ID: | 3114 |
From: | ge...@igor2.repo.hu |
Date: | Sun, 2 Jun 2019 06:45:19 +0200 (CEST) |
Subject: | Peter, very last warning (was: bbvia, Re: [pcb-rnd] discussion: |
in-reply-to: | 3113 from Peter Stuge <pe...@stuge.se> |
Hello Peter Stuge, Congratulations Peter, you are the first person getting the Last Warning Before a Ban. Read carefully, make the necessary corrections in your behavior and let me know if you understood everything. I am doing this in public after I did this a few times in private to absolutely zero result. Other readers can safely skip this mail, there's nothing pcb-rnd relevant in it. The reason I am doing this in public is not some kind of shaming, but transparency: if this ends up in a ban, I want it to be a public process, not some private mailing. I really tried to resolve this in private and as you see, it failed. Rationale: those who have followed geda-user@ for 10+ years, probably remember where this user attitude (always opposing, always knowing better, but not willing to help with the specific part) leads to. I do not want that to happen in pcb-rnd. Peter is repeatedly doing it despite of many very explicit warnings. I think I was extremely patient, but there must be a limit where I stop and say no. I hope we won't ever have this again, but if we do, this can be a precedent/reference/template for handling it. (I know Peter also does some useful bug reporting, and that's appreciated. But that does not justify his distracting mailing list activity.) On Sat, 1 Jun 2019, Peter Stuge wrote: >gedau@igor2.repo.hu wrote: >> >> At the moment we have full support for bbvias in the data model, in core >> >> and in the GUI, but we can not export them, simply because we don't know >> >> how exactly it should be done _in practice_. >> > >> >Here I may be able to contribute something; open hardware manufacturing >> >files (gerber) for an open hardware prototype (designed in KiCad, so KiCad >> >files are also available) which uses bbvias on a 4-layer board, and was >> >fabbed successfully by CONTAG AG outside Berlin. Should I send them over? >> >> Thanks, but what we need is the other way around: if I sit down to >> upgrade our gerber exporter for bbvias, we need to get the result tested. > >Sure. But the first step is to implement something that can be tested, >and for implementation I find known-good data points to be helpful. "first step": false. Sorry to say but you again demonstrated that you have no idea what you are talking about and didn't spend the minimum expected time looking up things before starting to oppose. Then you didn't ask, just stated your strong opinion on how the project should be ran. So let's see the facts. Here are the bbvia steps: 1. understand what bbvia is what it means to the EDA software (done) 2. figure what it needs from the data model (done) 3. design the data model to be able to handle bbvias in a clean way (done) 4. implment all the low level code for bbvias (done) 5. implement the native file format so bbvias can be safely saved and loaded, even when talking about subcircuits that may be placed on boards with different layer stackup (done) 6. implement the GUI that can present and edit bbvias and find.c and whatever other core code that need to deal with the special case (done) 7. implent export code (pending) 8. testing the exporter (what I said I need to see how would be done before I proceed with 7) The first 6 steps are done. For step 7, I am waiting to see how exactly step 8 would be done. This was my decision. I did not ask you to comment on this decision or to come up with a "better" one. _You_ find "good data points useful": are you doing the implementation? No. Are you leading the project? No. Are you paying for this effort to be done? No. Then why does it matter what _you_ would find useful for coordinating or implementing bbvias? I tell you: it does not matter at all, it is totally irrelevant. Believe it or not, I know what I am doing, and I know how to achieve well tested, working features in pcb-rnd. Unlike you, who probably spent a few minutes writing the mail and nothing else on bbvias in pcb-rnd, I did spend many dozens of hours specifically with bbvias in pcb-rnd already. I know what we need if we want a workig bbvia export support. At this stage, it's not data points. And I did make it very clear. Twice! You just ignored it, repeatedly. (So far it's the same pattern as with the recent kicad format coordinated testing, which you didn't contribute to either, but tried to force your opinion. In the same time Karl simply sat down, did what I asked and we already had about half of the coding job done.) My original comment (you are answering) clearly showed that I decide that I need to see how step 8 is done before I start working on step 7. I explicitly said that the first steps have been done. I explicitly said I did NOT need help with the implemenetation at the moment. Knowing you (I am not joking, I had specifically you in mind writing that part) I explicitly ASKED NOT TO start telling me how to do step 7: "I mean we know it in theory, (please don't link 200 documents explaining it), but we don't have any user or fab going to test it for us and bear with the extra cost and consequences of potentially botched boards." Despite that sentence, you did so, and I calmly explained _again_ what course of action I (as a project leader, as the person who implements the feature at the end, as the person who probably makes the investment in it) decided to do. But you can't stop forcing your opinion. So you started another round of opposing, just to press how everything should be done differently. Now I seriously ask you to step back and read again what I wrote originally, and what you answered. Try to look at it from my point of view: - you clearly don't look after things before you formulate your opinion (that's fine for me) - when you have huge gaps in your knowledge (e.g. how far bbvia support is gone, whether our native format supports it), you do not look afte things and you do not ask those who have information, but just fill the gap in with random noise and make yourself believe it's fact (fine by me) - ... and then you publicly state the random noise as fact (absolutely not acceptable in the pcb-rnd community) - and based on the noise you are trying to tell how to lead the project and/or how I should allocate my coding time (not acceptable) - in parallel to that, you did not understand my decision on how we are going to proceed with this from the first comment (acceptable) - I wrote it out again, using different words, as kindly as possible, but you totally ingore the decision and just go on pushing your opinion on how things should be done instead (unacceptable) despite the fact I have warned you about this exact behaviour at least 3..4 times already (sometimes in private, sometimes in public). It is about the _content_ of your posts, good manner doesn't help on this: this is like the 6th time you do these and I will not tolerate this any longer. > >> The problem is not the implemenation. The problem is how do we get >> _our_ implementation tested. > >An implementation that doesn't exist can't very well be tested. Once an >implementation exists then the testers may emerge. :) > >Considering the required effort I guess noone will commit to testing >before something is ready. > False. We have discussed this many times, and I did demonstrate the opposite with the project history of pcb-rnd between 2013..2015 (your proposed method, failed) and 2016..present (my method, works). If you can't accept this method, that's fine. But constantly trying to push the project to switch to your method is unacceptable and is not tolerated. >This feature might also be a good candidate for sponsoring, but there >too the gerbers and/or .lht should exist first. > What are you talking about? Native file format support for bbvia (.lht) exists already since 2017. GUI support too. find.c and othe core stuff do the right thing on bbvias (minus a few lurking bugs probably). You didn't spend 2 minutes to look at things before starting to write down your opinion. Again, for the 42th time. But you go on opposing decisions. Your attitude about this has not improved despite of all warnings. I am not joking, you are really a single step away from a suspension or ban. About the bbvia and what can help the case: There are a lot of things to be done and my time is limited. I need to decide what I spend it on. pcb-rnd most often follows an user-pull method for deciding. If nobody wants to invest in bbvia testing, it is simply not important enough to be done. This method is a well advertised thing pcb-rnd follows since 2016. This is exactly what got pcb-rnd grew out of a personal project into something that attracts users. I am repeating myself, but maybe with different words you will understand: I did invest enough time in the first many steps of bbvia implementation on the parts that did matter. The hard parts are all done. I will invest time in the export code only if I see that it will be tested because: - I don't want to end up having a excellon export that promises bbvia export and then is not tested for years and then some casual user (not subscirbed to the mailing list) just think it works and ruin an expensive board. - the cost of getting involved with the gerber exporter internals ("re-learning the code) is considerable; if I drill down to do it now, but then forget all details and have to revisit it in 2021 because that's the first time anybody signs up for actual testing (with fabs!), I spend that time twice. For no good reason, because the feature was just sitting there unused for 2 years in between. How to proceed: First you really need to think over your answer and why I am so upset about it. Especially that this is really about the 6th (and clearly the last) time. You have succesfully reached the final extents of my patience. In any mailing you are going to do, you need to demonstrate you have understood me. Especially do not try to talk yourself out of it (like you tried and failed so many times before, when I sent you warnings in private about this behavior). Try to understand what you did wrong and make the corrections in your behavior. If that's not possible, or I see any sign of not happening, I will suspend your mailing list subscription. As for the bbvia, as I said earlier: if you want to help, make some commitment for fab testing. By saving some money upfront for ordering test boards or potentially botched production boards, or by seeking your fab and get me an engineer contact who is willing to put in a few hours on looking at gerbers and excellons. If you can't do those, that's fine, but then: - stop opposing my decisions all the time, especially when you literally have nothing to back your oinion up with, other than personal preferences - stop telling me how to run the project - stop trying to allocate my time If you can't help with some part, don't try to invent how to do that part totally differently and especially do not force your idea, just lean back and wait until someone else helps. You don't need to send a mail about such decision. Regards, Igor2
Reply subtree:
3114 Peter, very last warning (was: bbvia, Re: [pcb-rnd] discussion: from ge...@igor2.repo.hu