- Boost - lists.boost.org

A history of the Boost Foundation
by Zach Laine 11 Sep '24

11 Sep '24

Dave Abrahams and Beman Dawes put up the initial money for BoostCon (which has since been renamed to C++Now) in 2007 (or maybe 2006?). This was something of a risk, since it was not certain how many people would show up. It turns out that the conference covered those costs, but they were still personally responsible for the conference if they wanted to put it on again the next year, or the year after. Beman had the idea that there should be an independent legal entity for this, rather than two individual people. I forget the exact chronology, but pretty early on in the conference history, we ended up with The Boost Steering Committee, under the Software Freedom Conservancy (SFC). That is, we called it "The Boost Steering Committee", but there was no such legal entity. The legal entity that represented the BoostCon conference and Boost in general was SFC. In terms of legal entities, "The Boost Project", or "The Boost Steering Committee" was just a single division within SFC. This is something they reminded us of from time to time. At the time, Boost had no real expenses. The "wowbagger" server that we use to host the (now, legacy) website, and on which the releases are built, was hosted at Indiana University. The downloads of the build Boost releases were hosted at Sourceforge (I think; anyway, wherever it was, it was free). They later moved to other places, ending up finally at JFrog. They're no longer hosted for free -- see below. BoostCon made more money than it spent for a lot of years. It accumulated some reserves, and people would sometimes ruminate on what we might spend all that sweet, sweet cash on. No large expenditures were ever taken on though. The SFC started very small. We were project #3 I think (maybe #2? something like that). Over a few years, it took on more and more projects, until Boost was one of 20 or 25 SFC projects (again, I don't remember the exact numbers). SFC had a small staff then, and probably still does. As they grew, it became increasingly difficult to get anything done. They were overworked and understaffed, so this is not some failure on the part of the SFC so much as a logistical inevitability. One problem we had was getting them to pay vendors for services used by the conference. IOW, we would use some service, the vendor would send us a bill, and then we would send SFC like a thousand emails trying to get them to pay the bill. It was a nightmare to deal with. Jon Kalb, who was the conference chair at the time, asked the Steering Committee if we'd be open to breaking free of SFC, and forming our own non-profit organization -- a separate legal entity. Everyone agreed, Jon did a bunch of research and paperwork (thanks again, Jon!) and we formed a new 501(c)(3) US corporation called The Boost Foundation. This is why the Steering-Committee-under-SFC became the Boost Foundation Board. This is also why the name changed -- we needed a new name for the new legal entity. There was no secret agenda; there was no smoke-filled room. There was, however, pragmatism. Sometime around then -- once more, I don't remember the exact timeline -- we started losing all our free stuff. Someone at Indiana University realized that randos (that's us, Boost) were running a whole server in their lab, and asked us to remove it. So then we had to pay for hosting wowbagger. Some time later, the download hosts stopped giving away bandwidth for free as well. No problem, right? The conference makes money, Boost now needs money, so let's use the conference money for the Boost expenses that now have to be paid. (Note that this is something the SC/Board took on as it came up; it was not part of the original mission of the SC/Board, because there were no significant Boost infrastructure expenses previously.) The problem is that the conference makes less and less every year, due to a combination of changes in the way the venue charges (it all went up a lot), inflation, the expenses from the covid-cancelled year, etc. During all this, we kept things running as best we could. Eventually, the downloads became too expensive, and the C++ Alliance offered to pay for those. So now they do. The Foundation only pays for hosting wowbagger right now. Hopefully that sets the stage. To me, the important take-away is that the Foundation has tried to "keep the lights on" as I like to say, without bothering anyone on the list about it. The Board does more than just Boost infrastructure stuff, such as organizing the C++Now conference, and more recently, paying the hosting fees for the Beman Discourse server. Sorry this is so long. :) I may have some of these facts slightly wrong. Like, did BoostCon actually start in 2006? I couldn't figure out definitively which year it was, even though I was there. But I'm much more certain about the major events, and the reasons for things. I also left out most of the non-asset-stewardship aspects of the Steering Committee/Foundation, since they are mostly not germane to the current review. I nevertheless refer to one such aspect below. In light of the history above, here are some points I'd like to make regarding things I've seen said on this list about the Foundation: 1) The Board does not impose its will on the developers on this list. It does sometimes make recommendations or official "calls" of some kind, like saying we should use CMake. It doesn't actually have the ability, as the Boost Foundation Board, to enact such changes. This is not a matter of restraint. It simply can't make concrete changes; only the Boost developers can. Also, there never has been, and doesn't appear to me that there ever will be, any appetite for meddling in Boost development, or forcing developers to do things. 2) The Board is self-selected, in the sense that the Board nominates and votes on Board seats. Multiple Boost authors are on the Board: Peter Dimov, Glen Fernades, Jeff Garland, and me. There are others who are only there for the conference business. There are others, who joined more recently, interested only in Beman. 3) I've read on this list about the desire to "get back" to the Boost Steering Committee way of doing things. This does not make sense. There is no difference between the Boost Steering Committee and the Boost Foundation, except for the name. Sure, there is a change in membership over time. That would have happened with or without the name change. Perhaps this is a desire not to have the current Board membership? If so, please get involved. Asking a volunteer project to do things differently, without volunteering to help yourself, doesn't usually work out. Which brings me to: 4) The Board is a volunteer project, like Boost is. As such, people work mostly on what they care about, and don't really do much for things they don't care about. So when I read, "The Board is not communicating enough," I think there are a couple of things to consider. 4a) Do we expect volunteers to do the stuff they want to work on, and then communicate that they did that to someone else that was not working on that same thing? Do we do that with our Boost libraries? I might communicate changes, including bug fixes, in release notes. I do this because it affects others. But if I make a non-functional change, who do I tell, and why? Note that, when there has been a loss of service, someone has mentioned it on the list right away, as far as I know. Do we also want people to mention when the Board took some action that kept a loss of service from happening? Moreover, I'm not sure what transparency is lacking, given that: 4b) The board publishes minutes of every meeting, and you can come to the meetings if you're interested. Except for occasionally needing to handle a sensitive topic that we do not minute due to its sensitivity, these are not closed meetings. For example, our latest Board member was asked to join the Board because he kept showing up to these open meetings out of his own interest, and would take on tasks from time to time. As you all know, this is how pretty much all open source/volunteer projects work. In short, I think if a Boost user wants to know the details of how Boost works, they subscribe to the Boost mailing list. If a Boost developer wants to know the details of how the Boost Foundation works, they read the minutes and/or come to the meetings. Zach

5 6

Re: [boost] A history of the Boost Foundation
by Zach Laine 11 Sep '24

11 Sep '24

On Wed, Sep 11, 2024 at 2:41 PM Vladimir Prus <vladimir.prus(a)gmail.com> wrote: > > >> > > 1) The Board does not impose its will on the developers on this list. >> > > It does sometimes make recommendations or official "calls" of some >> > > kind, like saying we should use CMake >> > > >> > >> > So the email about cutting ties with the C++ Alliance, which started all >> > this, was just non-binding minority report? >> >> I don't get it. Isn't this the thing the developers are voting on >> right now? How is the Board imposing a change on the developers? > > > The original email had no proposal of a vote. Nor did it sound like a recommendation. Right. But again, where was the impact on the Boost developers? I, as a Boost developer was not expected to, nor asked to, do anything differently from what I had done before. This vote was all about the Board not wanting to deal with the Alliance. We think that there would be an adverse impact on the Bosot developers, and think, "screw the developers!" If we expected our decision to adversely affect Boost development, we would have made a different one. Zach

2 2

Boost Asset Stewardship Review from Matt
by Matt Borland 11 Sep '24

11 Sep '24

DISCLAIMER: I am a C++Alliance Employee Prior to my employment by the C++Alliance I had been a contributor to Boost for a few years primarily in the Math domain. I enjoyed working with everyone, so when an opportunity to join the C++Alliance came around I took it. I agree with the foundation that the core of Boost is volunteers, but by adding a number of part time and full time employees we are able to handle much larger projects and sustained maintenance that may just be otherwise unfeasible. I don't think, for example, GCC would be better off without RedHat employees working on it because many people depend on it just like many people depend on Boost. I like that the C++Alliance is heavily invested in the future of Boost and it's infrastructure. Drone is fast and has a diversity of architectures that aren't available in Github Actions. We've been adding CUDA support to Boost.Math and a testing environment for that was happily provided since there is a heavy demand signal from our users. The new website is a pretty clear improvement over the current one. We also have Sam Darwin who is already helping the distribution of Boost happen behind the scenes. I see no reason to not want to work with an organization that literally puts their money where their mouth is. The fiscal sponsorship model and the new steering committee are the right choice for the future. The fiscal sponsorship model is not new and worked for Boost in the past. The three (possibly five) members of the initial steering committee all have public and sustained interactions with the Boost community. I trust they have the best interests of the project in mind since they have skin in the game. I would like to echo what Chris said in his review. Tensions have obviously been high over the past weeks and months between some rather large personalities. Once this referendum is complete I hope that we can have a peaceful transfer of power and return to a more cordial environment with a focus on how we can make Boost better moving forward. Bottom Line: I vote ACCEPT for the C++Alliance's Fiscal Sponsorship Proposal Matt

2 1

[Asset Stewardship] A request for clarification
by Andrzej Krzemienski 10 Sep '24

10 Sep '24

Hi All, I have a couple of questions to the C++ Alliance representatives regarding the Fiscal Sponsorship Proposal. First, the proposal is very clear and informative. Thank you for this. I still have some questions, though. In Appendix 1: Paragraph 2.c says, "all community programs [...] shall be the ultimate responsibility of Alliance. What is meant by "community programs" here? Paragraph 3 says "should Alliance be required to pay any taxes [...]". What taxes do you mean here? Can you give an example? Paragraph 3 also uses the term, "Project agrees". Is this a mistake, or intentional? I understand that the Current Committee can agree to something. Or that the Steering Committee can agree to something. But how can the Project agree? Additional two questions. 1. When I am a maintainer of a library, what does it make me in the light of this agreement? Am I part of the Project? Or a donor to the Project? Or something else? 2. I have a problem internalizing who decides and signs-off on the complement of the Current Committee. Don't get me wrong. I personally support the proposed complement (including the proposed amendments); I just wonder how it works logically/legally. The Current Committee cannot appoint itself. The Alliance cannot arbitrarily say "you, you, and you". Is this Boost Asset Stewardship Review a vehicle for the legitimization of the Current Committee? Paragraph 8.d says, "If no Successor is found, Alliance may dispose of the Project assets and liabilities in any manner consistent with applicable tax and charitable trust laws." This sounds very harsh to me. If the Project faces the necessity to find another sponsor, and this fails, is the Alliance allowed to offer the domain name for sale, get rid of the mailing list archives, or offer the logo to someone else? Appendix 3 -- I do not see how it illustrates the boost.org Domain Registration Expiration Appendix 6 -- It is very nice promotional material. Thank you for doing this. But the term "the Official C++ Language Slack Workspace" draws my attention, and I have always wanted to ask this: what makes this slack workspace "official"? Neither Boost nor the C++ Alliance represents C++ or could make decisions or blessings in the name of C++. So, is it not a stretch? Regards, &rzej;

3 5

Boost Asset Stewardship Review - Some questions received
by Glen Fernandes 10 Sep '24

10 Sep '24

The following questions were sent to me about the assets that the C++ Alliance already manages. My replies are included: 1. What are these alternate domains, and how do they differ from boost.io ? These are domains that the C++ Alliance provided for us when we had that issue where boost.org was down for a long period of time. At the time I wasn't sure we would even get boost.org back or what the timeline for recovery would be. I thought that we should provide a backup domain that users can be directed to until the issue is resolved. The C++ Alliance offered to help and registered boostlibraries.org and boostcpp.org that point to our website hosting. boost.io is the domain for people to preview the new website that the C++ Alliance built, developed by Fastly with content contributions from Boost users and developers. 2. Why do we need to pay to support downloads if the hosting is free? The storage costs are minimal but the bandwidth is not. At one point Boost downloads approached 60TB per month and Bintray couldn't afford to support us for free. JFrog generously offered to take over. Recently, our downloads started approaching almost 200TB per month, and now even JFrog couldn't afford to keep hosting us for free. They suggested we use a CDN to reduce traffic costs. I reached out to multiple providers to get a good deal. The best rate was from Fastly but this was also too expensive. The C++ Alliance offered to handle these costs and we subscribed to Fastly. Sam Darwin administers our Fastly account. For the list discussion, you can start at: https://lists.boost.org/Archives/boost/2024/01/255654.php Glen

4 4

Re: [boost] Replacing the ML with a forum
by Andrzej Krzemienski 09 Sep '24

09 Sep '24

śr., 26 kwi 2023 o 16:08 Vinnie Falco via Boost <boost(a)lists.boost.org> napisał(a): > Before we replace the mailing list with a forum (and implement a mail > gateway) can I get feedback on how you use the mailing list from your > email client? > Every message comes as a separate email to me. > > Are you using a threaded e-mail reader? Gmail groups all the emails from the topic as an email thread. > If so, do you care if > discussions are changed to be flat (like Discourse)? What does "flat" mean here? I am not familiar with Discourse. Could we get a sample or example? Regards, &rzej; What is important > to you (besides that you can continue to use your email client)? > > Thanks > > _______________________________________________ > Unsubscribe & other changes: > http://lists.boost.org/mailman/listinfo.cgi/boost >

5 7

Boost Asset Stewardship Review from Jakob
by Jakob Lövhall 09 Sep '24

09 Sep '24

I’m not affiliated with either the C++ Alliance or the Boost Foundation. I’m merely a boost user and sends occasional patches as contribution. The regulatory parts of the proposals are naturally US specific, and I want to make it clear that I do not understand the nuances of those parts. This is my short review. Boost Foundation: The bullet points and values stated in the proposal from the Boost Foundation seems strange to me and does not align with what I have been able to discern from public discourse. I haven't seen the transparency that’s being discussed. Engagement to provide said transparency seems lacking from the Boost Foundation. The emphasis on inclusivity over merit feels like a pathway to decline, as it explicitly excludes other groups beyond the intended inclusivity targets. This seems harmful to Boost and its community and should be avoided. The Boost foundation highlights code of conduct (and policing the discussion) as important going forward. I think it can be damaging, and should thus be avoided. Additionally, the Boost Foundation has not adequately addressed the infrastructure needs of Boost. In contrast, the C++ Alliance has stepped in to handle these concerns. C++ Alliance: Vinnie and others from the C++ Alliance have been actively engaging with the community on Slack and the mailing list—my two main points of contact. This provides transparency, builds relationships, and fosters discussions around technical issues, all of which are beneficial. The C++ Alliance has provided much-needed infrastructure assistance, particularly in cases where the Boost Foundation has failed to address long-standing issues (e.g., wowbagger). They support the maintenance (and at times creation) of libraries which I find useful. They promote on twitter, which I like, I've found good tools and libraries from twitter before and I expect others to also do so. having Boost on twitter therefore helps reaching those who might not have thought about boost when thinking about how to solve their current (C++) problems. The C++ Alliance focuses on merit over identity, which I believe is the right approach. Conclusion: I vote to ACCEPT C++ alliance proposal.

2 1

Regression Test Crashing
by Tom Kent 09 Sep '24

09 Sep '24

All- I've been pretty busy lately so just noticed this, but as of August 18 *all* the regression runs have broken during execution. The tail of the results log shows: - cxx11_char16_t : yes [47] - cxx11_char32_t : yes [47] - windows-sdk-supports-bcrypt : no [14] - windows-sdk-supports-bcrypt : no [51] error: Unable to find file or target named error: '/mnt/fs1/teeks99-06/run/boost_root/libs/charconv/build//stage' error: referred to from project at error: '../tools/check_build/test' Similar on windows: - cxx11_char32_t : yes [62] - windows-sdk-supports-bcrypt : yes [21] error: Unable to find file or target named error: '/C:/teeks99-09/run/boost_root/libs/charconv/build//stage' error: referred to from project at error: '../tools/check_build/test' Tom

4 8

Boost Asset Stewardship Review from Frank Wiles
by Glen Fernandes 08 Sep '24

08 Sep '24

Frank contributed the following review below, sent to me directly. > Vote: Transfer Boost Assets to the C++ Alliance > > Association with library: Occasional user. Long time OSS contributor > in other areas. Was paid to help C++ Alliance build the new boost.io > > Rationale: > > While my association with Boost is not large, I have seen first hand > the care the C++ Alliance has taken to both "do the right thing" and > drive for consensus with the Boost community. > > As someone who has been deeply involved in other software nonprofits > (Django Software Foundation directly, Python Software Foundation > tangentially) this proposal is certainly well thought out and > detailed. Most importantly it covers "end state" cases in the event > the Steering Council or C++ Alliance want to change or end the > relationship in the coming years. > > It seems clear to me that the C++ Alliance is driving the majority of > air in terms of breathing new life into Boost and the assets can > safely be within their control, if for no other reason than to reduce > friction in continuing to push things forward. > > -- > Frank Wiles > President/Founder of REVSYS > www.revsys.com > @fwiles Frank, thank you again for the review. Glen

1 0

Boost Asset Stewardship Review from Ali Asadpoor
by Glen Fernandes 07 Sep '24

07 Sep '24

Ali, thank you for the review. Everyone, Ali contributed the following review to glenfe -at- boost.org reposted verbatim here: On Fri, Sep 6, 2024 at 8:49 PM Ali Asadpoor wrote: > > Dear Glen, > > Hear is my review of the proposal doc: > > *** > Currently I don't see a bright future for boost library. Soon STL > will offer networking and later it might offer alternatives to other > awesome libraries in the boost. Seems STL is slowly catching up with > boost and it's killing it. > I was happy to read that The C++ Alliance might provide boost with > a second chance to glow again. Boost needs more human power to > process input from community faster and I think The C++ Alliance can > improve this aspect. > *** > > Best regards > Ali

2 1