I'm not holding anything against it that regard. The simple fact is that I've had two fairly low traffic MySQL databases become corrupted beyond the point of being usable within the last 3 years. The hardware wasn't at fault here (nor was it old or outdated).
I'm not sure what you're doing wrong here, but I think many of us have been running a lot more MySQL databases than that and never experienced corruption. Myself, I have been maintaining on average about 20 MySQL instances spread across 3 different servers for the last 10 years, and have never[1] experienced corruption that wasn't tied to a hardware failure. Sometimes after a system crash I need to do a myisamchk --recover, but even that's rare. I've never needed to do myisamchk --sort-recover, although I understand there are circumstances where this is necessary.
So my experience has been that mysql is, essentially, rock solid. It could be that we're using it for different applications and it works for mine but not yours, but I've also never seen any reports of problems with mysql corruption that cannot be fixed automatically by the tools mysql provides.
[1]: OK, this isn't quite true. There was one instance of corruption that was tied to a linux kernel version that had a bug where dirty buffers were sometimes corrupted prior to being flushed to disk. Hard to pin that one on MySQL either, though.
1) MySQL transactions are built into the table engines, and by default (last I checked, and meaning you don't install innodb, etc), the tables will not be transactional. This means that if you are building an inhouse app, you can trust it more than you can if you are distributing your software. In short, if you are distributing software you can't guarantee that it is running on a system with transactions without a great deal of headache........ The same goes for referential integrity enforcement.
It's easy enough to set up a database creation script that ensures the tables have the necessary support. Sure, if you're working on tables that somebody else has created you can have issues, but in your experience how often does that happen? And if it does happen, how often do you have a choice as to what dbms is in use? In my 10+ years as a consultant I've only ever had to do that once, and never with mysql (because the client's existing system I had to integrate with used informix).
2) Strict mode can be turned off by any application. This means that the more recent data integrity checks cannot be relied upon. This is an issue on both inhouse and distributed software because it adds quite a bit of overhead to the QA process internally, and can add support headaches in software for distribution.
On the whole, this is probably a good thing. If the application is under your control, you can use whichever mode you want. If you're relying on somebody else's application, forcing it to use strict mode when it wasn't written for this environment could introduce subtle bugs. Now, if you were to argue that the _existence_ of these different modes of operation was an issue, then I'd probably agree. But given the existence of the modes (and that's unfortunately a necessity for backwards compatibility reasons) the ability of the application to change the mode to the one it expects is crucial.
MySQL is a good db for single-app databases, where data integrity is not a tremendous issue or where you are deploying a separate MySQL instance on a different port.
Of course, you are describing something like 90%+ of all database applications there. Yes, this book should cover MySQL: it's a very popular database that is perfectly adequate for most uses. Sure, there are applications where it shouldn't be used, but that hasn't stopped it from becoming extremely widely deployed and being used as the database of choice by web developers everywhere (which is probably the target market for this book).
He's overstating the point. The best Li-ion batteries get about 150Wh/kg; the best LiFePO4 batteries get about 120Wh/kg. But, there's a crucial difference between the two that he doesn't mention: LiFePO4 batteries typically survive between 3 and 5 thousand cycles before they noticeably lose capacity; Li-ion, only slightly over 1 thousand. If you're building an EV you get a choice: spend $10,000 on Li-ion batteries and have a slightly larger range, but you'll have to replace them after 4 years, or spend $12,000 on LiFePO4 batteries, get slightly lower range, but have the batteries last about as long as the car does. Frankly, it's a no-brainer.
The claim sounds like hype, since it's energy density, not discharge rate that everyone is trying to ramp up lately.
No, energy density and discharge rate are both problems. Lithium batteries are clear leaders in energy density (at least some of the more unusual varieties like lithium-iron phosphate), but they have very slow discharge. This means that to get the 500A you can easily get out of a single lead acid battery, you need something like 20 batteries working in parallel. And making more small cells is more expensive than fewer large ones. A typical EV power system based on LiFePO4 will want something to be able to provide something like 100v x 200A peak, which means 8 parallel banks each containing about 50 cells. With this new design, that can be cut to just 50 big cells, which will probably halve the cost of production.
Wow, that's a lot of power. Here in America our outlets are rated at about 20 Amps. That translates to roughly 2kW. what a coincidence!
The problem is, though, that drawing 20A/110V will heat up your wiring twice as much as drawing 10A/220V (assuming the same resistance), thus wasting twice as much to transmission losses. Also, the voltage fluctuations you'll see due to changing power demands are twice as large. So higher voltage == better.
Besides, here in the UK we get 240v x 13A = 3.1kW, so we beat you both.:)
Example - dysons when I was selling them had lower engine power than, say, hoover(1200 vs 2000 watts iirc),
Having coincidentally tested my Dyson DC10 with a power meter just yesterday, I can tell you that the draw from mains is 1kw. Not sure what the official rating is, though.
How many years ago was it that we had credible airlines considering banning in-flight battery use?
Actually, merely carrying more than a certain (rather small) capacity lithium battery on a flight is banned by international convention. Most laptops squeak under, but officially you're probably not allowed to carry a replacement battery on as well...
When you hear about knife crime being down, bear in mind that the government have been caught fiddling the numbers to make themselves look better.
Yes. They claimed a 16% (IIRC) drop in hospital admissions with knife injuries, but actually the drop was only 8%. Point still stands, though: there has been a serious reduction lately.
Babylon 5 we appreciated the serial nature but couldn't get past how bad each individual episode was
How far in did you watch it? It gets seriously better about half way through the second season. Some of the season 1 episodes are painful to watch (I'm primarily thinking of TKO here), but by the time you get to season 3 that just doesn't happen any more.
I've never enjoyed Trek so much as I did in Season 4 of Enterprise, so I may be the wrong guy to talk to...
No, no, no. Season 4 of Enterprise was good. Don't worry about that. It's just that as the two of us are the only ones who actually bothered to watch it, nobody realised this. Everyone else gave up when they saw the alien nazis.
How so? I mean, really, I've watched both series in their entirety more than once. How did Firefly have anything in common with Buffy?
Amazingly compelling characters who have a distinctive use of language that you pick up and start using yourself. Funny one-liners. Individual episode stories combined with over-arching season plot.
So basically: they were both well written. Big surprise there.
Only about half of Voyager was like that. And looking at the listing on Memory Alpha, it was a half that this guy wasn't involved with. The episodes he wrote are the ones that took the characters and had them interacting in new and different ways (e.g., my favourite Voyager episode, "Workforce", where most of the characters have had their memory of being on Voyager erased and aren't aware they know each other).
I still don't get this. Why is knife crime suddenly such a big deal here in the UK? It seems like every other day some newspaper or TV news or something is talking about it. You hear phrases like "knife-crime epidemic" bandied about.
Anybody here in the UK who isn't reading Mark Easton's blog needs to add it to their RSS client now. I mean, before you even consider reading the next comment.
Unfortunately, most idiots who spout drivel like this don't even have a strong correlation in the first place. Sales of violent video games may be up, and knife crimes might be up, but is it even the kids playing the games committing the crimes?
This hysteria and panic is caused by, well, nothing. Except the fact that for some unknown reason over the last 5 years the media has become much more likely to report each and every incident of violence with a knife that they get to hear about.
So, if there is any correlation, it's a negative one: more video games, less knife crime.
Companies are considered non-natural person's for the purposes of taxations. You tax a company like a person is taxed (thought at a different rate). Companies have their own SS# (known as an EIN number ##-#######). After that the similarities between a company and a person stop.
Companies are also considered people in most other areas of the law, for instance you can commit theft from a company, despite most definitions of theft being described in terms of people, e.g. "[a] person is guilty of theft, if he dishonestly appropriates property belonging to another with the intention of permanently depriving the other of it" (UK legal definition). If a company wasn't a person, this definition means I could "dishonestly appropriate" its property without it being theft.
AIUI, the only areas where companies are _not_ considered people are laws that are specifically as only applying to a "natural person". Typically, this is only in legislation concerning voting rights and some human rights legislation.
Then, his argument breaks down because he'd be implying that individuals are analogous to machines...
People are machines, and the only reason they can't be patented at the moment is because there's too much prior art.
Wait until genetic modifications to people become a reality, then you'll start seeing patented people (tm). The human rights consequences are... interesting.
Unless you're masochistic enough to actually want to waste time in court, in the real world you treat contractual obligations as absolutes.
Yes, but consider this situation: you've already downloaded the data, and are being prosecuted. You have a choice; you can say that you authorized the download, thus violating your contract, or you can accept that the download was unauthorized. The court will assess the former on the same terms they assess any breach of contract; i.e. you will be required to pay actual damages + the opposition's legal costs; they will assess the latter on the same terms as any other copyright violation; i.e. you will be required to pay incredibly over-inflated statutary damages + the opposition's legal costs. Sounds like in the real world in this situation saying you breached the contract is better than saying you violated the copyright.
BTW, that's one reason I have a perfect record: I know how to draw people into disproving their own arguments
Yeah, but you only won because I phrased the original post sloppily.:)
Hold on...how can someone authorise themselves to "legally violate" a contract? A contract is an agreement between two parties, and any action precluded by the contract isn't legal unless both parties amend the agreement; one signatory can't arbitrarily decide that they aren't going to abide by the terms.
Yes they can. A contract is not absolute; anybody is free to violate the terms of a contract at any time they wish. Doing so opens them up to being sued for breach of the contract, and could be pretty expensive, but there is nothing that legally prevents them doing so (unless a court has previously ordered them not to).
For anyone who isn't aware, porn (particularly of the "gonzo", "reality" style) totally gives the finger to notions of plot and character development.
Depends very much on your porn. I'm quite nostalgically fond of 70s and 80s porn, which tends to have much more in the way of plot than modern stuff does.:)
a person who steals another person's property, esp. by stealth and without using force or violence.
"Steal v. i. 1. To practice, or be guilty of, theft; to commit larceny or theft. "(Webster's unabridged)
See definition of theft above, which downloading via p2p does not meet the definition of. See also definition of larceny:
"Larceny n. The unlawful taking and carrying away of things personal with intent to deprive the right owner of the same; theft." (Webster's unabridged)
Downloading via p2p also doesn't meet this definition. So somebody who does this does not steal, therefore is not a thief.
...Every time AI researches find a working algorithm for something that the human mind does, the ability coded on that algorithm stops being thought of as "Intelligence" and becomes "just a calculation that any computer can do".
So I guess pattern recognition in images is not AI anymore, right?
Doing it in a way that works is. Current facial recognition software has about a false positive rate of about 0.1% and a false negative rate of 1%, when comparing randomly selected people against a single sample. Obviously the false positive rate climbs as the number of sample images to match against climbs. In a school environment, I doubt they're getting better than about 3-5% false positives. This is clearly much worse than a person would achieve: looking at a person's face for 1.5 seconds while looking them up in a book of pictures of students should be a relatively easy task, and I'd expect substantially less than 1% false positives with no false negatives from a person doing this job.
Of course, these statistics (based on random face matches) are meaningless if somebody's actively trying to deceive the system, and I'd expect a person to better still in such a situation.
Slashdot Mirror
You throw transistors at your developers? ;-)
Yeah. I started with BC10s and 2N7000s, but it wasn't helping much. I found to get results, it has to be one of these.
I'm not holding anything against it that regard. The simple fact is that I've had two fairly low traffic MySQL databases become corrupted beyond the point of being usable within the last 3 years. The hardware wasn't at fault here (nor was it old or outdated).
I'm not sure what you're doing wrong here, but I think many of us have been running a lot more MySQL databases than that and never experienced corruption. Myself, I have been maintaining on average about 20 MySQL instances spread across 3 different servers for the last 10 years, and have never[1] experienced corruption that wasn't tied to a hardware failure. Sometimes after a system crash I need to do a myisamchk --recover, but even that's rare. I've never needed to do myisamchk --sort-recover, although I understand there are circumstances where this is necessary.
So my experience has been that mysql is, essentially, rock solid. It could be that we're using it for different applications and it works for mine but not yours, but I've also never seen any reports of problems with mysql corruption that cannot be fixed automatically by the tools mysql provides.
[1]: OK, this isn't quite true. There was one instance of corruption that was tied to a linux kernel version that had a bug where dirty buffers were sometimes corrupted prior to being flushed to disk. Hard to pin that one on MySQL either, though.
1) MySQL transactions are built into the table engines, and by default (last I checked, and meaning you don't install innodb, etc), the tables will not be transactional. This means that if you are building an inhouse app, you can trust it more than you can if you are distributing your software. In short, if you are distributing software you can't guarantee that it is running on a system with transactions without a great deal of headache........ The same goes for referential integrity enforcement.
It's easy enough to set up a database creation script that ensures the tables have the necessary support. Sure, if you're working on tables that somebody else has created you can have issues, but in your experience how often does that happen? And if it does happen, how often do you have a choice as to what dbms is in use? In my 10+ years as a consultant I've only ever had to do that once, and never with mysql (because the client's existing system I had to integrate with used informix).
2) Strict mode can be turned off by any application. This means that the more recent data integrity checks cannot be relied upon. This is an issue on both inhouse and distributed software because it adds quite a bit of overhead to the QA process internally, and can add support headaches in software for distribution.
On the whole, this is probably a good thing. If the application is under your control, you can use whichever mode you want. If you're relying on somebody else's application, forcing it to use strict mode when it wasn't written for this environment could introduce subtle bugs. Now, if you were to argue that the _existence_ of these different modes of operation was an issue, then I'd probably agree. But given the existence of the modes (and that's unfortunately a necessity for backwards compatibility reasons) the ability of the application to change the mode to the one it expects is crucial.
MySQL is a good db for single-app databases, where data integrity is not a tremendous issue or where you are deploying a separate MySQL instance on a different port.
Of course, you are describing something like 90%+ of all database applications there. Yes, this book should cover MySQL: it's a very popular database that is perfectly adequate for most uses. Sure, there are applications where it shouldn't be used, but that hasn't stopped it from becoming extremely widely deployed and being used as the database of choice by web developers everywhere (which is probably the target market for this book).
Anyway, a more important number is the capacity, how many Wh (or mAh) do these things hold? Is it 25kW for 10s? or for 1 hour? or what?
I assume, since they didn't mention a capacity, that it is unchanged from other batteries of similar chemistry, i.e. approx 120Wh/kg.
He's overstating the point. The best Li-ion batteries get about 150Wh/kg; the best LiFePO4 batteries get about 120Wh/kg. But, there's a crucial difference between the two that he doesn't mention: LiFePO4 batteries typically survive between 3 and 5 thousand cycles before they noticeably lose capacity; Li-ion, only slightly over 1 thousand. If you're building an EV you get a choice: spend $10,000 on Li-ion batteries and have a slightly larger range, but you'll have to replace them after 4 years, or spend $12,000 on LiFePO4 batteries, get slightly lower range, but have the batteries last about as long as the car does. Frankly, it's a no-brainer.
The claim sounds like hype, since it's energy density, not discharge rate that everyone is trying to ramp up lately.
No, energy density and discharge rate are both problems. Lithium batteries are clear leaders in energy density (at least some of the more unusual varieties like lithium-iron phosphate), but they have very slow discharge. This means that to get the 500A you can easily get out of a single lead acid battery, you need something like 20 batteries working in parallel. And making more small cells is more expensive than fewer large ones. A typical EV power system based on LiFePO4 will want something to be able to provide something like 100v x 200A peak, which means 8 parallel banks each containing about 50 cells. With this new design, that can be cut to just 50 big cells, which will probably halve the cost of production.
Wow, that's a lot of power. Here in America our outlets are rated at about 20 Amps. That translates to roughly 2kW. what a coincidence!
The problem is, though, that drawing 20A/110V will heat up your wiring twice as much as drawing 10A/220V (assuming the same resistance), thus wasting twice as much to transmission losses. Also, the voltage fluctuations you'll see due to changing power demands are twice as large. So higher voltage == better.
Besides, here in the UK we get 240v x 13A = 3.1kW, so we beat you both. :)
Example - dysons when I was selling them had lower engine power than, say, hoover(1200 vs 2000 watts iirc),
Having coincidentally tested my Dyson DC10 with a power meter just yesterday, I can tell you that the draw from mains is 1kw. Not sure what the official rating is, though.
How many years ago was it that we had credible airlines considering banning in-flight battery use?
Actually, merely carrying more than a certain (rather small) capacity lithium battery on a flight is banned by international convention. Most laptops squeak under, but officially you're probably not allowed to carry a replacement battery on as well...
When you hear about knife crime being down, bear in mind that the government have been caught fiddling the numbers to make themselves look better.
Yes. They claimed a 16% (IIRC) drop in hospital admissions with knife injuries, but actually the drop was only 8%. Point still stands, though: there has been a serious reduction lately.
Babylon 5 we appreciated the serial nature but couldn't get past how bad each individual episode was
How far in did you watch it? It gets seriously better about half way through the second season. Some of the season 1 episodes are painful to watch (I'm primarily thinking of TKO here), but by the time you get to season 3 that just doesn't happen any more.
I've never enjoyed Trek so much as I did in Season 4 of Enterprise, so I may be the wrong guy to talk to...
No, no, no. Season 4 of Enterprise was good. Don't worry about that. It's just that as the two of us are the only ones who actually bothered to watch it, nobody realised this. Everyone else gave up when they saw the alien nazis.
How so? I mean, really, I've watched both series in their entirety more than once. How did Firefly have anything in common with Buffy?
Amazingly compelling characters who have a distinctive use of language that you pick up and start using yourself. Funny one-liners. Individual episode stories combined with over-arching season plot.
So basically: they were both well written. Big surprise there.
Only about half of Voyager was like that. And looking at the listing on Memory Alpha, it was a half that this guy wasn't involved with. The episodes he wrote are the ones that took the characters and had them interacting in new and different ways (e.g., my favourite Voyager episode, "Workforce", where most of the characters have had their memory of being on Voyager erased and aren't aware they know each other).
I still don't get this. Why is knife crime suddenly such a big deal here in the UK? It seems like every other day some newspaper or TV news or something is talking about it. You hear phrases like "knife-crime epidemic" bandied about.
See 2008 crime figures:
Nothing to see here. Move on. Stop whining, and yes, Daily Mail editors, I mean you.
The British gov't has systematically distorted statistics and selectively presented data in order to advance its own agenda.
Indeed. And so, too, has the media. As a consequence, while violent crime has dropped 8% over the last year, 65% of people polled thought it had increased.
Anybody here in the UK who isn't reading Mark Easton's blog needs to add it to their RSS client now. I mean, before you even consider reading the next comment.
It seems the mods have never played Hello Kitty Island Adventure, either. Interesting?!
Unfortunately, most idiots who spout drivel like this don't even have a strong correlation in the first place. Sales of violent video games may be up, and knife crimes might be up, but is it even the kids playing the games committing the crimes?
It's worse than that. Knife crime is down. The number of people injured by knives and other sharp instruments is down (although not by as much as was previously reported). Incidence of violent crime in general is down.
This hysteria and panic is caused by, well, nothing. Except the fact that for some unknown reason over the last 5 years the media has become much more likely to report each and every incident of violence with a knife that they get to hear about.
So, if there is any correlation, it's a negative one: more video games, less knife crime.
Companies are considered non-natural person's for the purposes of taxations. You tax a company like a person is taxed (thought at a different rate). Companies have their own SS# (known as an EIN number ##-#######). After that the similarities between a company and a person stop.
Companies are also considered people in most other areas of the law, for instance you can commit theft from a company, despite most definitions of theft being described in terms of people, e.g. "[a] person is guilty of theft, if he dishonestly appropriates property belonging to another with the intention of permanently depriving the other of it" (UK legal definition). If a company wasn't a person, this definition means I could "dishonestly appropriate" its property without it being theft.
AIUI, the only areas where companies are _not_ considered people are laws that are specifically as only applying to a "natural person". Typically, this is only in legislation concerning voting rights and some human rights legislation.
Then, his argument breaks down because he'd be implying that individuals are analogous to machines...
People are machines, and the only reason they can't be patented at the moment is because there's too much prior art.
Wait until genetic modifications to people become a reality, then you'll start seeing patented people (tm). The human rights consequences are... interesting.
Unless you're masochistic enough to actually want to waste time in court, in the real world you treat contractual obligations as absolutes.
Yes, but consider this situation: you've already downloaded the data, and are being prosecuted. You have a choice; you can say that you authorized the download, thus violating your contract, or you can accept that the download was unauthorized. The court will assess the former on the same terms they assess any breach of contract; i.e. you will be required to pay actual damages + the opposition's legal costs; they will assess the latter on the same terms as any other copyright violation; i.e. you will be required to pay incredibly over-inflated statutary damages + the opposition's legal costs. Sounds like in the real world in this situation saying you breached the contract is better than saying you violated the copyright.
BTW, that's one reason I have a perfect record: I know how to draw people into disproving their own arguments
Yeah, but you only won because I phrased the original post sloppily. :)
Hold on...how can someone authorise themselves to "legally violate" a contract? A contract is an agreement between two parties, and any action precluded by the contract isn't legal unless both parties amend the agreement; one signatory can't arbitrarily decide that they aren't going to abide by the terms.
Yes they can. A contract is not absolute; anybody is free to violate the terms of a contract at any time they wish. Doing so opens them up to being sued for breach of the contract, and could be pretty expensive, but there is nothing that legally prevents them doing so (unless a court has previously ordered them not to).
For anyone who isn't aware, porn (particularly of the "gonzo", "reality" style) totally gives the finger to notions of plot and character development.
Depends very much on your porn. I'm quite nostalgically fond of 70s and 80s porn, which tends to have much more in the way of plot than modern stuff does. :)
So I guess pattern recognition in images is not AI anymore, right?
Doing it in a way that works is. Current facial recognition software has about a false positive rate of about 0.1% and a false negative rate of 1%, when comparing randomly selected people against a single sample. Obviously the false positive rate climbs as the number of sample images to match against climbs. In a school environment, I doubt they're getting better than about 3-5% false positives. This is clearly much worse than a person would achieve: looking at a person's face for 1.5 seconds while looking them up in a book of pictures of students should be a relatively easy task, and I'd expect substantially less than 1% false positives with no false negatives from a person doing this job.
Of course, these statistics (based on random face matches) are meaningless if somebody's actively trying to deceive the system, and I'd expect a person to better still in such a situation.