 And John, thank you for recording this. So more or less, I think the process would go like this is that we're going to, this is the first working group meeting. It's so we'll see how that goes. I think what the expectation is, is that we're going to see what you think your goals are. Then we'll make a kind of a list of what we expect to do in Q1. Then we're going to review that get back to you with the recommendations and at some point in the near future we're going to have a final list of Q1 goals and so Q1 will start. All right, so I think the plan here is that just every group would take about five minutes or so because we have many groups who would like to wrap up in one hour. And then just explain who is in that group, how you plan to organize and what you think you can do in Q1. So any questions on that? If no, let's go ahead and start. Let's start with the deployment group. Again, as we mentioned last time, we were hoping that every group would kind of a select a representative because obviously you don't want to talk at the same time. So that's where that representative should start talking. All right, let's start with deployment then. Is someone going to share the slides so that we don't all individually do it? I'll share the slides. This is the least I can do. John, can you let me share the screen? No, I don't have a normal admin settings in here. There is a shield at the lower. I don't see my normal stuff. Am I like not logged in as myself maybe? Host has disabled screen sharing. Okay, let me log out and log back in. All right, let's do that. I mean one of the most capable software development teams in the world here cannot figure out how to use goddamn zoom. The zoom is goddamn bullshit. No, it works. It works. It's the worst chat thing out there, except for all the others. Exactly. Stay on topic, people on topic. Yes, absolutely, Dave. Thank you, Dave. So I just locked my PSU account. I can't get into how was this meeting? Can you share? What was that? Can you share? No, it says host has disabled screen sharing. It doesn't think it's me. Okay, so what I'm going to do is I'm going to create new meetings. I'm going to put this in the chat right here. And then we're going to all log in. Would you like me to do that? No, I'll do that. It looks like I can. I don't know. I don't know. It gives me a button, but it doesn't work. Do you have like administrative privileges, Jen, to enable it? Or is it just let you know? I just saw the button. Oh, if you click it, it'll tell you you're not allowed, I think. Yeah. If you click on participants on the lower right, can you click claim host? No. You need the host key to the host. Yeah. Yeah. I mean, there seems to be. I can enter a host key, but where would I find that one? Okay. One second. So let's just. All right. So I can create a zoom and put the link in Slack. That probably work. Well, not everybody on this call has access to our Slack. So just hold on, hold on for one second. We should tweet it out too. I can also stick around here for 10 minutes and catch the late commerce. I feel like if you, if you tweet the link directly, you're asking for zoom bombers. No, don't do that. I wouldn't do that. Oh, wait, wait. I'm in. I'm in. I'm in one second. Let me. Right back in. I'm glad all this is being recorded. It's, it's a drama. Just that section is good for. So, John. Hey, I see everything. Wonderful. There you go. Sorry, sorry, everyone. That didn't happen in the yesterday's webinar. I think everyone can share now. So. Yeah, do you want to share? Okay. Deployment working group. Okay. I think I was a volunteered to. Lead this one. So. This is the, the deployment working group. The purpose of this group is basically not to develop galaxy itself, but to develop the solutions of how galaxy is installed. So, yeah. And we recognize that there are different ways that galaxy will be installed and run. Everybody's site is different. Everybody's technologies. Technology preferences and so forth are different. And so. We want to, to support those, but also. Unify the efforts of. We want to continue to support these things, but also. Not, not develop in, in sort of parallel. Without. Crossing over. Okay. Okay. So. So, we want to continue to support these things, which have mostly been working on Kubernetes and helm based deployments. And the, the more traditional group, which has been doing Ansible based deployments. Since there's a lot of duplicated effort, we want to try and continue to support. So, we want to continue to support these things, but also. Crossing over. Okay. So. I think I covered all this. But here's a group members. So from a number of different. Labs. And. I'll go ahead and go to the next slide. Our priorities for the, for the first quarter next year. We want to make sure that we have all the, all the production ready deployment of it's in the various. Ways that we deploy galaxy currently. So, especially for anvil and for public galaxy servers. We want to make use of it and figure out all the, the bugs and pain points with. Getting pulsar to directly interact with object store back ends. We don't have to go over this image, but essentially it means that currently for pulsar to stage data. For, for executing jobs. It has to go through. Galaxy to do that. And it would be much more efficient if you have an object storage system. Behind galaxy. So I rods or S3 or swift or stuff or whatever you were using. It would be more efficient if pulsar could communicate directly without object store. And so that'll be one of the things we'll work, we'll work on for, for first quarter. And then a third, third piece will be to finish up the work to make it possible to pip install galaxy, which should ease deployment quite a bit. So the, the it's, that's fairly well defined. The pulsar direct interaction backend object store. That sounds like a lot. It's a lot of work. For Q and no. It, it, it in theory works. Right. So, so, John did a whole bunch of work to make, to, to make this possible. But I don't think it's anyone has done it outside of John, you know, working in his development environment. I just really want to emphasize it. I mean, I, I'm not saying it's good or bad. I'm just, I'm just saying that it's a good thing. I'm just saying that it's a good thing. I'm just saying that it's a good thing. I'm just saying that we can do. Because not only this would benefit galaxy, it would also benefit this team. Once we actually see that we set the goal, we do it. We move forward. So just, just please. Think about. Actual. Because. Is that achievable? I'm not saying it's not. But just take that point seriously. For all working groups. Yeah. Yeah. Yeah. That's interesting. Like what is the. Deliverable or what is, how are we saying that it. Cause I think it does do that. Right. So what does it, is it a piece of documentation? Is it working on main? Is it. A default option in the cloud product with test cases or. Okay. Yeah. No, that's good. We should define it. I mean, I'll, we'll have to talk as a group, but personally I'd be satisfied with. You know, I think we're going to be able to do. Getting this set up and working in environment, other than just a pure development environment. And being able to. Explain to someone else how to set it up. Okay. I'm also not. Okay. We might do this for high glass. I don't know. Jupiter. It's a very well defined goal. We need to do this. So again, the idea is that we're going to talk about this. We're going to talk about the goals for all working groups, then we're going to have PI meeting. And I think what will come out of this meeting will be that. We would just like, for example, in case, the second goal here is that what state, as John said, what the actual sort of deliverables here, what's the. Break down. Concrete things. So any questions to that working group from anybody. Specifically from PIs. Okay. I guess you're concerned about the, you know, the two, I guess I'm concerned about what's needed for actually PIP install galaxy, right? So how close are we to that? So I have to come back to that. I think it should be easier now that so much work has been done on the packages. There are just a few cases now where we're reading from config files instead of defaults in the modules. And that was the main thing holding it back, right? Yeah. So you have to start galaxy from the root of the galaxy clone, which doesn't exist in a PIP install galaxy. So my, my estimation is we're like 95% of the way there and it would take someone, maybe a week. Two weeks, I guess I'm bad at estimates. Two weeks to get it working. I mean, I had it working in the past, right? Yeah. But a lot of stuff has been added since then. So there is stuff. Yeah. Yeah. I mean, that, that'll be a really awesome thing if that actually works. Yeah. Just for everybody else. Another point of this meeting is that so all working groups know what the other, because we cannot achieve anything if you guys, if working groups don't talk to each other. Yeah. So, you know, I think there's a chance to deprecate any of the install options as this one becomes online from the webinar, Tuesday. You know, we just, there's a lot of options, which just brings a lot of maintenance overhead. I think. Certainly, I think long-term, I wouldn't want to do like the run.sh and run, you know, method that we currently use. Like everything should be, we should provide some kind of, you know, a way of installing Galaxy and running it simply. So with like a pip install Galaxy, you know, it should just be pip install Galaxy and then run some command called Galaxy and it starts up and you have a fully functioning Galaxy server. The goal here then would be then to deprecate the clone and run. The clone and run would still be useful for development, but we don't need to make it easy at that point. You know, Yeah. So I think we should be talking about working groups before we get into too many of them. If people are interested in joining a working group, we might want to have the speakers mentioned how to do that or who to contact. I think you can, you can contact the working group, but so we had this artificial requirement of not having more than for a person not being more than in two groups has probably not going to work. So it's not really set in stone. It's just that a person being in all working groups, because then, you know, what's the point. Okay. Well, thank you, Nate. So the next is just to that just one more, maybe question be nice if, if working groups are going to have regular meetings to post that time on the slides. So that if somebody is interested or wants an opportunity, there's sort of that as a standing open door. So I think in addition, most of the groups already have Gator channels, but all groups should have that. Can we add a section to the events or a new calendar for these things too? Yeah, we can. I've also created a working groups page where I'll put a bunch of summary information pointing off to wherever stuff is. Great. So one of the things that we're going to do after this meeting is that we're going to follow up and you will have to tell Dave what your meeting times are. So we have this on the galaxy calendar. I think in the groups that I am, we've decided against meeting because it's across multiple times on some, it's just easier and faster to do everything on chat. Async first. Yeah. I can confirm that for the tools working group. It's even easier for people to just show up and read the history on Gator or whatever. In this case, it should be very clear what that Gator channel is. So Dave, we're going to have to create a page where all of that is aggregated. All right. Admin group. Formerly known as use galaxy dot star. I don't think we appointed a leader for this. If someone else wants to. To do that, that's fine. Otherwise I'll continue on. So right, we had a discussion and it seems like use galaxy dot star is sort of a sub project of admin in general, especially because there are a lot of plans that we have for use galaxy dot star that once we implement most of them, there won't be a whole lot more to do with it. Of course we can always come up with new stuff, but it seemed like the use galaxy dot stars is a more useful concept as like a project of the admin working group. So right, the goal here is to develop solutions for large scale production galaxy instances. And then figure out what it means to be used. We've already done this to some extent, but really formalize the process of the use galaxy dot star. Server experience. I think, I don't think made for you. It's a good idea to be head for both groups. You're going to go. I mean, you're already crazy. You're going to get crazier. So, so there should be, it should not be, we should definitely prevent one person being the responsible guy for, for two groups. So we'll, we'll discuss it internally down and someone can, someone else can be in charge of this. Next slide. Yep. So, so priorities for the, for the first quarter, there's probably a lot here, but there also are a lot of people, I think who are fairly motivated to work on some stuff. There are some issues, especially that you, his experience with pulsar that we want to get resolved within the first, you know, pretty, pretty quickly here. And then deploying the common tool set to, to CVMFS and actually getting our servers using those. All the tooling is in place and we should be able to do this, especially if we cut out conda, which is the plan. And so, so we can, we can move forward with that. Unifying the, the look and feel of the servers. Determining the, the suitability of using ref genie and maybe getting started on that process of, of a automated community driven way to, to get reference data in the galaxy and figuring out the conflicts with our existing reference data. As I said, defining what a use galaxy.star server is and formalizing that process. And then getting GTN tests that's currently already happening on you and AU, but not for the US server. And then testing tools that are deployed automatically updating tools that are deployed. And then regular meetings, although this is possibly the most difficult group to do it with because we're, we're spread out everywhere. So that sounds much more, much more than just Q one, but again, we'll, we'll refine that. A lot of it. So a lot of this stuff is done for like one or two sites, right? So there, some of these are relatively small tasks to do. Okay. Any questions? So I mean, so with the pulsars that just the issues that EU is experiencing or, you know, it's a very large thing, right? So, so we've discussed the idea that pulsar maybe needs a more full-time or, you know, dedicated maintainer. There are a bunch of us who work on it in little bits. That as, as we find problems and need things. But it might, you know, it's a, it's a sort of critical piece of our infrastructure at this point for all these sites. So we need to, we need to, we need to, we need to, we need to, we need to, we need to, we need to, we need to have a lot of attention on it. And, you know, I have a few wish list things as well that need to get done with it. And the back end working group has pulsar development tasks also. It would be amazing to have a full-time. Also developer. Awesome. Okay. And for those of you who are on this call, who are not usually part of these meetings also, please. Everybody's equal here. Please go ahead. Okay. I, is that Nate, you too? Yes, it is. So. This, yeah. So we discussed this maybe as a project also of the admin group, but more ongoing. Because there will always be new challenges with Maine. But we may keep it as a separate group as well. I'm not sure. We do need a place to sort of keep track of the priorities for Maine. Even if the folks in this group don't end up doing a significant amount of the development. It is something that probably should inform a lot of other groups and the work that gets done. So. Priorities for the, for the Maine working group. This is use galaxy.org for, for anyone who isn't familiar with the nomenclature. So. Interactive tools for course number one. Developing a storage policy. So right now we just keep everybody's stuff live on hot storage forever. Which is not sustainable. So in the first quarter of next year, we want to start storing new data in iRODS. We already have our test instance writing here and we've been able to do that. And we've been able to do that. We've been able to do, you know, things that come with that. So, so we're pretty far along that path and want to start putting it into production. Get training infrastructure as a service. Deployed on Maine. And. Use the data that we've been collecting on tool memory usage and do a better job than allocating memory for tools. And then we can use the data to increase the throughput. And then any sort of additional goals. So there is a lot of stuff. And this is why I talked about maybe folding this with admin because reference data is a huge. A concern for, for Maine. And that's an, that's an, a priority in the admin group as well. My understanding that you were talking to Bjorn. To basically not have that working group and do fold this into admin or somehow. You, I, you X. Okay. So this one's pretty self-explanatory. You, I, you X enhancements primarily to galaxy first, but in talking to folks, it seemed like it might actually be really useful to. Cross this group up or engage with like, you know, Cloudman developments and things like that in the future. So, so to be able to like reimplement a new Cloudman UI, so there's the members scope might grow to more generic UI, UX stuff. So the first, the two things that were suggested as priorities, we actually hope to have finished Q4. So before Q1, that's a history PR and IGV.js. The history PR and that is pretty much ready to go. Hopefully we can just merge that. And then I do anticipate that a significant portion of Q1 will be sort of iterating on little refinements there and, and, you know, kind of tweaking things and, and catching it up with changes that happened since it was first branched. And IGV.js will be a nice kind of prototype new visualization. So for Q1, we want to focus on tool form. Viewification probably is the largest single project. We're going to do a bottom up each of the components on the forms which are used in more than just the tool form. Those can be individual components that are used elsewhere in the application. So we can, we can sort of componentize and parallelize development here and hopefully have a completely view based tool form by the end of this, eliminating a whole bunch of old backbone code. Batch operations are another highly requested feature. So that's kind of like in Gmail, right? When you want to delete a whole bunch of emails and you've got a hundred thousand emails, you can say like match all emails with this search criteria. That's kind of what we want to be able to do in the new history with changing data types or, you know, like naming, adding tags, things like that. So that's Q1 task. And then the next two are a little bit more flexible. There are projects scoped out with like little sub components, but general library and collection UX. There is a whole project board for collection UX that has a whole bunch of issues will refine exactly which ones are going to be finished in Q1 and which ones are stretch goals or whatever over the next week or two. But same thing with libraries. Oleg already jumped on a big part of the library's thing and massively improved the recent changes to that. And then the last new feature is the new user welcome, which is something that first time users will see. And it'll just, we've had the issue open for a couple of years now, but it'll be a nice intro to Galaxy showing what's where, linking to resources and things like that. That's meant to be tailored to new users. And then you can always revisit it later, but kind of an intro to Galaxy tutorial built into the application itself. So given the previous history of history PR, do you really think this is going to be just part of Q1? It seems to me like it's the whole Q1. So it'll be merged in the next couple of weeks. And yes. And then so the way it's going to be merged and it'll be an off to end feature where initially nothing changes for users. But if you click in the little panel, you can say, I want to use the new history and you can see all of the new features. And I think that's what we really need to more fully flesh out what else needs to be done too. And to figure out what we need to refine. But like getting it merged will happen Q4. And then we'll probably spend the next quarter, you know, I would really encourage you to maybe stick to this too. And then we'll discuss this more because you know exactly what technical implications here. I don't. But it's just really, I know you're a Superman then, but it's nine of us. I'm Superman. But and super humans. But I really want to experiment with this idea that we sort of get something which almost looks like not enough work. And then, but really doing it well. Yeah. So the tool form was the big, it's, it's, it's. I struggled a little bit with balancing. Work that's exists, but it's almost done. So another thing that I could have sworn I put on this list, but fell off the bottom. It's on my other sheet here. Is the UI for workflow implications. Marius already has most of that done. And I would love for that. That's one of, it's actually one of the things that it just depends on the history PR because of using the same components. Yes. So a lot of this is kind of tightly integrated. We'll have to see. How well it works. Once it's all integrated, but. Any other questions. I don't want to dominate this conversation because from all this crew, I'm the least technically capable. So any other questions. I have a, just a. I guess a UI UX impacts everything and everybody predominantly the users. And so, I mean, this list doesn't include Marius, Nick and Sam, who are users of Galaxy or, or, you know, the main, the main experts. So, but I still feel like there's just a lot of times. You know, things get done that, that are very sort of opinion based. So I'd like just how can, you know, input be provided to the UX group. From, or how does the, the UX group both how can input individuals provide input and be listening to more sort of, or observing more user behavior to match the actual use cases. Especially, you know, there's two libraries and collections UX. Those are all like user driven things. So maybe some kind of study. I don't know what those. Yeah. I mean, so those, those two are already collections of feedback that are a whole bunch of issues or things like that all tied together, and they're all linked up in, in the relevant issues there. It's a good. Yeah. I mean, I'd love to do more proactive user ethnography, right? Sitting down, watching someone do, do something. at least my hunch was for this first term we have so more than enough to to more than enough on the plate right this second to yeah but you know once something is laid down like it stays there for at least a year before it's revamped again I guess in terms of the the layout and whatnot so so that that's one but like maybe having a tighter relationship with the training group so that as trainings are delivered people observe and their feedback provided back especially for first time users I don't know just just no that's a great idea that's that's something we talked about pre-covid it's a little bit harder to do right now to sit behind someone and watch them do anything but but yeah this is supposed to be how these these groups work right is that is that the training group is supposed to be able to say okay well this is these are the problems and now who who do I tell and how do I get priority for fixing these UI problems that we you know found during training oh I can talk to the UI UX group yeah so that should actually be improved under all this and I think feedback such feedback from training groups should get these these items that they identify should go on the top of that list because of real trenches I'd also like to add that the support team well which was me to start with which is now expandable out there later provide a lot of this information and so um and I think we have a pretty good pathway of doing that already but we can discuss yeah I was just going to note that I think the batch operation stuff came from you or yeah I think so I start yeah obvious everyone yeah so I'm very excited to see that in this case training or support group they need to distill this to the point or we need this button here this is not should be general statements like you know interface sucks it should be just you need to do this here so then it will it's easier to deal with it well there is eye tracking software so you don't to do an ethnography you don't really sit next to a person who is using a system you just have eye tracking software installed on the user's computer so it just needs to be a well-designed systematic study of how galaxy is used probably a lot of work but it can be done I believe Freiburg actually did such a study a few years ago I don't know if these results are still relevant but maybe Bjorn can resurrect this because Freiburg is much more collaborative place so they're actually they're interacting with people from what is it human-computer interaction group I don't know if Bjorn wants to... Could we have something on like record your first time using galaxy thing on on twitter and post your video or send us your video and you can see what the first five people that or I don't know first however many people choose to take this challenge they have to incentivize it somehow but just see what they do as Anton said so we have connection to our social social science department and they did such analyzers we also did some questionnaires we also observed galaxy users first time using during your workshop and so on so we have these connections we have initial material but in the end I mean it's a long-term such things are long-term projects and we need to guarantee that things are in the end implemented and so on if the UX team is interested in that I'm happy to to deliver the contact and to resurrect this effort but I think we should do that with experts so there are entire fields that only do that and this is this is a scientific field so this is nothing that you can just do this is a more or less long-term project where I think it's worthwhile to to get really experts in it so the science gateways community institute in the states offers such service what they will work with you we did that for cloud launch a couple years ago what they will work with you for like three months or so and they'll write up a report and what not about for a specific app um so the outreach I mean the sort of the so Dave uh we can actually do what and is proposed to ask users to submit uh first use videos where they encounter some bad situations and I can reprint all the versions of the galaxy unfinished genome calendar I have all of that so we can send this to maybe first 25 people they can easily do that okay can you make sure it does not get forgotten try so we'll just stick it into the agenda for the next Tuesday meeting and then I'll print the calendars I have the files I would I would still suggest uh maybe at least consulting uh the experts uh who worked with Bjorn because uh posting this to Twitter or asking new users would give a very uh skewed sample so the data we get from that might be not as we're gonna do both absolutely okay before we move I have just two quick questions how is IGV JS implemented so right now it's implemented as a charts style visualization where um we we were hoping you well eventually I want it to be the prototype for sort of a next generation visualization that doesn't rely on the charts framework that is view CLI based so you'll just say like view start galaxy bits or whatever and it scaffolds it out and you have your you know that's the that's the goal not q1 probably but uh anyway uh as a first pass I implemented as a charts visualization that uses the existing framework to you know pick data sets and add tracks and things like that okay cool thanks and new user welcome does that include any notion of activities at this point no okay I don't know where activities fits in here but it seems pretty important to me still that we can try to simplify as we speak about the user experience of galaxy simplify what happens when users first hit the page and how quickly they can move to an analysis and I wonder if a new user welcome could somehow be our gateway to doing something better with activities that's a great idea so we imagine that the new years welcome would grow into a so it's a new years welcome and then eventually a an existing user dashboard kind of thing that yeah you could imagine would grow in that direction okay how close are we to having the new user page the welcome page actually be integrated into view I mean I understand it's still living in an iframe and I know that you've been working on the housing for the application in view form how close are we to actually getting rid of that that damned iframe well so well we'll keep the iframe around for displaying particular types of things but actually I just talked to Alex about this yesterday for the for the pointers to just show a regular component there switching everything over to view today yeah wonderful right now like literally right now yeah I discussed that with him last week I didn't know whether you wanted to go ahead and do that now or if you wanted to perpetuate the iframe catastrophe we'll still have an iframe for other stuff but it's the way that panel works right it dynamically switches between iframe and adiv so it's not a big deal okay hello so this is the group that used to be known as IPCR but has been more broadly labeled cancer informatics and roughly the idea is that large data all over the place needs heavy compute and so this is a bit of a crisis particularly for the way historically galaxy is operated with data in some ways and so we seek to make that easier in a variety of fashions there are the members listed and on the updated slides there are there's the meeting time on this slide as well that would be monday's 1 p.m. pacific time and you know for eastern time and as an american i don't know that europe exists so i didn't do the calculations sorry about that next slide please um yeah so the main three legs of this are that we have various forms of protection schemes that need to be navigated and i mean i think everyone's had the experience of logging into 12 different systems in order to do an analysis so anything that we can do to streamline that for users would be good we're looking at approaches similar to an anvil situation where you would have one major login site and then smaller login sites that you may or may not refresh but ultimately we want this system to work on your behalf to get out of your way as much as possible with data local compute is important in that i mean data local compute is maybe people think of that as like it's on my hard drive with my cpu but in this sense we mean it as co-cloud local in order to use the uh higher throughput uh or bandwidth rather that uh cloud transfers permit so doing the analysis where the data lives as opposed to bringing the data to some monolithic hub of processing um and then there's the idea that there are common tools common workflows common flows of analysis that we can help people bootstrap their activities with by providing them with these tools and workflows prepackaged or easily accessible next slide please so to turn that into concrete priorities we want to have gvl as the central actor here since it already comes with a variety of tools that will help us with things like authentication or uh deploying jobs to specific locations and scalability that kubernetes enables so on and so forth so we want to have a gvl up and running on aws that's configured against a fence so that we can access uh crdc data fence being a uh service from gen three that serves as a gate uh as a fence for authorization and authentication um against resources that we want and we want to have the cyclic if workflows as a first sort of pilot workflow ready to go and analyze and these are our q1 goals questions comments i want to assure our european members that europe does exist i'll take your word for it um where are you going to develop the workflows like where are they going to end up is that something we can put on docstore or the intergalactic workflow commission yeah that's a good question that we are still um working on i think we just arrived at it would be a good idea to have a repository of these and i guess naively it would be like something bundled with the instance but that's only useful for a few canonical ones and as these wealth of workflows come into play something like what you're describing would be much more useful particularly if it takes us out of the equation so that user a can make workflow user b can use and we can just sit in the background and give a thumbs up i think we want them on both docstore and i wc i don't actually understand what it would mean to put a workflow on i wc at this point but i would welcome the opportunity to learn about that process because i've been thinking more about nf core for instance which is the next low analogous process it's not made it onto the priorities but it's very close at an initial pass with testing and publishing works um so i think that can still happen in q4 actually we would be happy to be gimme pegs for the i wc in this case we have some really important workflows that would benefit from this robustness that would presumably we would get by going through the i wc process that's the aspiration for sure i've got just a commentary i'm not in any group but because you mentioned retrieving data from image resources so i'm working on images mainly and we are now working also and that's relate probably more to the first element first working group sorry we we're looking at putting images in we've already put them in object store in s3 so that would probably yes with the work done in the first working group as soon as it's available we have quite a lot of example of imaging in a file format that is s3 compatible so that to yes with this type of workflow instead of fetching them from a standard resource using the api you can could directly get them from an object store are you just using docker images no images imaging data i'm sorry not docker images imaging data so like five the images that you currently retrieve so i think that's were mentioned of omero which is where the type of data will live in public image data resources basically uh yeah that's good stuff we certainly have a lot of working group synergy amongst other groups so anything that we can borrow would be fantastic as soon as well when the the connection in the first group was connecting the with object store i think we have already imaging data in a obviously it's in specific file format but that could be a workflow to explore potentially to pass that to the user instead of transferring data via an api or getting a local copy you can basically your l and try to you fetch the data those data when you want the plane when you want them yeah definitely on demand that's also that's also going to be a goal of the back end working group this concept of mounting data so to speak onto galaxy i think is really important for imaging because imaging data sets are so large and we do want to minimize the transfer so for now i think we're starting out with omero and trying to use the really nice bridges that have been built between omero and galaxy but in the long run if we can connect directly to a bucket or if omero can point us to bucket locations rather than having to transfer that image data across that would be amazing i just put a link of of an example of file that is currently in idr stroke or meru because the same thing but that has been converting into the s3 compatible format and you can actually fetch them right now and fetch the plane you want or that you as soon as it sends or i don't know if you're familiar with familiar with that but genre i'm giving a four for it that url is that expand oh sorry yeah maybe maybe there might be something something to do but that's the type of image i can so that's the image you will see when you look at it um awesome if we uh move to iraq it also has support for s3 plugin so it can directly talk to s3 um you know and we could get that connection via iraq it's needed sorry i got some yeah we can follow up on this discussion offline but i think this is really interesting and there's tremendous growth right now and interest in image analysis um so let's see if we can circle back around at some point and figure out how we can do this well with galaxy yeah it's yeah i said it's early days for us to work on that format but yeah it's not only the image data but also if you if you have track associated to it or there will be as part of the publication there will be in the bucket that would be fantastic if there are no more questions than the tools working group all right so i am a spokesperson for the tools working group we don't really have a specified leader more so that nobody really wanted to talk um so we decided basically that every every quarter we are going to take a an area of tools that we that we identify as lacking in galaxy we don't have the best ones we don't have a lot of them and specifically um dig in on those and try to expand our tool set in that area the q1 we're going for vertebra genome project tool set uh next we don't have one specifically for q2 but we'll look into that as it comes we also need to figure out exactly how we're different uh and how we relate to the iuc because that was a little unclear uh but those are our main projects we also are discussing a a more robust method of training for for future tool wrappers um there is just so you know i asked somebody to parse existing vertebra genome project workflows so i have a list of tools and so just get in touch with me on that that would be great we have a spreadsheet that is completely empty as of yet okay all right so that's well-defined q1 backend okay so um yeah the purpose of the backend working group is to design and implement improvements um obviously concerning the galaxy backend which is sort of a central thing and of course it also involves uh interaction with other groups and other uh future projects um so we've settled on sort of uh four major areas that we want to work on um that sounds like a lot but we're also many people in the working group um though so the first priority and they're not ordered by importance uh is to monitor modernize the galaxy api um so that involves um my so there there is a board if you can see here with uh seated with some initial issues um mostly this is about um adopting a um asynchronous web framework so currently we're looking into fast api um that would give us a lot of modern tools that includes open api schemers validation using type annotations so that makes the api much more uh usable and the documentation is great um there's validation so it will be harder to um add invalid payloads um sort of related to the history pr and like future iterations of history improvements is uh history update subscriptions of a web socket so currently whenever we want to check if something happened in the history we need to poll the history every two seconds I think um currently uh that causes a lot of stress and instead we can just um asynchronously listen for new things and then send that back out something else that has been limiting what you can do with galaxies the fact that we don't have a way to do uh background tasks things that run for a long time that shouldn't the user shouldn't wait for five minutes until there's time out there should just be an indication that this is happening in the background and you can go on about what you're doing currently um as part of this transition we would need to replace uh yuvisky and somewhat relatedly uh we're going to migrate to uh we'll need to migrate to a sequel alchemy first 1.4 and then enable the 2.0 features um which um brings some level of compatibility with asynchronous python next slide and this is q1 um that is q1 it's also already ongoing yeah um so uh we've grouped this under uh data framework so um most of it is work on object stores um so if you have multiple object stores you may need multiple quarters uh we want to provide uh easy and intuitive access to scratch storage meaning uh large storage that um you cannot count on the data being preserved but like you can you can get a terabyte instead of our standard 250 for instance um both storage so uh tape storage um that's sort of the different kind of um object stores um user interface for uh summarizing the usage of the object stores and so of course that would be in a collaboration with the front end group um related to organizing um object store usage is these different retention policies that goes together with scratch storage um there is a longer term project to also enable uh user object stores so users bringing in their object stores um and then there is a uh also another big thing that this would probably span multiple quarters but that would be completely remote object stores um where you can just work with references to s3 files and you could still see them in the galaxy history but uh there wouldn't be any metadata until the job has been run on the data set so that like galaxy doesn't need to attach the data it just gets a little bit of metadata um and then uh enhancing the new galaxy files plugins um so that's already part of the 209 release but um to really take benefit of it we can we can use this galaxy files plugin so that history exports and in the future workflow invocation exports can export to these file plugins meaning that you can put them on on s3 on ftp on whatever makes sense and whatever can be exposed as a as a file plugin which unfortunately I've forgotten the name of the library we use a py file system too so she can write a py file system to plug in your files can be exported there um and then the third item is performance improvements and testing so that's uh together with the testing deployment teams um trying to make sure that uh performance of difficult and large operations is not decreasing and um ideally increasing with every galaxy release uh enabling like larger scale analysis with with galaxy in terms of number of data sets and size and overall scenarios um and one more concrete goal for q1 is optimizing the job and data set state change optimization which takes up a good fraction of the time needed to finish a job so yeah is that realistic plan for q1 um which one all of them or yes uh so the first one is something that can be done incrementally um so the new um framework can coexist with old routes we don't need to immediately uh scratch everything um and the same is true for augmenting uh the api routes this doesn't have to be done all at once but for instance for some areas where we're actively working on I think it makes sense to uh build up those routes you and especially we have to do that anyway for supporting web sockets um and I have I have a proof of principle working as an open pr um q1 seems reasonable for the second uh one the um data framework things I mean they're they're uh they're more ambitious and they are multi quarter um for sure uh but there are deliverables we can we can accomplish in the first quarter um yeah for the data yes a lot of the objects sourced up there's two big prs already open that give a lot of that functionality at least the back end portion of it um and then the two of these things are labeled as multi quarter goals because I don't think we couldn't do them and then enhancing galaxy file plugins I mean hopefully that's that's something we can deliver yeah I want to make one comment to tie things together potentially I know that there are concerns that as we work with remote data the metadata that galaxy normally relies on may not be available to us and so hopefully to potentially ease this problem or at least make ourselves more cognizant of what the problem really is we could go go back to the uiux group and as you write that tool view viewfication or take on that tool viewfication task make notes about where metadata is pulled in and used in tools in general and whether you can imagine any potential workarounds if we don't have that metadata in place I mean I think I can already say that for a lot of these things that would be trivial because we have don't have these inputs available in workflows either because the data sets don't exist yet okay that's fantastic to hear it's I mean it's it's feasible it won't be as nice but I think we have the proper workarounds already in place and I hope I there's also the like some of the privacy-oriented galaxies like anvil and the ITCR one like the metadata and stuff should not even make its way back to galaxy like the peak just something I guess to keep in mind because of the the restrictions how where data can reside yeah I mean that is indeed something that we would then also need to discuss with the front-end group I mean disabling the peak doesn't seem like a big deal but yeah there's surely other instances any other questions to Marius yeah I was gonna suggest there's a rather important kind of overarching piece missing of the puzzle galaxy galaxy isn't cloud ready and it looks like a lot of people are trying to push it into the cloud but there's several architectural problems with it that need to be addressed before it can actually be put into a cloud environment I don't know if that's higher priority than some of these other ones so for example the only so you have the application process and then the job handler processes and arguably the workflow scheduler processes none of them can be horizontally scaled or recover from a failure without being thrown throwing the entirety of galaxy into a undefined state so for example is the application process if it's deleting datasets and one of the nodes is suddenly shut down that delete operation will just simply vanish the delete operation needs to be handled the same way as the upload operation where it's scheduled as a job the workflow handlers have an issue where if a job is assigned to a workflow handler and then that handler is shut down the those jobs are orphaned and never scheduled there are there are some other issues floating around I need to go back through my notes but so I think well Marius did you want to I was gonna I think a lot of this will come naturally with the decomposition of validatable tasks right so when we have salary tasks to do sort of these longer these not within the span of a transaction components of work that's going to change a lot of this and then that's on this roadmap here yes okay perfect but the horizontal scaling like there are some race like conditions but in a general sense horizontal scaling works like yes there are some like I guess windows in which if the handler dies within this window it's bad because galaxy doesn't cover it but well you also can't shut it down completely until it has no more jobs that it's handling like a handler is responsible for a set of jobs and no other handler can be can take those over I assume is what no one's talking about yeah I mean these are relatively minor improvements that we can do I don't think they necessarily have to be on the roadmap yeah I think that's I mean that isn't known we have an issue for that somewhere but and it's something I think that we've been moving towards especially with developments with the message messaging and all that kind of stuff I mean that there are two things so you know we can address the urgent need there and the longer term thing is building up message queue capabilities that would be more robust and maybe pushing some of the job handling also towards message queue which they are made for doing for doing that so yeah but that's not really Q1 we want to start with a smaller project for the message queue and then uh extend that Nolan are you in any of the working groups you want to join I I don't know if I have the time I I basically invest as much as I can based on the projects I'm working on but I don't actually have funding to work on galaxy directly thank you for all the feedback regardless yeah if we don't say that enough yeah I'm getting close to a hundred open issues now I think I'm at 89 right now okay to get a t-shirt at all so do we just need a Nolan priority there those issues or a working group yes okay so anvil I believe is going to be called human genetics working group yeah if you escape and then present again they'll be an upgrade yes yeah oh it's the main purpose of this working group is to I guess consolidate all the work from other pieces and especially the galaxy and committee's deployment and specifically purpose it for anvil on gke and all the integration points with terra and gently and other anvil components so the kind of like goals that I've written here are not set in stone we're meeting tomorrow with the brode team so I guess a little bit of background the anvil project is collaborative with the brode institute in Boston who is the institute that develops terra and then also a lot of other teams um but the deployment of galaxies handled by Leonardo which is a service that the brode has as part of the stack um so a lot of our work here also depends on them and coordinating with them so I put a bunch of goals if you go to the next slide I put a bunch of goals that are more like independent than the brode that we can mostly do on our own and that also touch up on some things that the other groups talked about um but we're also meeting tomorrow with the brode to talk about all the things that we need to do in terms of the actual deployment in leo um so high level priority is kind of the from the tool's perspective the idea is to have a transparent list of what tools come to the anvil by default in the anvil scenario everybody's an admin on their own instance so everybody can install tools from the tool shed and we can we're not taking on the guarantee that all the tools in the tool shed will always work in our versions but we do want kind of a guarantee that the tools that come by default in the panel are working so in order to do that we're gonna start from a list of working tools and then maintain a testing infrastructure so that the list is continuously tested and then document how we add things to those tools or whether we want to add it or we want if a researcher wants to request a tool to be added how to go through that process um in terms of integration this is the anvil fs plugin which uses the platform system to plug in um and that's working for ingesting data from terra um but we still need to do it the other way to push data back to the terra buckets um and then there's other integrations like the doxel workflow um which are working but might need some optimization um but i'm not sure if that's going to be q1 or not um then in terms of interactive tools and that's kind of we're gonna with the deployment group kind of get interactive tools working on Kubernetes more generally and then for the anvil specifically since it's on gke and there's the identity that's propagated from um terra and the fact that leo creates the proxy so we don't have control over the endpoint or it's going to be slightly different uh so after we get it working in a generic Kubernetes sense we're going to need to do a few more tweaks to be able to work with the proxy and all the integrations that and the identity from terra um and in terms of deployment um we need to continue maintaining the home chart that aggregates the galaxy home chart along with the dependencies and as the leases come up um like um update the chart to have the new galaxy also 21 a 21 or one coming soon we're going to have to update the chart to work with that um and then uh on top of the testing ecosystem for tools we also wanted to continuously know that and the galaxy is still deployable and ample so uh we haven't talked about like how often we would do this but i'm thinking maybe like once daily make sure an instance can still be deployed it can still run jobs and basically have a status knowing that if somebody's going to anvil today that it's working and not have to get a message from someone that nothing's working before we deal with it um and then in terms of launching so like we've optimized galaxy started up a lot and um there's like a lot on the back end group i guess uh that did all the caching um and so galaxy starts up in 30 seconds but before that there's like the about minute and a half two minutes of um all the other like the nfs cv mfs all of that coming up but even before that does seven minutes of the cluster coming up and that seven minutes is something that is almost entirely on the both side but we do have ideas on how to optimize it so we need to work with them to make galaxy come up faster and hopefully the baseline would be probably five minutes with everything um and then that's more on the technical side and then for the more science side of the group just in general developing more workflows to actually use a huge amount of data that the anvil makes available questions comments thank you that's very detailed is it on kubernetes in anvil like a quarter one goal specifically or ideally yeah so i'm hoping we do it since it's a q1 goal for the department working group that like say halfway through the quarter we can get working on kubernetes in general and then the last step would just be like a little bit of in-depth configuration and stuff for the anvil so hopefully should be doable yeah any other questions yeah well i guess i the it's have made their appearance at least three times i'm thinking these slides um but i don't think i actually noticed them on the ui ux team and apparently the ux absolutely sucks um with it's um for especially a newcomer so hey there are open issues but the point of us picking roadmap priorities was to pick a few particular things right it's not to say that other stuff won't happen um well there's an issue for it i think it is probably true though that if one of if the goals of the number of the other teams all right let's have these it's in production we are gonna have to get some of those u ux issues sorted out so i think well so i think at least part of it the the display was hopefully uh comes along with the the so the new history will will have a componentized view of the data set and we can have a little bit more flexibility in how we show the items in history and things like that and that'll resolve that particular part of it but yeah there's there is a master issue that shows uh i can link it up let me find it well it's uh i think you know once we have these priorities nailed down we need to absolutely should be like a graph representation of all the working groups with their priorities so we'll see how they should interact and that's yeah we need interface for it is i mean i icon should work at the minimum that's the one that was linked to the to the new history i think okay i think we're already way past the hour but we started late so uh support again are there any more questions on this so just make this quick uh this group is new because it used to be just one person so we are working on identifying people from different groups and um the support sort of touches everything else we kind of have to know what's going on with everything else sort of to address how users live questions so for end user advocates um provide scientific uses training support at various forums and chats and mailing rooms um and an important part of this is triaging those problems so if people report something is it a usage problem can we improve some documentation is it paper cut so we can improve the ux is it something that just needs more documentation is it something that needs to pull other resources together into a story so people can understand it and normally right now that happens um in q and a and instead we want to flip that so that those those resources are grouped together um and what we have right now is FAQs and instead access that resource and then help the user connect with their experiences with existing resources because the ones they have trouble making that connection so right now it's uh me, Pristo, Dave C. from Outreach is involved and he will cover more about what some of the immediate stuff is about and then Ignacio was nominated late last night early this morning so um and then we have a getter chat people are welcome to join that that is not public because we do not want people to be asking questions there so it's not a q and a place it's to discuss how to give support um and so our priorities are onboarding finding out who's going to be involved um we have been having regular meetings but they've been just a few people we're going to uh do uh larger ones and that will include some gtn people and uh the the video stuff that kind of thing um there's there's just sort of like a bigger overview inside the chat right now people want to kind of get an idea of what this involves i think that's too detailed for this particular meeting there will be more in drive soon we're working that out um part of this is the priority three is that we need to decide where we have these existing FAQs right which are not probably in the right home they're not consumable uh as assets that we could say link into the interface right a little snippet of how things work and and and then links to other resources um so we need to decide how we're going to do that and where we're going to do that and that's going to involve some work with the core development teams to make sure these things are consumable and i know john had mentioned that that was something he was interested in but the way that the way that these are now are too centric about uh use galaxy.org um they are not automatically linked to anything they are maintained personally i'm probably the only person who understands the content of all of it it's everything the other one using it um and so all of that kind of needs to be redesigned a little bit and then reorganized standardized formats um a way to update it and a way to integrate these into other things and and how much of that could be can start to be migrated needs to be decided and it's going to decide and we need to get the first three done first and then we can think about where the rest now i know that um when Dave talks he's going to talk about how we want to turn some FAQs into tutorials in the GTN which is great for some set so that may be the first thing that we do um as as a translation but we also need to this people have questions i don't want to know how to do analysis x right and this this can link to multiple resources to help me do something or i problem x there's multiple resources and so those need to be summarized together and then also broken out into smaller bits wherever they wherever they should live and they should be uh version controlled they should be linked to do uh releases for new features you make sure we cover those in a timer way um and i think i'm going to stop there so if you're interested in what the bigger picture is right now you can go look at the chat and um like i said this is all like brand new we've been just having and we have to remember people who are working at support every day are dealing with users as 90 percent of what they're doing right now but that won't get easier until we start to make some of these the other things better especially now that it's more than one person doing it so um we want to migrate resources into something that is expandable and forward thinking in a smart way and do it once so right now they're gonna look at it we probably want to represent them from australia on that group too yes we do and so uh simon if you we can't see the speaker notes but the speaker notes say yes simon has been paying to we're working on getting finding out who that would be so yeah that's part of it questions well just i guess for the future when itcr and and for start getting more activity we're probably also going to need support for users there um i agree so yes so i i think that the scope of what these support assets are should increase as well so it's not just and that that's where the update is and we can discuss how to branch that out and there's going to be some kind of ontology to help things get organized and um and then a relational um aspect so that things are just it's not a linear branch right there's a lot of cross connections to different assets and the ultimate goal is to be able to provide some of this directly inside of galaxy itself small blurb a little question mark it needs to i think that having another website is not such a good idea um maybe we have a repository under galaxy called support materials and then those get published out to the galaxy help forum but they can also be consumed directly on the application and so how that gets organized you know come at it um like everyone's feedback about so it's not going to happen in q1 but yes ultimately anvil and atcr will okay any questions all right thank you jen and uh how many do we have okay so we have training oh sorry we have okay we have testing and hardening uh testing and hardening um those are the people there's our gitter channel we decided on first Monday of the month we're gonna have meetings i forget the time probably like noon eastern uh we're gonna on the second monday of each month we're gonna do like a testing hackathon code fest we're gonna start by focusing on like transiently failing tests so i've started sort of tracking those in the github and for galaxy um just a place to start and get people hacking on things and reduce the single to noise ratio just make the test more useful so probably something that would be documentation also um and then along with this like sort of focus on tests we we're gonna sort of stick with the release uh testing that we did in 2009 or that surgei sort of led up surgei maria said and sort of broaden the scope and cover more usability testing in 2101 um and so surgei is putting together some ideas um related to uh like a a rotation of people to sort of um sort of help out with release testing so maybe like everyone could help out once a year and we could have four teams of people and then the teams can also have specific uh you know uh deep dives they do for for a release so like maybe the admin team once a year the admin team can just walk through the admin ui and make sure everything there is really solid with that release uh remember the next slide and so our quarter one priorities um something that came up this last release we wanted to do performance testing we didn't quite get there uh we had some issues around this and so uh we roped alex m into sort of helping us um set up some infrastructure um so we've broken it into like three big parts and we'll get and hopefully in quarter one we'll get the pieces set up and then in quarter two we'll start sort of integrating them at scale but the pieces include being able to launch jet stream instances to preview galaxy from a from a galaxy pr and sort of test and play with the release um we're going to get some performance tests in things that that that you know use galaxy's metric collection and and run some big jobs um and then we'll sort of write the pieces to connect those things and run those jobs um and so that's sort of getting that that performance testing infrastructure in places the a big quarter one goal uh the other big thing we have on the plate is sort of selenium and publishing and production instances so olig is sort of spearheading this work around um getting screenshots uh generated for training materials running our integration tests against production instances annotating the test better so you know what tests can run where and what they require um and hopefully these will be like tools that we can give production instances um um uh yeah and that the sort of hopefully these are like very specific uh small quarter one goals that we can uh get done these seem fairly reasonable um any questions about any of that okay awesome so uh i'm going to talk for the attrition training group uh which includes all these wonderful people this is on the right uh so we have decided to try to begin with uh to talk on discord instead of gitter uh that for some reasons uh one of these is you don't need to have a gitter account a gitter account uh to connect to discord uh we can have an unlimited message history for free uh which are to find uh some document that has been leaked before uh we are trying to install some rules as well so we can ping people related to one project without pinging the whole group um from the two time and one of the big advantage is also that we have uh voice channel where we can pop no meeting without sending uh invitation and everyone's connected can join the meeting at any time um talking about meeting we're going to try to meet once a month so we're going to have two meeting um uh one day per month one at uh I think eight or nine a.m uh eastern time and one at five p.m eastern time so we can have uh everyone in australian europe as well uh we will record the meeting um so people who can't attend can see what's happening uh next slide please so uh part of the priority we have uh in the next three months for q1 is uh transfer some some of the faq into uh training material uh like what do we do if you have a red set or this kind of style uh since gen can't possibly do that all alone uh we're going to try to do that during either co-fest or some extra hackathon uh to transfer all of this uh we have nick working on community website package uh to replace the hub uh that can be modular and deployed for different use galaxy dot star uh so the goal is to have it deployed in q2 but having some at least something to show up for q1 uh I think so gay is working on deploying uh tis on uh all the use galaxy servers uh so we have some training video that are out on the youtube channel for q1 the goal is to have them uh linked in the gtn website and uh also start talking about the automatic screenshot integration in gtn with I think Oleg is working on uh that uh for q1 there's a galaxy for higher education that has that is being tried uh in couple universities in europe uh so maybe not for q1 but they will apply for ects once they have some first results with that what is doubt of higher education and what is ects ects are European credit for degrees okay uh so they uh I don't have a lot of detail on that but they have been starting to try some material uh with some teacher uh see how it goes next slide please uh so for the three months next three months we're also going to start in storing an extra call uh for the co-fest so we're going to now have three uh call one with eu uh with berenice and sascha leading leading the conversation in the u.s uh jen and I and in australias Simon and Anna uh so the gtn co-fest will be happening in the same time as paper cuts uh I think right now we are not timing off every three months but we're thinking about moving to every two months for the co-fest we also want to encourage the participation to the open life science project which is organized in part with berenice which pairs mental and new projects to help them navigate through open science using github and open resources uh so uh Dave Beatrice and I have been uh mentor and participants for OLS2 and we are thinking about encouraging more people of the community to join in that uh the gtn paper has been uh in the machine for a bit uh but we hope to be able to send it before the end of january uh there's a deaf training happening uh sometime in 2021 and uh it will be discussed more at the january round table um and I think that's oh no there's an extra slide uh we also like uh to organize some uh hybrid meetings with the working group because uh as gen we're saying uh we have some topic that we need to discuss notably with the yx group like uh linking the tutorial the help and the videos uh to to the interface for example and also uh we have a lot of feedback from users on how they uh leave the the experience not most notably for storage management uh we want to develop some training for how to manage your space efficiently and we're thinking about how to make it more gamey for people to be interested in cleaning the space thank you okay is that all right so this is the last so I think this is it so just some before we wrap this up just a few observations says I don't think all groups have gitter channels it doesn't have to be gitter channel but there's to be some channel where this communication is possible and we need to once we're going to be preparing um page on the hub which lists all working groups that should all be there because otherwise I don't know how are you guys going to communicate um and again um we so the PIs we're going to take a whoever's typing please mute uh so the PIs we're going to look at these goals that you have in these slides but we're going to do this next Tuesday so if you guys can look at them again before next Tuesday and A again check if they're reasonable for Q1 and that they're um sort of a spec that they're specced out nicely uh and obviously we we should avoid one person being head of three groups as Nate does so you guys need to rethink that in the first three groups and um we need we need to have point people for each groups we have that and we also need to create some communication channel with them so we can communicate directly with uh with these representatives any other notes Anton we have a few new users in the beginning can you go to slide two maybe one second and just maybe sum up what these working groups are for the new people okay just a second I lost my which slide slide two you need to refresh probably uh yeah do you see that Bjorn yep okay Bjorn would you like to talk through this slide okay or I can it's up to you I wrote this and please do it Nate please uh please do it Jeremy sorry better Nate's way more savvy than I am um but I'll do my best so as a P as a set of PIs we talked about working groups over the course of a couple weeks and decided it was important to try to create some structure because the galaxy community has grown so big so quickly and it's become incredibly diverse as well and so in order to put some structure around that and in particular to hopefully make everybody who's actually doing the work and has hands on day to day happier more productive more autonomous and provide opportunities for growth these working groups are meant to ideally improve not only the function of the entire team but also your happiness as individual developers admins trainers and so on so certainly please let us know if that's not the case but that really is the goal here it's to give you as part of a working group someone to talk to to bounce ideas off of in a small group setting so you don't feel like you're all alone and it's to provide the opportunity to build something that you couldn't do on your own by gathering folks who are like-minded so by having these groups which to be fair existed informally throughout years now we've kind of formalized those so there have always been groups that have been interested in deployment or the back end now it's just a little more formal and so for work group organization we've created these work groups around big topics for now we think these are good working groups if there are others that we missed and I know we talked about some of those today we can certainly revisit that and try to manage them what we want is to try to come up with a sweet spot where we have enough working groups to cover all the major work work points or work topics but don't get to the point where we're stressing people by having to participate in too many work groups and so we don't want to create a work group unless we can really have folks that are going to dedicate themselves to being able to contribute and pushing the work group forward it is absolutely okay to say that we can't get work done in a certain quarter or that this this work is important but it's got to go into a priority queue at some point and so perhaps instead of creating a working group we take some of this important work that has fallen out and move it into another group and list it as a priority that we will get to and certainly concretely the goals here as I kind of articulated at a high level are to promote communication and collaboration around joint projects to provide entry points for new members I know there are a lot of new folks on these calls and we're getting new folks every day these entry points I already saw for instance that someone asked about potentially contributing to the back end and they were pointed to the getter back end channel and that's fantastic and hopefully they'll find some traction there so provide entry points into the community is an important thing that we're doing here improve the communication between the experts that we do have across our community I broke group but it really is community and then clear communication points as well across these groups ideally so that we can get some some improved throughput so rather than having everything quite so distributed streamline it just a little bit so that we get important people talking to each other when we have these features such as interactive tools that cross cut all these different places did I cover everything during perfect thanks any other concerns please I mean if there are no concerns that's fine too but just if you think that they really don't get that but I just don't want to say it so I had one because I mean something that you really stand on which is the scope of a working group I felt that there is this issue that some of the some of the tasks that each working group has taken on might be maybe too large for a quarter and maybe to make sense to have one like a reasonably you know reasonably large task that's accomplished or the working groups tries to accomplish within a quarter which you know that is their main task somehow they will achieve that task and thus it keeps the ball moving forward and then to address the concern that Nolan raised that there are some bugs and stuff that might be left behind maybe you could also include like one single overarching task like a large task and a bunch of say 10 bugs or something or five bugs at the working group will try to fix within a quarter I guess what in part we're trying to solve the issue of small bugs we have paper cuts but but there might be bugs specific to expertise of a particular group yes yeah I don't know if we need a better way for so to should we have like a leader board of things that people are concerned about before we do the whole you know working group meeting and setting the priorities that way so for for an example the UI UX working group would know okay everyone really wants you to focus on it UI UX for for this this quarter it's a blocker before we sit down and make priorities or I don't know how to better organize that I guess I think there are two concerns we're trying to address here number one is whether we're too ambitious in our goals and my suggestion is that we let the working groups decide and sketch out what they want to do in q1 and if it if it turns out that they're way too ambitious then they're going to scale back so this is part of the autonomy and trust in your teammates is let's let's trust that everybody can scope something out and if not they'll adjust the second question of how we prioritize work is a really interesting one and to be clear it's one that the PI's continue to discuss and it's brought up some really nice points about how we can try to incorporate user feedback into this loop and I don't think we have the answers at all right now suggestions about how to gather feedback from the entire community which is worldwide at this point and and prioritize that work in addition to the work that we've agreed upon for our grants is kind of what we're balancing right now to be really clear about from the PI level that's what we think about is we want to serve the community we have some deliverables as part of our grants and we and how do we merge those two together on the topic of scope for a quarter I think that suggestion of having you know roughly 60 percent of your effort sort of assigns these working groups if you break that down if you're a member of two working groups that means one and a half days per week roughly contributing to the work for the working group so if you you know multiply that out roughly 12 weeks in a quarter means 12 days maybe 15 days of work so what can you accomplish in 15 days of dedicated full-time work is what basically may help inform what gets put on the individual tasks for a given quarter for a given working group this might be a dumb question but does the galaxy project have a project manager not really huh because one of the things that I always had the impression of is that there were like these larger internal meetings that were deciding direction and kind of just normal like traditional project management stuff and I always wanted to like have a peek at the Gantt charts and what it's not a traditional project in many ways and it's been operating that way for like 12 years already so we sort of never felt the need because it was it used to be fairly horizontal in the way it worked and so this is probably the first time that we feel that need and this is why we're trying but we still don't want to you know one of the Jeremy mentioned that word autonomy so we the goal of working groups is sort of make these decisions as as they're seen by developers we don't want to impose a lot of these things yes of course we have a set of priorities we have a goals for the grants but we don't want to make this traditional top heavy sort of a paradigm I was thinking more uh taking them taking on a role where they facilitate the project planning and management for each of the working groups uh rather than dictate just to take the burden off the developers because I'm sure they have their own tasks that they absolutely this is what the PIs historically have done so I think I guess program managers here are the PIs and that's that's how it's been I don't know if it's good or bad I'm just telling you how it is okay uh it's been one and a half hours I mean we're the one and a half I have one thing very briefly um I just I just like to volunteer something here uh I I've been doing a little experiment with Alex uh uh with some of the tasks that he's been assigned to view which is the front end framework that we're using um I'd like to just volunteer to mentor anybody who's trying out front end tech for the first time and isn't sure exactly what to do I am more than happy just ping me I will I will help you do your PR you know what I mean like uh and I I don't know that this is done frequently in Galaxy but this is done frequently in software development everywhere else um I get I am looking at a lot of our work product and wishing that I had done this sooner you know I am wishing that I had teamed up with people when they tried to make their first PR in the front end arena um there are definitely some basic concepts I think that have been missed and we're discussing that in the UI UX channel too a little bit but I am more than happy to help anybody who is struggling to make their PR in a technology that they are unfamiliar with um to add on to that I think it would be awesome if uh you you know if you feel you have expertise that you can review um pull requests you may not have a binding vote but it would certainly be appreciated well I was going to say to add to that as well now that we have more formalized working groups we can we we can have you know okay the UI UX team has to look at this or the backend team needs to look at this and not just can look at it but can also be a resource when you when you want to implement something you know who to talk to and where to get help and I think that's a it's a good move in the right direction and I'm always up for mentoring people with tool wrapping as I've done before I would just much rather spend some more time up front than like for instance I'm having to to to fix some things now which is natural when you get people working on tech for the first time uh if we could reduce that churn that would be great and if it means pinging me and bothering me and having you show you how to do something do it please just bang on my door any other ideas how we can improve all of this I just like to say that this is the first time that you're doing this right so we can't expect that everybody's scope is going to be perfect like it's kind of human nature or at least work culture to be overly ambitious so um it'll even out it's like an iterative process you got to let it coast so again take a look at your priorities see if you can either scale them down or expand them a little bit more and then we will review them and then we'll go through another iteration of that and eventually we're going to have a solid set of goals for the next quarter what when is that like when is the how long do we have to refine these for you on tuesday okay thanks yes this is the next pi meeting is on tuesday thank you everybody thank you for preparing for this we'll definitely be in touch uh john please uh send us a link once the recording once you receive the recording since you're the host we'll do all right all right thank you everyone I will save everyone thank you