Automated fraud detection is fine. But acting on a number which is basically a statistical guess, without some follow-up manual review is plain irresponsible. The credit industry does this and regulation differs by country. For public services, the justice standards should be different. I don't think we are hearing the whole story though. It is hard to believe that a 20% error rate is accepted by European standards of fairness.
> The product did function on some level however below advertised promises.
Pardon my ignorance, I did not read anywhere that the product worked at ANY level. Please tell me a single innovation that Theranos produced that could be scientifically verified. I was under the impression that it was all hush-hush, smoke and mirrors.
> They were able to sell it to investors
Investors are not scientists. Did they manage to convince scientists that they had ANY breakthrough, even one.. lets completely ignore profitability of this supposed breakthrough and viability. What exactly did Elizabeth Holmes discover or invent?
All I saw was this argument: These investors were rich and must therefore know what they were doing. They must have checked the science behind it.
No, the strawman is pointing to scripting languages like Python and Javascript and saying they won't cut it for writing an OS. Of course, they don't. Who is even suggesting?
Again, I am not arguing against use of C for kernels. I am arguing against using C style programming (trusting humans with transparent memory access) unless it is absolutely warranted.
I am also not getting the sense that you properly tried Rust yet.
It does do safer constructs with no extra cost. It does safe without compromising quick.
> Any language is an expression of your intention. If your intentions are poorly thought out, planned, described and executed, that's a fault common to all languages and little can fix that (expecting the "grammar" or "vocabulary" of a particular language to save you in that instance is not sensible).
I completely disagree. Tools (languages in this case) have a great bearing on your ability to express yourself. There is a huge amount of psychology research on this general area; and this is a psychology issue, not a CS issue. Intentions aren't enough. That is the whole point of programming language research.
> All things aside - trusting SOFTWARE to enforce security is a nonsense. When software wants to access out-of-bounds, it only has to change a number. > We need to design *hardware* to enforce memory security. Things like "per-process/page encryption" are starting to do that for us
Of course. I completely agree that the current security scenario is broken and these are all band-aids. My point is that because we don't have that utopia, we do need to compensate until we get there.
This is mostly a strawman. None of those languages were ever considered real candidates for OS or driver development, except Rust (which generates the same machine code with safer high-level constructs).
As much as I dislike C, I am fine with it for kernel and driver development. C can be mostly safe IF you pour a lot of human resources to review it. People always assume they are better at things than they actually are... and with C that is more of a problem.
Not all that makes an OS and presents vulnerabilities is kernel and driver code. Avoid C whenever you can (especially network facing code, bottleneck is mostly IO anyway), but otherwise keep it. If you are not writing hardware level code, one perhaps needs to reconsider its use entirely. Keep it legal and rare.
Rust is much harder to learn than C. That makes it a not-yet-ready replacement, not performance costs.
I had hoped it would succeed more than it did. But somehow, it never quite caught on. SageMath did not do for Python what Anaconda did.
Perhaps, it was (still is? - did not install in a while) the fact that it was distributed as a VM image for Windows. With Anaconda, it was easier to setup than plain Python; not even admin rights were needed. The package management was familiar. And it was just Python. SageMath was its own system and was intimidating for new users.
It also targeted scientists (and I hardly knew any that switched from MATLAB to SageMath), which are a much smaller demographic than data analysts.
Most of the new users since perhaps around 2012 came for the data analytics side.
IPython, NumPy, SciPy had been around for a while, but with maturing Jupyter, Pandas and TensorFlow/Keras, it really caught on. Other NLP and Machine Learning libraries probably helped too.
My use of Python today is completely different from how I used it earlier, nearly two decades ago, when it was mainly seen as a better Perl, back when Perl was THE scripting language. Now it is seen as a better MATLAB or a better R, even though the base language isn't itself vectorized as the others. The language and the standard library didn't improve much towards this. It was mainly the third party libraries that emerged and matured.
Speaking purely from a language standpoint, Julia has all right features for the analytics side, but the scientific community is right now with Python.
I don't recall ever worrying about breaking an old Java app on a new JVM.
It is easy to have multiple JVMs on a single machine and have any app use any of them. Sure, you will have to manage updates for them of course, which isn't much of a difference when you have a legacy JVM, out of support period, for a specific app anyway.
Java, in its early days, did deserve the scorn. It was a hog, pretended that the features it lacked were a good thing, only to get them later. The language also used to be painfully verbose and the frameworks were usually over-engineered.
Conquer? Go takes no more than a day or two to learn, obviously longer to be fluent. It is the least challenging language today. It lacks several features that we take for granted elsewhere and pretends, just like Java did, 20 years ago, that you don't really need them. Give it another 10-15 years, maybe.
Earth-1218. Which universe is this where Javascript is pervasive on the Desktop? I have a couple of Electron based editors and a couple more communication apps. That's about it. That is no more penetration than Java RCP apps of the last decade.
Go is adding an average of 5 modules per day. That is quite low. Compare that to 118 for Python and 128 for Java. Even Rust is growing faster.
Much of Python's recent resurgence has been in "data science". Go community barely even attempts to solve anything here. It's appeal to Python programmers is in a rather narrow class of problems. Static binary? Against Java 11 with the upcoming GraalVM, I don't see much in favor of Go.
> Python isn't dead, but it was gravely wounded by the Python 2/3 debacle.
That is absurd. Python is more popular than ever. Take a look at historical ranks. Python transition was planned for a long time (Python 3K) and was extensively discussed. The transition was expected to take a long time when introduced and was managed as such. Its community managed breaking changes better than most languages.
You might have had that case with Perl 5/6. It lost all its clout in the transition.
Both Java and Python have grown tremendously. It was just that no one saw Javascript to have the resurgence it did in the meantime.
The problem is not so much that this content is available by search on demand; it is that it is often amplified and promoted on the landing page if you watch a few videos of this nature, initially only out of curiosity (yes, I am aware that you can tell Youtube to not show similar content). I am now in a developing country. All the landing page content is utter junk and the people (especially the low information users who know Internet only through their phones that form the majority) know no better. People seem to think whatever Youtube content is popular or is promoted is probably true. Most of these don't seem to actively seek junk, but passively consume whatever Youtube says is right for them. Internet is not enlightening them; it is doing the opposite, reverse of the early promise of the Internet when it was all academic.
This is a hard problem to solve. We all know the issues. The real solution is ultimately better education, but the current recommendation algorithms are amplifying ignorance. Their goal is purely to promote engagement. Google would not exist without the Page Rank algorithm and while that is often gamed, it does mostly work. They need to come up with an equivalent of the Page Rank for Youtube landing page.
Whatsapp seems to be way worse with respect to propagating junk in developing countries. The new effort that limits forwarding to 5 is at least a start.
> supply of people who know something like R well is smaller than people who have a good understanding of statistics
Learning R is trivial. There are no difficult concepts at all. It is just a matter of getting used to it. Learning to think statistically isn't. Not everyone is even built for probabilistic thinking.
Mathematical skills are always at a premium over coding skills.
Data Science is not just mining social network data. It applies to any data. The data science I am in contact with is all about publishing papers and little about monetizing anything. And this is stuff being done for ages, except that this is a new umbrella term since there are a common set of algorithms and technologies that everyone finds useful.
> I think my main objection so far to Rust is that I can't just download a tarball, unzip, set up some environment variables, and run. It seems that you have to run an installer, which I really dislike! But that's me.
> I think my main objection so far to Rust is that I can't just download a tarball, unzip, set up some environment variables, and run. It seems that you have to run an installer, which I really dislike! But that's me.
From what I remember (it has been a while since I ran rustup-init.exe), Rust installer does not need admin privileges and does not have a registered uninstaller. All it seems to do is download, unzip and just adds to path (No new environment variables by default. You set them if you want to override).
Not true. MOST of what was claimed from religious frameworks was wrong.
It is quite uncommon for old religious claims to be held up by science. It is just that it makes news when it rarely does.
This is also obvious if you give it even a minute of thought. Since Truth is singular, all religions must then make identical claims. But they do not.
This is however true of Science. There are no multiple versions of Physics or Chemistry, but you can have an infinite variations on religion.
> Citation: religious societies survived the evolutionary selection process. Non-religious did not.
They survived on social and cultural grounds, not ontological grounds. Survival fitness requires very little Truth. Humans aside, least complicated creatures survive better than more complex, not the ones that "know" most.
It is generally reproductive fitness, not intelligence that contributes more to survival. Religions with an active or aggressive propagation component survive better than "more true" or at least less false religions. The job of the priest/theologian/apologist is partly to deny deviance from known reality and that the texts were never wrong, even against overwhelming evidence.
Create a sophisticated philosophical religion with no proselytization element and it has no chance against an emotional one that appeals to convert as many as you can to its fold through power (wealth or violence).
Asserting health benefits with no data is cheap, even when occasionally correct. Showing with data and analysis is actually knowing (science). There is still quite a bit that needs to be investigated with regards to diet. The basics are well understood. The long term effects are less well understood since that research is harder (to control).
> You seem not to belong to the "western world" anymore with your retarded science.
The only "retarded" arguments seem to be yours. What western world? I have not cared to divide the world into East and West here. What makes you think I am from the "western world" in the first place? I don't remember declaring my location. Where I am from/at, religious fasting for the public usually simply means eating lightly for dinner, abstaining from the main staple diet, not complete caloric abstinence, which is left to monks. In the data-centric modern world we collectively inhabit today, east and west is irrelevant. One measures. One verifies. Biological realities have little need for philosophies and theologies.
> in most religions fasting was done during periods were they had no food anyway. > were fasting was artificially introduced, because they simply copied it from Christians. > But they were smarter, as they had irrigation and food: they were allowed to eat after sunset. Hence: Ramadan is actually a big party.
What does all that have to do with my statement that religious fasting has little to do with health? Culturally, the point was an offering or sacrifice, that you cared about it enough to endure some discomfort and participated in a collective activity which perhaps contributed to some degree of social cohesion.
Most of what had been "known" for millennia was wrong. That is the whole point of doing science. But even a broken clock can be right twice a day. The scientific process for doing dietary studies though is less than stellar today. But that can be fixed. Science is working, not "struggling". Never before in human history has so much been uncovered in such a short time.
> "religions" are just archetypes of human biology pattern
If so, we would not have so many religions. Religions occasionally are codifications of natural human behavior - social, more than biological. Other times, they deny natural human behavior.
The fasting traditions in religion didn't make health claims. Fasting causes acidosis, which leads to mild euphoria. People also fast to promote the odds of transcendental experiences. It was an act of discipline.
But no religion had the liver, let alone circadian clocks in mind. I wouldn't say religion *knew* fasting. It used it for an entirely different purpose.
Slashdot Mirror
Automated fraud detection is fine.
But acting on a number which is basically a statistical guess, without some follow-up manual review is plain irresponsible.
The credit industry does this and regulation differs by country. For public services, the justice standards should be different.
I don't think we are hearing the whole story though. It is hard to believe that a 20% error rate is accepted by European standards of fairness.
This is a great post.
I wish Slashdot had more thoughtful posts like yours.
> The product did function on some level however below advertised promises.
Pardon my ignorance, I did not read anywhere that the product worked at ANY level.
Please tell me a single innovation that Theranos produced that could be scientifically verified.
I was under the impression that it was all hush-hush, smoke and mirrors.
> They were able to sell it to investors
Investors are not scientists. Did they manage to convince scientists that they had ANY breakthrough, even one.. lets completely ignore profitability of this supposed breakthrough and viability. What exactly did Elizabeth Holmes discover or invent?
All I saw was this argument: These investors were rich and must therefore know what they were doing. They must have checked the science behind it.
No, the strawman is pointing to scripting languages like Python and Javascript and saying they won't cut it for writing an OS. Of course, they don't. Who is even suggesting?
Again, I am not arguing against use of C for kernels. I am arguing against using C style programming (trusting humans with transparent memory access) unless it is absolutely warranted.
I am also not getting the sense that you properly tried Rust yet.
https://ruudvanasseldonk.com/2...
It does do safer constructs with no extra cost. It does safe without compromising quick.
> Any language is an expression of your intention. If your intentions are poorly thought out, planned, described and executed, that's a fault common to all languages and little can fix that (expecting the "grammar" or "vocabulary" of a particular language to save you in that instance is not sensible).
I completely disagree. Tools (languages in this case) have a great bearing on your ability to express yourself. There is a huge amount of psychology research on this general area; and this is a psychology issue, not a CS issue. Intentions aren't enough. That is the whole point of programming language research.
> All things aside - trusting SOFTWARE to enforce security is a nonsense. When software wants to access out-of-bounds, it only has to change a number.
> We need to design *hardware* to enforce memory security. Things like "per-process/page encryption" are starting to do that for us
Of course. I completely agree that the current security scenario is broken and these are all band-aids. My point is that because we don't have that utopia, we do need to compensate until we get there.
This is mostly a strawman.
None of those languages were ever considered real candidates for OS or driver development, except Rust (which generates the same machine code with safer high-level constructs).
As much as I dislike C, I am fine with it for kernel and driver development. C can be mostly safe IF you pour a lot of human resources to review it. People always assume they are better at things than they actually are... and with C that is more of a problem.
Not all that makes an OS and presents vulnerabilities is kernel and driver code. Avoid C whenever you can (especially network facing code, bottleneck is mostly IO anyway), but otherwise keep it. If you are not writing hardware level code, one perhaps needs to reconsider its use entirely. Keep it legal and rare.
Rust is much harder to learn than C. That makes it a not-yet-ready replacement, not performance costs.
I had hoped it would succeed more than it did. But somehow, it never quite caught on.
SageMath did not do for Python what Anaconda did.
Perhaps, it was (still is? - did not install in a while) the fact that it was distributed as a VM image for Windows.
With Anaconda, it was easier to setup than plain Python; not even admin rights were needed. The package management was familiar. And it was just Python. SageMath was its own system and was intimidating for new users.
It also targeted scientists (and I hardly knew any that switched from MATLAB to SageMath), which are a much smaller demographic than data analysts.
Most of the new users since perhaps around 2012 came for the data analytics side.
IPython, NumPy, SciPy had been around for a while, but with maturing Jupyter, Pandas and TensorFlow/Keras, it really caught on. Other NLP and Machine Learning libraries probably helped too.
My use of Python today is completely different from how I used it earlier, nearly two decades ago, when it was mainly seen as a better Perl, back when Perl was THE scripting language. Now it is seen as a better MATLAB or a better R, even though the base language isn't itself vectorized as the others. The language and the standard library didn't improve much towards this. It was mainly the third party libraries that emerged and matured.
Speaking purely from a language standpoint, Julia has all right features for the analytics side, but the scientific community is right now with Python.
It would be funny... if any of that is anywhere close to being true.
My first Java dev machine was a 400MHz single core with 64 MB of RAM.
My first Eclipse machine was a P4 with 256 MB of RAM.
Java and Java IDEs are a lot more efficient than these JS runtimes and JS IDEs today. Good luck doing Android Studio over Atom.
I don't recall ever worrying about breaking an old Java app on a new JVM.
It is easy to have multiple JVMs on a single machine and have any app use any of them. Sure, you will have to manage updates for them of course, which isn't much of a difference when you have a legacy JVM, out of support period, for a specific app anyway.
It is a click bait article alright.
Java, in its early days, did deserve the scorn. It was a hog, pretended that the features it lacked were a good thing, only to get them later. The language also used to be painfully verbose and the frameworks were usually over-engineered.
Today, most of those things stand fixed.
> I leave it to the new kids to conquer it.
Conquer? Go takes no more than a day or two to learn, obviously longer to be fluent. It is the least challenging language today.
It lacks several features that we take for granted elsewhere and pretends, just like Java did, 20 years ago, that you don't really need them. Give it another 10-15 years, maybe.
Earth-1218. Which universe is this where Javascript is pervasive on the Desktop? I have a couple of Electron based editors and a couple more communication apps. That's about it. That is no more penetration than Java RCP apps of the last decade.
> Many talented folks who jumped ship from Python are now Gophers.
You might get that impression if you read some enthusiastic bloggers. But the stats don't seem to reflect that at all.
http://www.modulecounts.com/
Go is adding an average of 5 modules per day. That is quite low. Compare that to 118 for Python and 128 for Java. Even Rust is growing faster.
Much of Python's recent resurgence has been in "data science". Go community barely even attempts to solve anything here.
It's appeal to Python programmers is in a rather narrow class of problems.
Static binary? Against Java 11 with the upcoming GraalVM, I don't see much in favor of Go.
> Python isn't dead, but it was gravely wounded by the Python 2/3 debacle.
That is absurd. Python is more popular than ever. Take a look at historical ranks.
Python transition was planned for a long time (Python 3K) and was extensively discussed. The transition was expected to take a long time when introduced and was managed as such. Its community managed breaking changes better than most languages.
You might have had that case with Perl 5/6. It lost all its clout in the transition.
Both Java and Python have grown tremendously. It was just that no one saw Javascript to have the resurgence it did in the meantime.
I wonder if this Program manager ever considered to apply this logic to MS Windows and Unix like Apple has.
The problem is not so much that this content is available by search on demand; it is that it is often amplified and promoted on the landing page if you watch a few videos of this nature, initially only out of curiosity (yes, I am aware that you can tell Youtube to not show similar content). I am now in a developing country. All the landing page content is utter junk and the people (especially the low information users who know Internet only through their phones that form the majority) know no better. People seem to think whatever Youtube content is popular or is promoted is probably true. Most of these don't seem to actively seek junk, but passively consume whatever Youtube says is right for them. Internet is not enlightening them; it is doing the opposite, reverse of the early promise of the Internet when it was all academic.
This is a hard problem to solve. We all know the issues. The real solution is ultimately better education, but the current recommendation algorithms are amplifying ignorance. Their goal is purely to promote engagement. Google would not exist without the Page Rank algorithm and while that is often gamed, it does mostly work. They need to come up with an equivalent of the Page Rank for Youtube landing page.
Whatsapp seems to be way worse with respect to propagating junk in developing countries. The new effort that limits forwarding to 5 is at least a start.
> supply of people who know something like R well is smaller than people who have a good understanding of statistics
Learning R is trivial. There are no difficult concepts at all. It is just a matter of getting used to it.
Learning to think statistically isn't. Not everyone is even built for probabilistic thinking.
Mathematical skills are always at a premium over coding skills.
Data Science is not just mining social network data. It applies to any data.
The data science I am in contact with is all about publishing papers and little about monetizing anything.
And this is stuff being done for ages, except that this is a new umbrella term since there are a common set of algorithms and technologies that everyone finds useful.
> I think my main objection so far to Rust is that I can't just download a tarball, unzip, set up some environment variables, and run. It seems that you have to run an installer, which I really dislike! But that's me.
This might interest you too
https://forge.rust-lang.org/ot...
> I think my main objection so far to Rust is that I can't just download a tarball, unzip, set up some environment variables, and run. It seems that you have to run an installer, which I really dislike! But that's me.
You could start with
https://play.rust-lang.org/
No need to download anything at all.
From what I remember (it has been a while since I ran rustup-init.exe), Rust installer does not need admin privileges and does not have a registered uninstaller. All it seems to do is download, unzip and just adds to path (No new environment variables by default. You set them if you want to override).
By that you mean the carbohydrate based food pyramid that was driven by farm lobbies and not sound science. I get that.
That not withstanding, the basics are still considered resolved some 50 years ago. The basic biochemistry here has not changed radically.
> "Much", and a tiny, tiny portion was wrong
Not true. MOST of what was claimed from religious frameworks was wrong.
It is quite uncommon for old religious claims to be held up by science. It is just that it makes news when it rarely does.
This is also obvious if you give it even a minute of thought. Since Truth is singular, all religions must then make identical claims. But they do not.
This is however true of Science. There are no multiple versions of Physics or Chemistry, but you can have an infinite variations on religion.
> Citation: religious societies survived the evolutionary selection process. Non-religious did not.
They survived on social and cultural grounds, not ontological grounds. Survival fitness requires very little Truth. Humans aside, least complicated creatures survive better than more complex, not the ones that "know" most.
It is generally reproductive fitness, not intelligence that contributes more to survival. Religions with an active or aggressive propagation component survive better than "more true" or at least less false religions. The job of the priest/theologian/apologist is partly to deny deviance from known reality and that the texts were never wrong, even against overwhelming evidence.
Create a sophisticated philosophical religion with no proselytization element and it has no chance against an emotional one that appeals to convert as many as you can to its fold through power (wealth or violence).
Asserting health benefits with no data is cheap, even when occasionally correct. Showing with data and analysis is actually knowing (science).
There is still quite a bit that needs to be investigated with regards to diet. The basics are well understood. The long term effects are less well understood since that research is harder (to control).
> You seem not to belong to the "western world" anymore with your retarded science.
The only "retarded" arguments seem to be yours. What western world? I have not cared to divide the world into East and West here.
What makes you think I am from the "western world" in the first place? I don't remember declaring my location. Where I am from/at, religious fasting for the public usually simply means eating lightly for dinner, abstaining from the main staple diet, not complete caloric abstinence, which is left to monks.
In the data-centric modern world we collectively inhabit today, east and west is irrelevant. One measures. One verifies. Biological realities have little need for philosophies and theologies.
> in most religions fasting was done during periods were they had no food anyway.
> were fasting was artificially introduced, because they simply copied it from Christians.
> But they were smarter, as they had irrigation and food: they were allowed to eat after sunset. Hence: Ramadan is actually a big party.
What does all that have to do with my statement that religious fasting has little to do with health? Culturally, the point was an offering or sacrifice, that you cared about it enough to endure some discomfort and participated in a collective activity which perhaps contributed to some degree of social cohesion.
Most of what had been "known" for millennia was wrong. That is the whole point of doing science.
But even a broken clock can be right twice a day.
The scientific process for doing dietary studies though is less than stellar today. But that can be fixed.
Science is working, not "struggling". Never before in human history has so much been uncovered in such a short time.
> "religions" are just archetypes of human biology pattern
If so, we would not have so many religions. Religions occasionally are codifications of natural human behavior - social, more than biological. Other times, they deny natural human behavior.
The fasting traditions in religion didn't make health claims.
Fasting causes acidosis, which leads to mild euphoria.
People also fast to promote the odds of transcendental experiences.
It was an act of discipline.
But no religion had the liver, let alone circadian clocks in mind.
I wouldn't say religion *knew* fasting. It used it for an entirely different purpose.