closed-access mailing lists https://lists.riscv.org/g/main/subgroups

59 views
Skip to first unread message

lkcl

unread,
Oct 2, 2019, 6:44:13 AM10/2/19
to Community, RISC-V Community Response Team
https://lists.riscv.org/g/main/subgroups

this is an ongoing issue which, unfortunately, needs to be expressed, discussed, and a solution thought through.

as has been highlighted a number of times over the years, it can be seen that these lists are all closed, secretive lists.  not even the archives are accessible, which automatically prevents and prohibits participation in innovation from commercial projects that have "full transparency" as part of their mission-critical business objectives.

it has also been highlighted that the perception of such commercial projects is that because they are "Libre", these two requirements - Commercial *and* Libre - must be so completely diametrically opposed and mutually exclusive [along the lines of: "no Libre Project could EVER raise the USD $10m to reach silicon"] that their proponents must be either insane, outright liars, or "Religious Frothing Zealots", with many people expressing "amusement" and "patronising tolerance" to cover what basically amounts to an extraordinary and shockingly disrespectful tone towards anyone with such objectives.

worse than that, with only the *public* lists being available to groups with these two objectives, and with the majority of participants on those same lists having signed the RISC-V Membership Agreement and having direct access to the (secretive) information and resources, such signatories to RISC-V Membership have, understandably, been getting quite frustrated at having to repeat things time and time again, with the "Outsiders" being effectively forced to *BLACK-BOX REVERSE-ENGINEER* [aka "guess"] the restricted-access information, get it wrong, and thus frustrate signatories even further.

even worse than that, witnesses over the past two years to such black-box reverse-engineering efforts, in combination with the ongoing disrespectful attitude shown by some participants, has led to increased frustration and, in some cases, outbursts of such outright vitriolic and violent hatred that it leaves some considerable concern for physical safety about ever meeting such people at Conferences.

as if that was not bad enough, *nobody has called these people out*.

to repeat that: there has not been a *single* message from anyone in authority within the RISC-V Foundation to indicate that:

* the disrespect shown towards Libre Business Innovators is not acceptable
* the expression of violent hatred of other members is not acceptable
* the ongoing effects of keeping the lists and other resources "closed" is recognised and being considered.

instead, we have total stone-cold silence, which, unfortunately, can only mean that either the RISC-V Foundation *doesn't care*, or that they actively *endorse* such behaviour.

the most obvious solution would be to open up access to these resources.  the last time that this was raised, a response was given from Krste Asanovic, that the lists were closed for the protection of the RISC-V Foundation.  this response was great, and *would* have been the entry-point for further dialog... except there was absolutely no follow-up engagement.  further questions were once again met with stone-cold silence.

can we please have some dialog here, some acknowledgement of this issue *as* an issue, because it will go a long way to undoing the damage to the public reputation of RISC-V.

what solutions are available, here?

l.

lkcl

unread,
Oct 2, 2019, 7:50:06 AM10/2/19
to Community

https://groups.google.com/a/groups.riscv.org/forum/?nomobile=true#!topic/isa-dev/Wt7ts281qJ8

To clearly illustrate the extent of the problem, the above discussion can be monitored as to whether there is any response.

Our group is *actively* helping to review RVV - an official RISCV Specification - yet there is consistent and persistent ongoing total stone cold silence, even when serious ambiguities and flaws, not to mention opportunities for improvement, in the RVV Spec.

Likewise the main BitManip author also consistently fails to respond or engage.

This leaves the entire RISCV Community worse off.

It may seem "perfectly reasonable" to exclude contributors who will not use "allotted resources" because, of course, they are "causing inconvenience for everybody else", however I can assure you that Trademark Law takes a dim view of such discrimination, as will the wider public.

L.


lkcl

unread,
Oct 6, 2019, 2:47:59 PM10/6/19
to Community, con...@riscv.org


On Wednesday, October 2, 2019 at 7:50:06 PM UTC+8, lkcl wrote:

https://groups.google.com/a/groups.riscv.org/forum/?nomobile=true#!topic/isa-dev/Wt7ts281qJ8

To clearly illustrate the extent of the problem, the above discussion can be monitored as to whether there is any response.

So, sadly, as expected, it is with much regret that I have another Conduct Violation to report.  Again, it is the subtle kind which is so hard to characterise, by virtue of what has *not* been said, rather than what has.  Again, I remind you that I should in no way even have to be raising these issues.

Let's go over the messages, comparing them where appropriate, to similar messages, illustrating the difference.

1. "hello there is a problem with RVV Spec, it is not clear".

expected response: "thank you, appreciated, will fix it"
actual resonse: none.

2. "hello i did some work writing some pseudocode, to get clarity on the spec".

expected response: "thank you, appreciated, let's work TOGETHER on this"
actual resonse: none.

3. "hello I spent more of my sponsors money tracking down the simulator, the lists everyone else has access to I cannot join, so you wil expect you to make allowances for that, but the simulator does not have these instructions so we cannot check the spec"

expected response: "whoops yes we know, it is a pain, we are wotking to fix that, in the meantime thank you, let's get the spec fixed at least, work together on that".

actual response: "lkcl you are a moron for wadting everybody's time by not knowing that SiFive has had RVV in mainline simulator for months. I am going to ignore the reason why you raised this issue, make sure that I focus on distraction, make you look like a total dick as much as possible, in the hope that everybody else piles in and accuses you of being a disruptive dickhead too".

4. "err Bruce, the proprietary code you said contains an independent simulation is incomplete, the mainline RVV doesn't have these misding instructions either, you completely ignored the main issue of the spec ambiguities, the list goes on. There was *nothing* useful, welcoming, or constructive about your reply".

expected response: "ah whoops, sorry, yes, my mistake for not picking up on the main issue, and many apologies for attacking you, i am having a bad day that's all. I will alert people on the internal lists to come and interact here and work with you to fix the ambiguities that you kindly raised"

actual response: stone cold silence.

Now, by contrast, Andrew raised a spec review, someone responded, and what did Andrew do? Within 24 hours he said "thank you for your contribution, I will fix this".

You see the difference? You see how much less time and energy it took to make an improvement?

That's all it takes. A simple "ah thank you for spotting this, we will get it fixed".

This says "your input is valued. you are a valued contributor and part of this community. we work together and do so for everyone's benefit."

These constant systematic attacks are poisoning the entire community for everyone. People reading these archives, which they will be able to do for the next 10, 20, 30 years, will see these ongoing systematic and extremely cleverly disguised community violations with both horror and sadness, that such a fantastic idea was poisoned and turned to ash.

Why are people unable to say "thank you for your valuqble contribution"?

When is this "poisoning" going to stop?

When is someone going to start saying "this ia completely unacceptable, we are dealing with it immediately"

What is it going to take for people to wake up and realise the damage being done to RISC-V's reputation?

Do you understand why I said that this is not a healthy community?

When I asked people what kind of community they would like this to be, it was and still is a genuine question. The problem is that most good people are horrified at the behaviour of those with close ties to SiFive and the Founders, and have walked away and formed "closed lists" amongst themselves, in order to shield and distance themselves from these lists and get work done.

This in and of itself creates a fragmented community, undermining the entire spirit of why RISCV was set up in the first place.

Enough.

L.

RISC-V Community Response Team

unread,
Oct 16, 2019, 6:11:24 PM10/16/19
to lkcl, Community
  Hi Luke, thank you for expressing your concerns. There are several discrete issues presented here. We will attempt to address them as well as we can.

The RISC-V Foundation was set up as a standards body in 2015 to develop and protect the RISC-V ISA copyright and the RISC-V trademark while still enabling development through open processes. It was designed to encourage lots of companies to join and support, as otherwise RISC-V would have remained an academic project. After a long legal development process, RISC-V was set up as it currently is, with a set of bylaws and a membership agreement that protects both the developers and the intellectual property. As Krste has said, this is why the development community requires participants to sign the membership agreement.

The Foundation has done everything it can to balance the needs of the organization and legal structure with the needs of the community. For example, while for-profit enterprises are required to join as a paying member, any individual person can join as a member at no cost as long as they are willing to sign the membership agreement and abide by the Foundation's bylaws.

The working task groups are closed to the public for legal reasons based on the Foundation bylaws that were put into practice when it began. There is an ongoing effort to change this restriction, but those changes also need to reflect the needs of the membership who are responsible for the project's IP as well as their own.
 
The public lists were introduced to give non-members an opportunity to discuss the RISC-V ISA and the hardware and software being developed around it. Foundation members are not required to participate on them, though many do, and as you note, projects use these groups for periods of public commentary during the development of new specs.  Until recently these groups were not directly moderated.

The Code of Conduct provides a guideline for acceptable behavior and a framework for dealing with people who go beyond what is acceptable - it isn't perfect in this regard but we intend to develop it as we move forward, in conjunction with feedback from the community. However - we need to be clear that the Code of Conduct does not restrict criticism, as that is not its role nor does it mandate that people respond to comments individually. All open source communities need to provide some degree of freedom for participants to express their views, yours included.

By our research, no one from the Foundation has ever referred to libre RISC-V efforts as insane or religious zealotry. If you have encountered this kind of rhetoric on our lists since we introduced the CoC, please give us pointers. If you are referring to opinions expressed outside of the RISC-V groups, that is outside our scope.

The community response team is made up of volunteers. and responding to messages like this is not anyone's full-time job. Responses may not be fast, but we will endeavor to answer all concerns addressed to the Foundation on this list. We appreciate your patience.
_____________________________
RISC-V Foundation Response Team  

lkcl

unread,
Oct 16, 2019, 7:23:48 PM10/16/19
to Community
Hi folks, really appreciated the insights and the discussion. Am in transit and v. busy so responses are limited, apologies in advance.

Alex Bradbury raised the question last year after Samuel Falvo very reasonably pointed out the damage caused by ITU Style Standards development. Krste did indeed respond as you reiterated... except he then walked away from the conversation and failed to engage further. This is clearly extremely frustrating and has caused a lot of ongoing problems.

There are several solutions here: the Foundation needs to be prepared to actually listen to constructive advice ( instead of viewing it as "ad hominem attacks")

If the Khronos Group can get this balance right, then so can the RISCV Foundation.

In addition, developers obviously have to interact when it comes to LLVM, gcc, linux kernel and so on.

It is nonsense that there is anything "secretive" about the UNIX Platform WG for example, and excluding input from key people unrelated to RISCV *is* going to result in harm and mistakes.

Likewise people who would otherwise be happy to contribute to the CLIC WG have contacted me privately in total despair on discovering the overwhelming complexity of what is transpiring, there.

45 days to "respond", catching up on years of development, devoid of context, is completely unrealistic.

All of these things and more have been said many times and we have never received a response or an acknowledgement that someone is actually listening (until today).

So, thank you.

L.

lkcl

unread,
Oct 17, 2019, 2:01:04 AM10/17/19
to Community
On Thursday, October 17, 2019 at 6:11:24 AM UTC+8, RISC-V Community Response Team wrote:

Hi I have some additional time available.

> As Krste has said, this is why the development community requires participants to sign the membership agreement.

This argument was stated last year. When it was pointed out that this is a conflation (logic error) - not by me - there was no redponse.



> The Foundation has done everything it can to balance the needs of the organization and legal structure with the needs of the community. For example, while for-profit enterprises are required to join as a paying member, any individual person can join as a member at no cost as long as they are willing to sign the membership agreement and abide by the Foundation's bylaws.

This is where the bias against Libre Commercial projects comes in.

The assumption has been made that individuals are incapable of raising the substantial funds necessary to develop a processor outside of a profit maximising Corporate structure or a University setting.

Anyone claiming to be capable of doing so must therefore be a liar or deluded and thus a source of amusement at best, tolerated whilst the fun lasts, and, if they persist in their "mental delusion" and do not go away, they are to be attacked, ostracised and treated with further disrespect.

I have never found the condescending attitude towards the goals of the Libre RISCV Project to be in the least bit funny.

Rather than *ask* us "whoops we in no way anticipated this scenario, how can we work WITH you to ensure that your valued and valuable contributions and insights can be part of the innovation in RISCV" we have stone cold silence and thus have had to investigate and deploy a "Laches" strategy
which ultimately nullifies and even invalidates the RISCV Trademark in order to ensure unimpeded forward progress and commercial success of the project.

We currently have *nine* separate Extensions to RISCV under development in order to enter the 3D and Video markets, and that is expected to go up (including Texturisation, Z Buffering and so on).

In addition, several mistakes and omissions in official extensions have been pointed out over the past 18 months: without exception they gave been ignored, "justified away" or goal-post-moved, anything, absolutely anything done to ensure that under no circumstances is there a reason to include a member of our team in the Acknowledgements as a valued Contributor.

All it would take is to *ASK* us "what would you be happy to sign or agree to" and yet rather than do that, we receive condescending offers to pay the USD 100 Membership fee.

If someone actually bothered to ask us we would likely be able to come to a suitable agreement on the Extensions that we have developed and curated (with valuable input from many people in the community), yet we have condescending and highly embarrassing attacks by SiFive employees on the reputation of our project instead to deal with.

All of which is deeply unimpressive.

So no, the RISCV Foundation has *not* gone to adequate "trouble" to be inclusive of our highly unusual and unique business requirements, and thus, if this silence persists, is in danger of losing the Trademark.

The only reason I tolerate the condescension, public derision and silent treatment is because it strengthens a "Laches" Trademark defense.

Clearly this is in nobody's best interests to have this stone cold silent treatment continue, and definitely not the RISCV community on whose good will our commercial success critically depends as much as any other profit maximising Corporation does.

We are acting in good faith and it is absolutely ridiculous that *we* have to be the ones acting responsibly and yet teying to *reverse engineer* and derive best practices and procedures which have already been documented and to which we are *prohibited access* due to the Foundation failing in their duty and responsibility to ASK us what the issue is.

This is the repeated pattern of behaviour that the Founders engage in. As Academics they tell people "how it is" and completely fail to listen to outsiders. Academia has always been thus.

It conflicts extremely badly with the responsibilities and duties of a Trademark Holder and I am getting very fed up of having to point this out.

I would like to see RISCV be long term successful and inclusive of ALL innovative efforts regardless of origin.

L.

Reply all
Reply to author
Forward
0 new messages