yeah, I think this is just a place where it’s hard to have precise terminology across languages, in particular what ‘static analysis’ actually means here. Personally, I wouldn’t call what the language server does “static analysis”, I’d just call it “text-based guessing”, and what I’d call tooling for static analysis in julia would look a lot more like what’s present in smalltalk / common lisp.
Our existing tools for static analysis like JET, really are dynamic and depend on introspection of the real language runtime. We call them static because within a fixed world-age, they are doing basically the same thing as what a statically typed language does. But yes, IMO these tools absolutely need to be aware of the actual program state, need to be able to expand macros, and understand the real method-table, etc.
My understanding of what Keno said is that there is no (significant amount of) money laying around waiting to be distributed by someone. The core developers prioritize their work according to the needs of the companies that employ them. The community does not decide what private companies do with their money (THANKFULLY). So while I agree that budget assignment is a function of priorities, there is no public/community budget to assign in the first place.
Very nice interface in this package! Tracing is one of the quite common demos that people show with compiler overlays, but your approach to filtering the trace is quite nice and easy to use.
Yeah, it seems like there is no one, anywhere, who is at liberty to allocate additional resources–monetary or developer time–to the debugger or IDE, or any other area that the larger Julia community cares about. Except by donating personal time voluntarily. Everyone is locked in by personal or business needs.
It’s a bit of a “tragedy of the commons”, except, of course, lots of people are contributing to the community. But they just can’t or won’t contribute according to what The Community deems important.
I couldn’t agree more. I think this is definitely worth a more detailed conversation, probably warranting another thread. As I said further up, that lost of folks is many years old (is Logan even still involved?). But the fact that it hasn’t been updated in a while makes me wonder what the role of that group even is.
We have other committees (around things like juliacon) that I think are more diverse, and there are other (unaffiliated?) organizations like Julia Gender Inclusive.
I kinda know what the role of the Stewards is, but as mentioned above it’s nothing to do with money / resources. I don’t really know what is involved in “governance”, my guess is it’s pretty ad hoc - every juliacon there’s a call for volunteers, and people raise their hands and do what they can do. Is there anything more formal?
I believe that working on issues related to scaling Julia applications up aligns with a sizable portion of the communities interest.
While a good IDE is also important for managing larger code bases, what we currently have is probably good enough for the very experienced Julia folks who currently happen to work at JuliaHub etc. Out of curiosity, is the most common IDE used in those companies also VS Code and what is the experience on said very large code bases?
As Julia is used in more and more companies, invariably it is picked up by teams that work with a diverse set of technologies. If you only work with Julia 10% of your time and need to interact with a large codebase, you feel the shortcomings of the IDE much more acutely. This is exacerbated by the fact that a lot of interoperability stuff does not exist yet or has bitrotted, so a lot of manual cross-language work is often required.
I really look forward to the changes in the compiler toolchain mentioned by others that will make it easier to introspect the code.
Stewards is largely an appeals body for resolving disputes in the community and enforcing community standards. The Stewards do not claim to have any other influence on the project.
The major governance body is the JuliaCon committee, which is directly related to the project’s finances and has an open recruitment process. In fact, getting involved through JuliaCon is a great way to take on a larger role in the community (outside of code).
However, as has been mentioned elsewhere - there is no formal committee in the community or at NF that allocates funds towards Julia development - outside of GSOC / JSOC / Community manager / CI spend.
At a technical level, there is the triage call that happens every other week. That is another way to get involved in core Julia itself. Of course, triage does not direct resources or allocate projects, but participating and helping with triage is a good way to gain an understanding of how the project actually gets things done.
While this conversation has generated all kinds of very good (and a lot of tangential) discussion, let’s make this concrete.
Do we have any devs who have a clear vision on where the VS Code extension needs to go, and a proposal for a budget they would need to make it happen? If we have a capable dev and a reasonable budget, that would be a great starting point to find the funds to make it happen.
While I do not have a clear vision on this question yet, I would be happy to organize a Zoom call to discuss this topic with all possibly interested people. So if you like this idea and want to participate, then reply to this post. My idea would be to gather high level requirements that could possibly be implemented within 6-12 months. My personal preference would be:
less bugs
not more features
clear documentation of the different components and the way they work together
I suggest emailing juliacon@julialang.org. I believe that this is roughly around the time when the JuliaCon committee will start its work in the earnest. In the past calls for participating in the committee were made (I forget if it was only at JuliaCon, or also on discourse). @ranjan and @rayegun will probably make an announcement soon.
The JuliaCon 2024 committee is listed on the website. That will probably be a good staring point for the 2025 committee.
This is definitely a bit outdated and needs refreshing.
The signatories have no role beyond the original signing at NumFocus. I believe John Myles White was also one of the signatories at that time. @aviks has managed project finances at NF for a long time, even before he was with JuliaHub - and this is a volunteer role. @mkitti has been helping out greatly as well. As @Keno pointed out, the majority of the finances are around JuliaCon receipts and expenditures.
@StefanKarpinski is no longer on the advisory council. There was a request to vacate the spot at one point and we were happy to. @logankilpatrick is a board member, but not in his role as a community manager (which he stopped being a couple of years ago). I am sure Logan has Julia project interests close to his heart, but he is not a representative of the Julia community on the NumFocus board.
Some folks also may think that there is a large sum of money at NF that can be directed to various Julia projects. On the contrary, we pay NF management fees (10-15%) from our collections (largely JuliaCon) to be an NF project and to have NF be a fiscal sponsor. This does make us eligible to receive NF’s small project grants, which in the past have gone to a few people (and should be listed on the NF site).
Does it make sense to create an old-fashioned mailing list where conversations are kept open and widely accessible on the internet for future reference?
I am happy to learn that the Stewards list needs an update and that the JuliaCon committee plays the most important role.
The juliacon@julialang.org mailing list is not for project discussions, but rather for people to contact the JuliaCon committee. I don’t think it makes sense for conference discussions to all happen in the open. However, that is a decision for the committee. That said anyone interested in participating will be welcomed.
I don’t think I said the stewards list needs an update - I said the governance page needs an update to reflect the latest. The work of the stewards is of a very sensitive nature, and it is the only group which is not open for volunteering (at least for now).
The most immediate way to make this happen is through GSOC. The major bottleneck in GSOC is to have a mentor. Thus, we need one of the core devs to mentor a student (which means updating the GSOC project description page on julialang.org to start with).
The mentors do not necessarily need to be @davidanthoff or @pfitzseb (which would be great), but at least someone they would trust.
While I understand your suggestion, I still would like to suggest to have a Zoom call first to discuss and possibly define the high level requirements of the community. As a second step I would suggest to think about how to make it happen.
Is the JuliaCon committee the independent (cross-company, cross-country, …) committee that has the authority to address issues related to funding, human resources, networking, local community building, etc?
Say I need to raise an issue about VSCode like this present issue, or need to organize a local meetup in Brazil. Can the JuliaCon committee make decisions about these matters without previous approval of the Stewards, JuliaHub, or any other organization?
No - The JuliaCon committee is focussed on JuliaCon, which is where the bulk of the collections and expenses happen (and hence is the most important financial body). Discussions around JuliaCon Local do happen in the committee.
The community section of JuliaLang.org is where meetups and community things are listed. If you want to set up a meetup in Brazil - just send a PR there.
The groups - stewards, JuliaHub, MIT, NF, or any others - do not drive any of the topics you brought up. You do not need anyone’s authority to do any of the things you mention. You are welcome to leverage the julialang.org resources to list your meetup, your working groups, put meetings on the shared calendar, etc. You just need to make a PR to julialang.org.
Any new ideas and improvements are already discussed here.