digitalmars.D - The process described in the linked article could be a good thing for
- max haughton (4/4) May 24 2022 https://sharedphysics.com/everything-is-important/#step-1-create-a-unifi...
- forkit (11/15) May 24 2022 ummm..divide and conquer.
- max haughton (10/27) May 24 2022 We don't have people enough to need to do that, what I mean is
- forkit (11/43) May 24 2022 You put too much faith in 'systems on servers' ;-)
- forkit (7/9) May 24 2022 Boyd's loop is targetted towards decision makers (i.e.
- max haughton (24/35) May 24 2022 I'm saying we have enough humans to do stuff we are just not good
- forkit (18/35) May 25 2022 This is why you need managers. Not to 'manage people', but to
- forkit (5/5) May 25 2022 On Wednesday, 25 May 2022 at 08:44:47 UTC, forkit wrote:
- forkit (3/3) May 25 2022 On Wednesday, 25 May 2022 at 09:09:40 UTC, forkit wrote:
- max haughton (11/15) May 25 2022 And yes I do agree that we have been ridiculously slow when it
- max haughton (5/11) May 25 2022 We have recently starting making a push to get everything under
- forkit (9/12) May 25 2022 Ahh. The holy grail for programmers ;-)
- max haughton (13/20) May 25 2022 I am specifically talking about building things on servers,
- forkit (21/43) May 25 2022 Nothing you've talked about will occur, without leadership and
- max haughton (14/16) May 25 2022 This is exactly the kind of thing that I am trying to stress
- forkit (5/22) May 25 2022 Well, I agree that we disagree ;-)
- max haughton (7/9) May 25 2022 I am not saying this can be avoided, I'm saying that this a
- forkit (11/21) May 26 2022 Umm.. psychology is a science, with plenty of peer-reviewed
https://sharedphysics.com/everything-is-important/#step-1-create-a-unified-view-of-all-existing-work We need step 1 (and then the others), but step 1 in particular is something we're missing. Suggestions as to how to implement this are appreciated.
May 24 2022
On Tuesday, 24 May 2022 at 18:07:30 UTC, max haughton wrote:https://sharedphysics.com/everything-is-important/#step-1-create-a-unified-view-of-all-existing-work We need step 1 (and then the others), but step 1 in particular is something we're missing. Suggestions as to how to implement this are appreciated.ummm..divide and conquer. to get a unified view of a business, for example, you divide it into its components (sales, marketing, customer-service, logistics, I.T, finance...etc). Then you ensure there is management, and competent management, in those areas. Then you bring that management together, and try to get the big picture. You replace those management who are unable to assist in this process ;-)
May 24 2022
On Tuesday, 24 May 2022 at 22:44:36 UTC, forkit wrote:On Tuesday, 24 May 2022 at 18:07:30 UTC, max haughton wrote:We don't have people enough to need to do that, what I mean is that we need to build systems on servers rather than brains to keep track of what people are doing and observe the effects. If we take Boyd's OODA loop as an abstraction to target we currently are struggling with step one and two: Observation is somewhat informal but not impossible, orientation is very hard because we can't easily keep track of what's happening in way that's accessible to either external readers or would-be volunteers.https://sharedphysics.com/everything-is-important/#step-1-create-a-unified-view-of-all-existing-work We need step 1 (and then the others), but step 1 in particular is something we're missing. Suggestions as to how to implement this are appreciated.ummm..divide and conquer. to get a unified view of a business, for example, you divide it into its components (sales, marketing, customer-service, logistics, I.T, finance...etc). Then you ensure there is management, and competent management, in those areas. Then you bring that management together, and try to get the big picture. You replace those management who are unable to assist in this process ;-)
May 24 2022
On Wednesday, 25 May 2022 at 00:15:56 UTC, max haughton wrote:On Tuesday, 24 May 2022 at 22:44:36 UTC, forkit wrote:You put too much faith in 'systems on servers' ;-) There are six git repositories, and one issues repository. At the very least, there should be a (separate) human in charge of managing each, and also attracting, organising, and delegating, to other talented humans, within each repository. To the extent that the 'systems on servers' do not meet the needs of those managing each repository, that is a matter for them to attend to. Managing is the key to success, not 'systems on servers'. If there's failure in the D community, it is failure of managing.On Tuesday, 24 May 2022 at 18:07:30 UTC, max haughton wrote:We don't have people enough to need to do that, what I mean is that we need to build systems on servers rather than brains to keep track of what people are doing and observe the effects. If we take Boyd's OODA loop as an abstraction to target we currently are struggling with step one and two: Observation is somewhat informal but not impossible, orientation is very hard because we can't easily keep track of what's happening in way that's accessible to either external readers or would-be volunteers.https://sharedphysics.com/everything-is-important/#step-1-create-a-unified-view-of-all-existing-work We need step 1 (and then the others), but step 1 in particular is something we're missing. Suggestions as to how to implement this are appreciated.ummm..divide and conquer. to get a unified view of a business, for example, you divide it into its components (sales, marketing, customer-service, logistics, I.T, finance...etc). Then you ensure there is management, and competent management, in those areas. Then you bring that management together, and try to get the big picture. You replace those management who are unable to assist in this process ;-)
May 24 2022
On Wednesday, 25 May 2022 at 00:15:56 UTC, max haughton wrote:If we take Boyd's OODA loop as an abstraction to target .... ...Boyd's loop is targetted towards decision makers (i.e. management) - ie. humans. If you're saying D does not have the humans to put into management positions, then D does not have the humans to make decisions, and thus, pursuing Boyd's loop (or any other form of managing).. is kinda pointless ;-)
May 24 2022
On Wednesday, 25 May 2022 at 03:16:05 UTC, forkit wrote:On Wednesday, 25 May 2022 at 00:15:56 UTC, max haughton wrote:D contributors are indeed humanIf we take Boyd's OODA loop as an abstraction to target .... ...Boyd's loop is targetted towards decision makers (i.e. management) - ie. humans.If you're saying D does not have the humans to put into management positions, then D does not have the humans to make decisions, and thus, pursuing Boyd's loop (or any other form of managing).. is kinda pointless ;-)I'm saying we have enough humans to do stuff we are just not good at keep them pointing at the right stuff, for reasons similar to those described in the linked article. In a project like D naively aiming to manage people will just piss people off, we need to make progress observable and make it obvious where actually needs work and what other people are trying to do. Being able to measure some notion of progress, even if fuzzy, is good way of getting more of that thing: In a very simple sense, which types of D outreach bring the most traffic to dlang.org? How many people have downloaded the nightly build on github etc. (hint: not many). Managing programmers is a full time job even if you're paying them to work for you. I think a lot of these nicely packaged ideas like OODA are bullshit, I just think there is obviously progress we can make either for individuals or for D as a directed programming language in the rough concept of the loop. The specifics don't really matter. This concept has an added benefit in allowing people new to contributing to find something to do without having to actively seek out people they may not know exist yet. Similarly we have no data on whether people try and fail to contribute, get stuck, etc: sticking a little survey somewhere is better than nothing.
May 24 2022
On Wednesday, 25 May 2022 at 03:44:08 UTC, max haughton wrote:... I'm saying we have enough humans to do stuff we are just not good at keep them pointing at the right stuff, for reasons similar to those described in the linked article.You need better managers then.In a project like D naively aiming to manage people will just piss people off, we need to make progress observable and make it obvious where actually needs work and what other people are trying to do. Being able to measure some notion of progress, even if fuzzy, is good way of getting more of that thing: ...This is why you need managers. Not to 'manage people', but to manage these issues.Managing programmers is a full time job even if you're paying them to work for you.One does not manage programmers ;-)This concept has an added benefit in allowing people new to contributing to find something to do without having to actively seek out people they may not know exist yet. Similarly we have no data on whether people try and fail to contribute, get stuck, etc: sticking a little survey somewhere is better than nothing.Again, this is why you need someone managing this. It won't happen otherwise. Try asking the most important questions first. Who is managing this repository, and that repository? At the moment, I don't know. Do you? Is it Walter? Does he do it all? I cannot currently go to a manager and say, hey, I think these areas need to be improved, and here are some ideas on how it might be done. Just having the idea and spraying it around the NG won't make much difference to anything. If D wants to set itself up for a new phase, it will need good managers, not just good programmers.
May 25 2022
On Wednesday, 25 May 2022 at 08:44:47 UTC, forkit wrote:Perhaps D needs a formal, 'engineering services' team? (with the emphasis on 'team', as opposed to 'person') case study: https://azure.microsoft.com/mediahandler/files/resourcefiles/devops-at-microsoft-innovating-on-open-source/DevOps%20at%20Microsoft%20-%20.NET.pdf
May 25 2022
On Wednesday, 25 May 2022 at 09:09:40 UTC, forkit wrote:I think this says it all ;-) https://wiki.dlang.org/?title=Special:Search&search=Vision%2F
May 25 2022
On Wednesday, 25 May 2022 at 09:20:12 UTC, forkit wrote:On Wednesday, 25 May 2022 at 09:09:40 UTC, forkit wrote:And yes I do agree that we have been ridiculously slow when it comes to producing vision documents. Part of this I think is Michael being a perfectionist but more generally there is a framing issue in that I think a vision statement has to be a quasi-living document tracked in git and published loudly somewhere rather than a wiki page. But underneath the overall vision there is still a vast amount of things to organize and measure. The level of granularity I want to attack is not initially the big picture stuff, that's not where the real development work has issues.I think this says it all ;-) https://wiki.dlang.org/?title=Special:Search&search=Vision%2F
May 25 2022
On Wednesday, 25 May 2022 at 09:09:40 UTC, forkit wrote:On Wednesday, 25 May 2022 at 08:44:47 UTC, forkit wrote:We have recently starting making a push to get everything under one IaC roof when it comes to infrastructure. The whole idea behind doing that is to eliminate much of the need for a team at all, it's just code.Perhaps D needs a formal, 'engineering services' team? (with the emphasis on 'team', as opposed to 'person') case study: https://azure.microsoft.com/mediahandler/files/resourcefiles/devops-at-microsoft-innovating-on-open-source/DevOps%20at%20Microsoft%20-%20.NET.pdf
May 25 2022
On Wednesday, 25 May 2022 at 09:27:23 UTC, max haughton wrote:... The whole idea behind doing that is to eliminate much of the need for a team at all, it's just code.Ahh. The holy grail for programmers ;-) In the end though, it will always come back to leadership, and interactions between people; not 'systems on servers'. "It seems clear that the success of armies, athletic teams, business enterprises, universities, religious organizations—any collective activity—depends on the leadership of that collectivity." https://www.hoganassessments.com/blog/personality-theory-and-the-nature-of-human-nature/
May 25 2022
On Wednesday, 25 May 2022 at 10:17:38 UTC, forkit wrote:On Wednesday, 25 May 2022 at 09:27:23 UTC, max haughton wrote:I am specifically talking about building things on servers, because that is something we can actually do. Almost everything you have mentioned so far just isn't doable in the short term. Or actually relevant to what is blocking current progress. I am happy to coordinate and manage programmers but I just don't think it's a realistic model for the D community because the vast majority are working for free. The dynamic just isn't the same as a company. Similarly I want incremental progress to be measurable, not to impose a grand model onto the process. This is going to be a matter of working hard and smart, not a battle of management philosophy.... The whole idea behind doing that is to eliminate much of the need for a team at all, it's just code.Ahh. The holy grail for programmers ;-) In the end though, it will always come back to leadership, and interactions between people; not 'systems on servers'.
May 25 2022
On Wednesday, 25 May 2022 at 10:26:14 UTC, max haughton wrote:On Wednesday, 25 May 2022 at 10:17:38 UTC, forkit wrote:Nothing you've talked about will occur, without leadership and interactions between people. Better systems don't just suddenly appear out of nowhere. If you're not focused on this, you may be working 'hard', but not very 'smart'. Also, you do not need to 'coordinate and manage' programmers. You only need to set a clear agenda, with priorites, and measureable goals; which again, requires leadership and interactions between people. You are correct that better systems are needed (as well) - cause inviting programmers to contribute, by saying ..there's a bunch of bugs in our bug repository, go pick one... is not very motivating ;-) If there's going to be 'a next phase' for D, it's going to come about because of leadership and interactions between people. That is the only way it can occur. So I have to disagree with your assertion, that this is 'not relevant to what is blocking current progress'. If it were me, I 'lead' by saying D3 will be safe by default. Now, let's get to work on it.On Wednesday, 25 May 2022 at 09:27:23 UTC, max haughton wrote:I am specifically talking about building things on servers, because that is something we can actually do. Almost everything you have mentioned so far just isn't doable in the short term. Or actually relevant to what is blocking current progress. I am happy to coordinate and manage programmers but I just don't think it's a realistic model for the D community because the vast majority are working for free. The dynamic just isn't the same as a company. Similarly I want incremental progress to be measurable, not to impose a grand model onto the process. This is going to be a matter of working hard and smart, not a battle of management philosophy.... The whole idea behind doing that is to eliminate much of the need for a team at all, it's just code.Ahh. The holy grail for programmers ;-) In the end though, it will always come back to leadership, and interactions between people; not 'systems on servers'.
May 25 2022
On Wednesday, 25 May 2022 at 23:02:54 UTC, forkit wrote:If it were me, I 'lead' by saying D3 will be safe by default. Now, let's get to work on it.This is exactly the kind of thing that I am trying to stress isn't the scope I'm talking about: You can't coordinate an open source project like this, it doesn't work. The overall strategy isn't the issue, I want to improve the smaller things underneath that. The apparatus for tracking smaller projects simply isn't there. Have you contributed to D? There is a huge amount of stuff at the bottom of the stack that needs improving and tracking, the actual design of the language is the easy part as far as I'm concerned. Even ignoring that you're talking about a level of structure that is probably impossible unless the D Foundation itself had a significant number of full time employees, it's just not what I'm talking about.
May 25 2022
On Wednesday, 25 May 2022 at 23:32:36 UTC, max haughton wrote:On Wednesday, 25 May 2022 at 23:02:54 UTC, forkit wrote:Well, I agree that we disagree ;-) You want to advance the D project using 'systems on servers'. I think it can only advance, through better understanding human and social psychology.If it were me, I 'lead' by saying D3 will be safe by default. Now, let's get to work on it.This is exactly the kind of thing that I am trying to stress isn't the scope I'm talking about: You can't coordinate an open source project like this, it doesn't work. The overall strategy isn't the issue, I want to improve the smaller things underneath that. The apparatus for tracking smaller projects simply isn't there. Have you contributed to D? There is a huge amount of stuff at the bottom of the stack that needs improving and tracking, the actual design of the language is the easy part as far as I'm concerned. Even ignoring that you're talking about a level of structure that is probably impossible unless the D Foundation itself had a significant number of full time employees, it's just not what I'm talking about.
May 25 2022
On Thursday, 26 May 2022 at 01:03:00 UTC, forkit wrote:I think it can only advance, through better understanding human and social psychology.I am not saying this can be avoided, I'm saying that this a completely different level of abstraction to what I'm trying to suggest by linking the article (more specifically the first step). This is the absolute basics. There's no point in waffling about psychology if you have almost no way of judging the effect of any changes you make.
May 25 2022
On Thursday, 26 May 2022 at 01:27:45 UTC, max haughton wrote:On Thursday, 26 May 2022 at 01:03:00 UTC, forkit wrote:Umm.. psychology is a science, with plenty of peer-reviewed scientific research to back up the claims it makes. I am unaware of any management/leadership/motivation/decision-making (etc) theories that don't draw on such research. They all involve humans afterall. If anything I've said is unrelated to your goals fine, ignore it, but it's not 'waffling' - it comes from peer-reviewed, scientific research. Good luck with your strategy. I hope it works out ;-)I think it can only advance, through better understanding human and social psychology.I am not saying this can be avoided, I'm saying that this a completely different level of abstraction to what I'm trying to suggest by linking the article (more specifically the first step). This is the absolute basics. There's no point in waffling about psychology if you have almost no way of judging the effect of any changes you make.
May 26 2022