.

So don't feel bad if you have to revert someone else's PR. Don't feel bad if your PR gets I mean you can think about what you should have done differently, but um it's not, you know, you shouldn't be too attached to your um code or your PRs.

Awesome. Yeah, thank you. That's that's great advice. Nice. Uh yeah, that's lovely.

We have uh last two minutes to go, folks. Uh feel free to use it uh to ask some more questions, please. Um, and do get some uh some more thoughts, more advice from Sally. Use the opportunity, please. Got some people leaving can come in as I see.

So, if you just came, feel free to feel free to ask. anyone otherwise I will keep asking okay so was it one thing and um I'm sorry if this is too personal like feel free to stop me there but uh I think in in interop that I learned that you raised couple of kids during your career as well and I I find it so crazy like being able to handle family and a big family and we already had a question on this and the work life balance but uh like dealing with my my partner for example like uh uh she's she's so scared to have kids when focusing on the career right like it's just uh it's hard to balance so I'm wondering how did you how did you manage like you already had a similar question but like from the perspective also of family and having like so much responsibility on both ends like uh uh you you're really superman there like being able to uh to handle handle both career and and uh and children and and the household like how do you Yeah. How do how do you manage like what is what is your hack there?

Um so my hack there is lower your expectations. So like you know uh if if you think that you have to have like perfect children and a perfect house and a perfect um you know everything like it's it's unachievable. So, um, like my house my house is really messy and my kids make their own lunch box and I don't I don't care or see what they put in it. Um, you know, we don't know that's that's that's uh what I would suggest if you have children is like that they are capable and they can do a lot for themselves. You don't have to do everything for them even from a young age.

Um, I would say they are more capable than a lot of people give them credit for. and you know they get to make mistakes and um uh have autonomy and um make their own decisions and stuff. Um yeah, but yeah, lower your expectations. Uh that's what I would say if you're thinking about having children.

Oh, that's that's awesome. That's very interesting advice. Thank you so much. Yeah, I I appreciate that. Um yeah, I mean a lot of people in the ecosystem just getting older having kids now and uh and it's uh it's it's it's new challenge that comes in life.

That's it's very interesting and and yeah, thank you so much for saying I very much respect that. Yeah, I mean I I come from a single parent family. So my mom working hard, I had to also take care of myself when I was when I was little sort of uh um preparing my own lunch and stuff. So I I totally agree. I respect that.

That's uh that's interesting advice. Awesome. Yeah. Yeah. Thanks so much, Sally.

A bit more out of Ethereum question maybe, but uh maybe let's get back to back to the Ethereum and core development folks. If you have any any uh any thoughts, any more questions? Uh Irene. Yeah.

Yes. Another question, a bit more technical this time. So what you want to see uh changes in the team over the next two years like uh something that you would like to be improved or more developed or something like that.

Um yeah so I think there's been a lot of a push for scalability and sort of pushing the limits of um performance and numbers and stuff. I think um at some point we you know like we'll we'll get to a place where we've we've done that enough like it's good enough and we should be focusing on other things. I think we're maybe not quite there, but we're maybe getting there. Like I think Glamsterdam is going to be a bit of a focus on other things like EPBS and Fossil and um um for me I think Ethereum like it's sort of becoming a I don't know if this is controversial a bit like we're losing a little bit of the decentralized um ideal with the push for you know more and more um you know scalability, throughput and you know solo daycarees are um being pushed out a bit. Even the mev and um the public me is um being starved of transactions and I I think this is where enchanted meals is um uh maybe going to help.

So that's that's something that I'm a little bit excited about. I don't know a lot about it but um it sounds cool.

Perfect. Thank you. So in terms of scalability you mean like uh like increasing the the throughput so that like we can like reach the mass adoption so or scalability in terms of uh

um so pushing the gas limit like that's something that we're really focused on right now. Um, but like I don't know like can we just keep pushing it higher and higher? I don't know if that makes sense. Maybe.

Totally. Thank you.

Okay. Thank you. Uh, Rahul has a question.

Rahul, if you were uh if you're with us, am I audible now?

Yeah.

Yeah. Uh so I have this question like what do you think about this whole push for risk v uh 5 uh architecture like the whole like everyone pushing for that. What are your thoughts about that?

Um yeah so that's an area that we're in research stages of at the moment with BESU. So I don't have a firm opinion on that but it's definitely something we're looking into.

Got it. Thank you so much.

Okay, thank you. Yeah, another question on the future of Risk 5 there. Like it's all uh it's all still in the stars, guys. Yeah. Um right now we have the proper execution client still checking along with regular gas.

Um and uh oh yeah, maybe that's uh that's one thing I was curious about. I also asked a bit about optimization there but maybe if you have some insight into how you guys handle the new gas increases like it was the whole performance research and uh going now to 45 million 60 million uh what bessu specific changes you had to do what were some interesting changes or optimizations that you that you did in the client to be able to handle it.

Yeah. So um at the Berlin interrop mod x was the big one that a lot of clients had to um make changes to and optimize that one. Um we're looking at EC recover easy add. There's a few um there's a few mod op codes, but yeah, we're uh we're just going by the benchmarks um and they're sort of highlighting different um use case or different test cases that um the different clients need to need to optimize to go to like the next level um of gas. But yeah, some of it's some of it's native um optimization, Java native, some of it's um up codes, some of its pre-ompiles.

It's just kind of methodically working working their way through um whatever test case sort of pings the alert um and optimizing. So yeah, it's it's kind of it there's a process of looking at the benchmarks, gathering the data, figuring it out what optimization to make, and then rerunning the benchmarks, making sure that you've warmed up your uh JVM. It's very important. Okay, that's interesting. Thank you, Sally.

Awesome. Okay, um last few minutes, folks. Uh I give you the opportunity for one last question if there is anyone who wants to ask please do use it. Uh okay Rahul has another one. Yeah.

So maybe another live question. So uh h like uh how do you think uh like personally what do you find is has better output like being a generalist or like contributing to one specific niche thing uh yeah I mean that's a question like um

um what has better output it probably depends how you measure output I I think in a team it's really good to have a mix of both. So to have specialists and generalists. Um I think if you're an individual working on an individual project, you probably need to, you know, dive deep into whatever thing it is that you're working on. So you probably need to be a specialist um in that thing. Uh but every now and then you might need to sort of zoom back out and look be more you know have more of a generalist view of how does this fit in with you know the landscape or the ecosystem or other things that are coming up or whatever.

So you probably need to do both as as an individual but it's hard to do both at the same time. So you might need to you know schedule focus time and then schedule um project management time or something like that.

Got it. Thanks. Yeah. I mean personally I had this problem because I have to drop off some things because of some personal reasons or some XY Z reasons but yeah I mean that makes a lot of sense. Thank you.

Okay, thank you. Um thank you so much for the last question there. Uh we are at the full hour now so we should wrap it up. Uh I'm just going to ask one more time for any last thoughts or advice or anything you would like to share before we finish. Um

uh no I think we covered we covered a lot we covered a lot of ground. We covered some technical uh technical questions and some philosophical questions and some work life balance questions.

Yeah. Awesome. Yeah, it was really great. Thank you so much, Sally. I think we we learned a lot.

We got inspired. Uh I really appreciate you uh uh taking the evening to to chat with us. Um so yeah, thank you so much everyone. Uh thank you for being here and uh uh yeah. Um, have a great rest of the day.

Um, I hope to see you around again in some of the some of the interups maybe or uh around uh in PRs, right?

Yeah, for sure. Um, well, thanks for having me and uh, you know, appreciate the great work that you're doing, Mario, coordinating the EPF uh, program and thanks everyone for joining and doing the great work and learning learning everything from the get-go.

Awesome. Yeah, appreciate it. Thank you so much everyone and yeah, bye-bye.