清空筛选条件
清空筛选条件

No matter what I try, I cannot get these plots to come up in separate figure windows.

39 次查看（过去 30 天）
Richard K Collins
Richard K Collins 2021-3-22
评论： Jan 2021-3-27
This should have taken five minutes.
Url = 'http://service.iris.edu/irisws/timeseries/1/query?net=AV&sta=DLL&cha=HDF&start=2011-01-01T00:00:00&end=2011-01-01T00:03:00&scale=AUTO&format=ascii1&loc=03';
Text = webread(Url);
Lines = strsplit(Text,'\n');
N = length(Lines) - 1;
values = zeros(1,N,'double');
for i = 1:N
values(i) = str2double(Lines{i+1});
end
figure('name','first');
plot(1:N-1,diff(values,1));
figure('name','second');
plot(1:N-2,diff(values,2));
figure('name','third');
plot(1:N-3,diff(values,3));
figure('name','fourth');
plot(1:N-4,diff(values,4));
figure('name','Raw');
plot(1:N,values);
---------
windows 7 64bit professional.
It is a trivial problem. Read the data from IRIS, split into lines, extract the first column (only one column), plot the first 4 differences, plot the raw data.
But it keeps putting the figures into the code in the live editor. I have been running other things all day where the figures come up on my other screen.
Yes, I cleared command, output, workspace. Exit and return. All the usual. Some damn convention some made up and the passed along as verbal history. Horrible trying to find anything online. Yes, I like talking to people, but I hate having to ask a human for every tiny little bit of what did someone make up for this situation.
---------------------
I tried this below and it would not plot anything and gave no error messages.
for k = 1:4
figure(k);
plot(1:N-k,diff(values,k));
end
----------------------
Going to the output and manually opening leaves the figures on the screen, but running again won't update those figures. It keep shoving them into the output area next to the live editor. I know this is trivial and stored in someone's head. I hate memorizing stuff like this. And it is impossible to search for. There has to be a better way.
----------------
I went through the 130,142 files and 14,031 folders in my Program Files/MATLAB folder to see how it is organized. It is a jumble of 207 different extensions, several that are ambiguous. Most of it hand built. Many automated files in text and binary with no map as to how they are connected. Fragmenting things into so many pieces means file access is slow. It needs to all be in a database, indexed and mapped. All these data files and no organization. No wonder ever trivial question takes hours because all the answers are in people's heads or written in incomplete human to human text.
  8 个评论
显示 6更早的评论
Jan
Jan 2021-3-22
编辑：Jan 2021-3-22
@Richard K Collins: Simply type "@" in the comment and the forum's interface suggests the names of the participants of the discussion.
"I went through the 130,142 files and 14,031 folders in my Program Files/MATLAB folder to see how it is organized." - This is not a useful strategy to master Matlab.
"I put "clear figure" because I was trying random things" - this kind of shotgun programming is not efficient, but will increase the confusion level.
"I don't see anyone in the MatLab community with any sense of urgency for anything." - seriously?!
It sounds like you are tired. Take a break. Drink a coffee. We are all working on the same trillion of dollar project. Welcome in the team. Matlab is a tool. Use it, if it helps to solve your problem. But if you want to pull a nail into the wall, use a hammer instead. Blaming Matlab not not hit nails is a waste of time.
Richard K Collins
Richard K Collins 2021-3-22
@Jan - Yes, I am tired, and thanks for your concern. But you are complaining about random things, and not answering the question I had about why MatLab graphs do not seem to work consistently. If you want, I can not ask anything, or share anything. You have to hear my voice and listen to what is in my heart and mind. I am not blaming MatLab the people, but the systems that everyone is relying on now for the Internet, for programming, for systems for large communities. I think MatLab takes people years to master. And it is no different conceptually or cabability wise than any of hundreds of similar tool kits. I thought it was better integrated into devices, and I thought more stable and refined. I remember when MatLab started. It is a good community, it deserves better. When I hear criticisms of systems, I take that and remember. And try to fix it. Matlab is too good to throw away, but it is a lot of work to clean it up and make efficient for people to learn quickly. I am weighing that now.
Current Covid deaths are 2.72 Million in less than a year. I project about double that, now that everyone is being exposed. The developed country vacines are not enough. So while we talk more people die.
Have you found how to fix the communication problems on the Internet regarding covid? If you already did it, then I can rest, assured. But I think not, and I will just keep going. I am looking for help on serious things, and tired of these little things that get in the way. There is no easy way to trace problems in MatLab. It is not open so I cannot check the source code. It is not laid out well to read the files and figure it out. Not sure what best to do, but stopping when things get hard is not an option. I cannot pay people to help me, so I solve problems for groups and hope they are grateful enough to help. Not the best strategy in the world, but time is not kind.

请先登录，再进行评论。

请先登录，再回答此问题。

采纳的回答

Image Analyst
Image Analyst 2021-3-22
It works fine for me. I get 5 figures, though they are all on top of each other. You might want to use subplot(2, 3, k) to put the plots in different subplots on the same figure instead of in separate figures.
Try this:
Url = 'http://service.iris.edu/irisws/timeseries/1/query?net=AV&sta=DLL&cha=HDF&start=2011-01-01T00:00:00&end=2011-01-01T00:03:00&scale=AUTO&format=ascii1&loc=03';
Text = webread(Url);
Lines = strsplit(Text,'\n');
N = length(Lines) - 1;
fprintf('There are %d lines of text in the URL.\n', N);
values = zeros(1,N,'double');
fprintf('Please wait . . . Getting values from the text....\n');
for k = 1 : N
values(k) = str2double(Lines{k+1});
end
fprintf('Done getting %d values from the text.\n', N);
figure('Name', 'My 5 plots', 'NumberTitle', 'off')
subplot(2, 3, 1);
plot(1:N-1,diff(values,1));
grid on;
title('First Plot');
subplot(2, 3, 2);
plot(1:N-2,diff(values,2));
grid on;
title('Second Plot');
subplot(2, 3, 3);
plot(1:N-3,diff(values,3));
title('Third Plot');
grid on;
subplot(2, 3, 4);
plot(1:N-4,diff(values,4));
grid on;
title('Fourth Plot');
subplot(2, 3, 5:6);
plot(1:N,values);
grid on;
title('Raw');
  6 个评论
显示 4更早的评论
Walter Roberson
Walter Roberson 2021-3-26
"Automatically recognize code" is in tension with "separate code editors that can apply to sections inline.". MATLAB Answers has a separate code editor that applies to sections inline; it just has to be activated (the '>' button in the toolstrip)
It is not possible to create an editor that recognizes and handles all computer languages automatically. This is not just a problem with trying to catalog all computer languages and their syntaxes: this is a problem that there provably exist programs that are valid programs in multiple programming languages. See https://en.wikipedia.org/wiki/Polyglot_(computing)
I agree that XML is too verbose for many common purposes -- but control characters is not the way to go. csv does have problems with distinguishing commas that are in quoted strings and commas that are acting as decimal points. Some of the special characters such as ` are used less often and not used as part of numbers or common numeric expressions like "100%" so practical problems could be reduced if we switched to, for example,
2001-03-25 18:58:03`9,1`0`sunny, with cloudy periods`8,1
but any printable character is going to need to be escaped somehow . Would people be willing in practice to switch over to ` delimited files? I doubt it. Comma delimited has success because comma is commonly used as a list separator in many human languages. Semi-colon delimited has some success for the same reason. But ` has no history related to lists, only history related to quotations. I suspect you would have more success convincing people to use ` as an escape-char in csv files, such as
2001-03-25 18:58:03,9`,1,0,sunny`, with cloudy periods,8`,1
Humans can (often) adapt to "this comma does not indicate the end of a field" by context, but of course "context" does not thrive nearly as well with machine processing of files. Correct machine processing of data needs more things spelt out explicitly... making for more verbose files that are less palatable to humans.
Correct machine processing and correct indexing of data files quites meta-data... standardized meta-data. At which point we get into XML / SGML again, and other representations that attempt to solve the same kind of problems that XML / SGML try to solve. Human-readable, structured information, information made explicit. Good ideas in theory, but not how humans like to operate...
Heck, it is hard to even get the same person to be consistent on date formats over extended periods :(
Richard K Collins
Richard K Collins 2021-3-26
@Walter Roberson
I am going over your answer again. Trying to squeeze in MatLab among other things I have to do.
We are actually talking about human-computer communication. Or rather human to the humans controlling the development and operation of software (more and more that many people use). I am trying to document the social and economic and technical impacts of the Internet. Particularly on the human species. But also on all things. I have looked at as many formats as I can, and experimented with interconverting them. I got a lot of help from my brother, Clif, who spent his life converting GIS and large databases for companies and agencies. He and I have hammered on core issues for the last ten years almost daily. He wrote a Javascript native compiler and we have constant arguments and discussions about what languages to include. I keep saying that we need to include all the commonly used languages in a universal editor that gets used everywhere. I say, just have the human say "Now I am going to be using Fortran for a while" or "this is going to be mostly Visual Basic" or "try treating this as assembler". Or this is a binary format from a geophysics groups, try some of those first. But probably by having common actions as tiny icon buttons and just clicking or dragging or saying. During exploration the needs are different from curation and merging, are different from simulations and calibration, ar different from free running preparatory searches to gather basic information. Watching a human working at full speed, it might be hard to separate the classes of actions, but it can be done. Humans are good at some things and not at others. It varies with the human, their age and interests. A good intelligent algorithm (IA) can do a good job of helping a human find things, and get through life. Memorizing the things they need. Applied to work that means remembering all the textbooks and stuff needed for a professional or technical or general degree. The humans are not hired for their direct memorization, but how to find and move information around, discover new connections, and quantify and calibrate algorithms.
This editor is better than many, but it lacks part of the things I would expect from a universal editor. Before I quit talking about human-computer relations, I should say that a common waste on the Internet is large sites and groups of sites where everything was written by hand, So there are hundreds of different ways that content from humans is accepted, and almost as many from computer sources. Standardize that and a site collapses to layouts, content of various known or partially known forms, and humans moving things around and clicking where appropriate. The best in the world editors could be a long list, but I want to set a basic standard for the Internet where anyone can use their favorite and it just substitutes whereever one is needed. We do this now with people choosing their own editor. But treat all webpages as layouts on displays, and it can be simplified. Whatever tools the persons need, they are part of the hardward, operating system and tools installed locally, or globally, and those standard parts talk to other nodes. Some headless nodes (computer-computer) and some human in the loop nodes (human-computer), some human monitored nodes( computer-human), and some human only nodes (human-human). This categorization might seem trivial, but it helps me to more quickly separate out complex situations. I have been reading the Internet in as big a chunks and as deep as I can for 23 years now, and had been building sites, including a global site before that. I am just trying to see if I can add "the value of human time" to the bookkeeping and accounting for software developers who most often now give zero weight to humans they serve.
A software tool is only as good as its authors and support network. I am looking at MatLab because it handles a lot of technical information. But it is small group and easily eliminated. The world is not kind. Large numbers of people depend on something, the company gets bought out and the software tanks. You have seen it many times. These open groups are volunteers or sponsored, they usually get taken over if they get good at anything. If you are lucky the new management cares about what the groups were already doing. But that almost never happens. So I am memorizing the syntax enough to get by, learning the groups and ways of having to ask for help enough to get by. I have a lot of rather unique experiences and I have spent a couple of decades, about 100 hours a week, trying to test ideas and understand the whole of what people are doing on the Internet. But I see groups who could do things that ought to be done for everyones' benefit, and I am too old and tired to convince them to change. Covid is like that. I know exactly several problems across the Internet that are interferring with all websites and all those (HH, HC, CH, CC) transfers. The biggest is HH when it is going HCHCHCHCH. I am trying to ask all site owners to help (for covid and research genearlly) to collapse those long strings between groups out to millions or billions of humans to HHHHCHHHH, I cannot write it in text (but I don't have a diagram editor or a draw program or a live animated 3D editor to show you the process.
I am looking at MatLab because they did a decent job on a few problems. How many time have I written from scratch the matrix library, the vector and transform libraries, the string parsing libraries, the save and retrieve libaries. You know the drill. The language dies and you need basic things. You cannot use you old tools because they have been made obsolete. There were no standards, so many of the old formats and tools, that worked extremely well, won't run in the new environments.
I don't want the whole of the human species to be charged what might be serveral times the global GDP just because a few companies were given Internet monopolies, or because "business as usual" is not fair to the human species, while it might benefit a tiny tiny few.
I could not find a > for code editing. There is a light gray CODE at the top, but it is a useless categorization of the icons. Those words should be in the hovers. The help won't open in a separate tab. I have lost months, cumulatively, from pages that wipe out what I am working on when you press help. Things are that bad on the Internet. I average about 30,000 page views a month, and go much higher than that when I am working on a difficult and systemic problem. I deliberately try all entry points on a site and often check if the human-computer and computer-human nodes are all talking to each other. But the usual is [ HCCC, HCC, HC, CCCH, CCH, CH ] and none of the central Cs are talking to each other. It generates nice income for the support groups. If you have a human manager whose incentives are tied to number of staff managed, or total budget, or number of calls, that is a prescription for what I call "Internet pathology".
I think I am going to add a basic process editor to my list. For schedules, circuit diagrams, process descriptions, business processes, program and data flows. We were trained in that in the old days and there used to be good tools for that, and some good program language to diagram and diagram to program tools. But things so profilerated and every young person or group wanted to try their own, it got to where nothings is stable or permanent. How many large corporate merger and acquisitions are simply not possible because they use incompatible and costly to translate information systems and methods. Multiply that by millions to get to the scale for the whole Internet. I have hacked at those global problems for 55 years total, and the last 23 years full time. I see some things that might help with covid. But I am just one person. I am seriously thinking of just stopping. But when I was asked (after a million people had died) to set up a famine early warning system (FEWS.net), I wished they had asked me a year earlier, or even 6 months. I am that good in finding the data streams, remapping the variables and interfaces, finding what the users have in their backgrounds that they all can use, finding the training for people to fill in the gaps, finding what the people are supposed to be doing but cannot and preparing for that, writing simple to use workstations for everyone. I call it "not hard, just tedious". But I had one programmer and a small staff when I set up the central Economic and Social Database for all those international datasets, and I had a dozen people and several subcontractors (NASA was a data feed, NOAA, UN, all country embassies, State and everyone in every country. and on and on. Once a basic workstation is available on the Internet, it is not hard to add new groups and people. It is just tables in databases. Now it is hard because all the sites are so messy and hand written. But my plan is to give everyone involved a basic workstations. The ones who need mathematical symbolic equations, simulation tools and modeling tools. I thought I might be able to adapt Matlab, but now I am not so sure. It is so limited in parsing and managing large numbers of things. For the ESDB I used SAS because it was available. But they are nowhere near the same. I see it again and again. A great software company and dedicated community gets taken over by marketers and people who do not share the original vision and all the people who depended on that are hurt.
Even if I am rather tired (up at 3 am to start my day) I am good at what I do. I am a Director almost always. I see what is needed, how it all fits together, and what each person needs to get their job done. And I can keep the whole of billions of people and their individual needs in mind because I spent decades working on problems where that is the main criterial for doing something in the first place. The global economy can expand a hundred fold, if solar system colonization is done as a global project. The economy of every country, including the "developed" countries can come to a global equilibrium for all countries evenly covered at about 4 to ten times current global levels, even if there is no solar system colonization. It is not hard, but it takes humans working with computers on a global scale. Since it is virtual, countries cannot as easily be excluded. But I have seen the worst of human behavior as I trace economic and social and financial and technical and scientific and job and education and cities and countries and groups and goals and projects and all types of human endeavors on the Internet.
I am starting with education, which is humans exploring the world through direct senses and sensors, and exploring what has been codified and passed on from before. Because I read hundreds of complex papers and dissertations and studies across all field, I want to add that to the toolkit, so no one need blanche at picking up the most complex "paper". They will know that it can be reengineered (most of the time) back to useable models, or usable data. At worst they use it as a rough map for adding what else is found. Since every topic on the internet is duplicated roughly 10,000 times, if you are patient and persistent enough then those old papers and books and websites can offer clues to how things fit together and what is important. I don't know if all the old stuff will be translated. It is so easy to use statistical computer methods (algorithms) to discover ways to predict and model the behaviors of any data stream or combinations of data streams. The AI methods I dislke because they too often use smoke and mirrors and marketing. And if you dig into their core, they are looking at large training sets and samples of data to discover basic variations that correlate with things in the real world - that correlate human sensory data with computer sensor data. Again HH, HC, CH, CC. To begin. They are all mostly statistical, linear algrebra and basic and sometimes more advanced constrained optimation toolkits. Often just one or a handful.
So if there is a button or tiny thing, or a series of clicks and drags and typing I need to perform, some magic chant or dance, tell me how to open the live editor from here.
There is a Live Editor link at the bottom. I looked at it. and it goes to yet another spash page that talks generically and market-speak about the wonders of Live Editor.
I finally found the [>] symbol. It is so light gray on gray that I could not see it. And you have to hover it to discover what is means. And it says, "enter a line of code" and I have almost never in my life written just one line of code. I breathe code, I express code, I paint with code, I fill whole canvases with code to sketch real things in the real world. So if someone offers me one line and a keyboard shortcut, rather than an inline or connected development tool, I really don't want to even get started. Having been down so many hundreds of thousands of pathways that lead to dead ends, I hate doing it again.
this is a line of code in a different font that I can barely read, in colors on a gray background, with no word wrap, and
function Guest(a,b,c);
end
It works, but no neat exit. The box should be outlined, the box show be undockable so I can work full screen.
The box should have its own tool panels and they should be dockable and movable and
the layout saveable. If I need libraries, they can be shown with visible lines or links
I bet the people writing this editor are not working with the layout paradigms from oal
the other editor groups. What a pain. it works, but it is not enough. Do I wait
another few years for them to get it right? I am not going to do it for them.
"Clicking that [>] thingy again" did not get me out. No way to change the color of
"quoted strings". No database tools, no object inspector. Those might somewhere else
always somewhere else, more pages, more page views, more complete context changes
Alt enter. So you have to memorize an exit magic word. Remember that cave game, Adventure. We are still exploring dark caves and pathways - with obscure commands, magic words and hidden treasures. I want a map, a light, and something worthwhile. Trying to guess this groups editor conventions is wasting me time, and the time of every person using the Internet. The HC and CH paradigm has been hacked at, but it needs to be done on a global scale so everyone can have the current best in the world tools if they want them -- without cost, as part of the Internet access standards I am trying to write. Alone it seems.
Richard Collins, Director, The Internet Foundation

请先登录，再进行评论。

更多回答（2 个）

Jan
Jan 2021-3-22
编辑：Jan 2021-3-22
"There are probably ways to read a text string from the web and pass it directly through CSV tools rather than looping as I had to."
Correct. You can replace:
N = length(Lines) - 1;
values = zeros(1,N,'double');
for i = 1:N
values(i) = str2double(Lines{i+1});
end
by:
values = str2double(Lines(2:end))
"Putting a variable into the figure(k) does not seem to work." - for other like me it does work directly by copy& paste of your code. Maybe teher is a confusion with Live Scripts or the local exectution.
The meaning of the line "clear figure;" is still not clear. I assume you should omit it.
If Matlab is not useful fpr your needs, use another programming language. I'm satisfied with it for many problems, but of course not for all.
  1 个评论
Richard K Collins
Richard K Collins 2021-3-25
@Jan Thank you, that is more compact. I will use that.
My concern for the whole Internet is that the conventions used for indexes and objects in the MatLab language are not universal. And, among the few dozen groups who have tried to write simple ways to index, MatLab does not provide translations.
By the way, this question of how to skip the first element of an array is hard to search - in the MatLab help on Windows (where I am working and testing), in the MatLab online documentation and community help, and on the Internet as a whole. The difficulty being that is becomes one of those, "everyone ought to know this" sort of problems that make efficiencies in a small group, but delays and wasted time on a global scale. MatLab seems to have only a few million users globally. I asked "How many people use MatLab" and got 4 million. My guess would be a few times that number tried it and gave up from the search costs and memorization burden.
I wish there were "Like" and "Donate" or "Support" buttons next to each thing contributed. And beside each person and their suggestion or contribution. With no way to know what people are doing, and how well, it is hard to find anything on these MathWorks pages. It is almost two millions pages of text, with no overall curation, indexing, de-duplication (a much more compact respresentation), or index by value. MatLab cannot afford world-class indexing, and has not tried by itself. Relying on Google and search engines is iffy, since they make their money by clicks and not by value provided to society.
What is a suitable value for your time? If I spend an hour searching, then find an answer here in a few minutes, my best strategy would be to come here first, get a quick and precise answer, then donate to support the ones here giving their time. The problem being it taps a human brain for that information, which won't work if there are more than one person asking at a time. And, I see it every day on the Internet, as soon as money becomes expected, then the processes get gamed to make things more complex and more delayed to raise the price and income for a human worker.

请先登录，再进行评论。

Richard K Collins
Richard K Collins 2021-3-25
编辑：Richard K Collins 2021-3-26
The comments and answers were helpful. But my original question still has not been answered. The figures still stay in the output area and are not coming up in a separate window that I can move around. I would like the Output Area to be a undockable. It does not even have the standard capabilities. The controls to "output on right", "Output inline", "Hide code" are too limited. If all areas are treated equally, then the Output Area would be a separate area that could be docked, moved, resized, minimized. Particularly, it would be a separate window that I could actually work with, no shoved indiscrimantly into some second class area part of something else.
Here is my version as of 25 Mar 2021 about 2:40 pm Houston Texas time:
This is a horrible editor. It should recognize code and make it alive. Why am I having to cut and paste code into text - it is not words and text, it is a live MatLab program that works. Or would, if you just treat it properly. That business of the A = S + '\n'; not working is just bad. I tried fprintf(S + '\n') and it would not work. The people writing the parsers have this one possibility per situation mentality. If the intent is clear enough to offer only a few possibilities, then let the person choose, or just warn and try your best. The world is more using one time code that gets later improved. During development be more flexible. When you program as much as I do, my time is more valuable than syntax. One syntax programmer can make rules that waste millions of hours for users. Or save them. At $32 per hour for the current value of human time on the Internet, save millions of hours and it should show up in your bonuses and earnings and stock options. Or just fame.
----------------------------------
Url = 'http://service.iris.edu/irisws/timeseries/1/query?net=AV&sta=DLL&cha=HDF&start=2011-01-01T00:00:00&end=2011-01-01T00:03:00&scale=AUTO&format=ascii1&loc=03';
Text = webread(Url);
Lines = strsplit(Text,'\n');
N = length(Lines) - 1;
NoteNum('There are %d records', N);
Note('Please wait. Extracting values');
values = str2double(Lines(2:end));
NoteNum('Done getting %d values from the text.\n', N);
Layout(values);
function Layout(values)
set(0, 'DefaultFigureVisible', 'on');
N = length(values);
figure('Name', 'My 5 plots', 'NumberTitle', 'off');
Plot1(2,3,1, 1:N-1, diff(values,1),'First Plot');
Plot1(2,3,2, 1:N-2, diff(values,2),'Second Plot');
Plot1(2,3,3, 1:N-3, diff(values,3),'Third Plot');
Plot1(2,3,4, 1:N-4, diff(values,4),'Forth Plot');
Plot1(2,3,5:6, 1:N, values,'Raw Data');
end
function Plot1(rows, columns, Place, Xs, Ys, MyTitle)
subplot(rows, columns, Place);
plot(Xs,Ys);
grid on;
title(MyTitle);
end
function Note(S)
fprintf(S + "\n");
end
function NoteNum(S,Num)
fprintf(S + "\n",Num);
end
  6 个评论
显示 4更早的评论
Richard K Collins
Richard K Collins 2021-3-27
@Walter Roberson
Thanks. But what I was referring to are the example scripts, documentation with examples and many other pages on MathWorks.com and elsewhere - no way to know who wrote the page, if they are actively working, or just randomly posting things. If they have backgrounds that warrant reading what they posted. This comment springs from going through what were supposed to be examples of using MatLab -- that I know are just wrong. Poorly done and that will give the wrong answers. Anyone following that is going to get into trouble with real data. They were slinging an app at a problem. And then I looked at a few more and found the same thing. The authors (inside MathWorks? other people?) were not documenting the context and and where the method is applicable.
I literally have to deal with problems like how to train billions of people. Tiny mistakes can waste billions of years of human time. And I can measure these impacts, and see where these kinds of mistakes are made. So it is not just academic. I spent years modeling economic and social impacts on a global scale. So it bothers me to find large groups using bad methods. It is kind of like finding a bad teacher who is affecting many students - giving them the wrong tools and methods. The MatLab community might be self-correcting. I am going to look. But there is a lot of material on the Internet. It might be out of date and they will say, "Oh, that is old stuff, we know better now". But the community is not cleaning up.
On Wikipedia it is common for well written and correct articles to be "rewritten" and come out worse. Or to take a well covered topic and spread it out into dozens of fragments. I know why it happens, because anyone one can change what has already been done and there are no checks and balances, except for random people. The same things happens in the DNA genealogy community, where there are tens of thousands of copies of the family data for a person. Or where they try to make one, and only one. There are rules and methods to stabilize those situations. But it is not easy.
People use MatLab to solve problems, but there are a lot of people trying to use it to learn methods. If they post their work, and it is wrong, and you cannot locate them to gently suggest they change, That hurts anyone who just sees it as an example -- "of course, it is published by MathWorks, it must be right". Do you get what I am hinting at? There is not even peer review. Or does not appear to be. That is why I said I will see if the community is self-correcting. I am working on dozens of other things, and only have a few hours a day to devote to the MatLab community. But I want tools for myself for some things that need to be simplified. I find large communities (tens of millions) where they are wasting a lot of time using inefficient methods. The usual thing is someone finds that and then wants to make money from selling a bit of knowledge. I inject the answers or ideas or memes and they automatically self correct. All these global scale communities have internal and personal needs. But they also have a purpose and organization, usually subconscious at a social level, that causes global changes that can be tracked.
You have heard that story about an idea bubbling up. "If Newton had not lived, the ideas would have been there anyway"? I can quantify that now, because the speed and interconnectedness of the Internet allows for finding groups that are all working on a common problem. I have done that kinds of looking almost 20,000 times in the last 23 years. And it is almost universally true that if you can think of something "new or better", there are already tens of thousands who have worked on all the parts of it, for many years. But they have not found each other and are not working together.
So, with MathWorks not formally tracking contributions to the MatLab community, nor organizing the Internet material, nor checking to see what the rest of the world is doing - you have a community of a few million people acting as a blind organism. And prey to more adept and agile groups. Or just a waste. You can see why I make short comments "that is bad"? I have this amount of stuff in mind and each piece can take days or months or years to map and clarify. I work 120 hours a week usually and have for the last 23 years since I am working on this full time. I almost never forget anything large. I thnk it is important to solve these global problems, to find ways for people to work in groups of millions or billions for the good of the whole human species. But I am forced to use tools and a weak human memory for much of it, that are less than perfect.
There are millions of calculators and models running on the Internet now. It is one of the fastest growing items. They all use different methods, different referencing methods (they use equations and calibrations from somone else), and often are wrong or broken. I was thinking that MatLab or Mathematica or another tool could replace all or most of them. Now practically a good toolkit would be picked up by those online authors (writing toolkits for that sort of thing from scratch) and over time the system would heal itself presenting a few interfaces and human-computer paradigms to the whole of the Internet community - rather than the mess of non-standard and time wasting methods now. The advantage of just a few toolkits is that it would be possible to merge all the models and have a global and complete set of all the phenomena, engineering, mathematical, scientific, and other profession rules and guides avaliable to everyone. And it would be possible to check and cross check all of them. I can find the core models and put them together. What I cannot do is get al the bad copies and partial one to refer to the best-in-the-world few. It is a delicate balance. Too few and monopolies and manipulation always occur. Too many and so much time is wasted in finding things, it is easier to develop things from scratch. Right now massive duplication and parallel reinvention is going on - in every part of human life and society. It is wasteful, especially when billions of people get trained in the wrong things, or are not prepared, or simply cannot find answers to things that have been well known and well understood for hundreds of years or decades, or just that last few months, but in thousands of group globally.
Covid is the the same, just it only has a relatively small number of people and groups using quantitative methods and models. So I have to track data flows and human-based rule processes and procedures. That is a simple Internet problem, but socially I cannot force people to change. Even with lives at stake. No one listens, and too many people benefit from things the way they are.
Richard Collins, Director, The Internet Foundation

请先登录，再进行评论。

类别

在 Help Center 和 File Exchange 中查找有关 Startup and Shutdown 的更多信息

标签

产品

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by