I thought that not needing to worry about memory management was supposed to be one of the advantages of Java. If you need to take care about which structures you load into memory when, then you are better off writing the whole thing in C, where you have a more finely tuned way to control it.
And, where as a bonus, you will fuck it up over and over in all sorts of random places and you'll never be able to stamp it out. Don't optimize prematurely, bro.
To be fair, they issued a red notice, which is odd given that the crime he is charged with is actually not a crime in most other states.
He's been accused of rape. That's a crime everywhere. The whole idea that he's not been charged with rape, and instead for "consensual sex without a condom" is bullshit that his lawyers have made up and managed to get the press to repeat.
There is nothing in the NYTimes article that justifies this elaborate portrayal of the events. Is there another source that justifies this?
Oh, yeah, I forgot that one. It comes from The Daily Mail, which claims copies of the women's police statements as the source for most of its reporting.
Though not for this one tidbit, which was the source for my description: "One source close to the investigation said the woman had insisted he wear a condom, but the following morning he made love to her without one."
When exactly did either woman say "Stop"? Where did you get your information?
Sweden Issues Warrant for WikiLeaks Founder
By JOHN F. BURNS and RAVI SOMAIYA
Published: November 18, 2010
"According to accounts the women gave to the police and friends, they each had consensual sexual encounters with Mr. Assange that became nonconsensual. One woman said that Mr. Assange had ignored her appeals to stop after a condom broke. The other woman said that she and Mr. Assange had begun a sexual encounter using a condom, but that Mr. Assange did not comply with her appeals to stop when it was no longer in use."
I ask because your interpretation doesn't square with the factual record.
And the problem is that your "factual record" probably is sourced from Assange's lawyers, who are going around and bullshitting about this case.
For example, Assange wasn't wearing a condom when he fucked Jessica, so how could Jessica withdraw consent because of a broken condom?
That's not how the allegations go. It's more like this: (a) she sleeps with him first on the night, he uses a condom; (b) he is going to sleep with her again on the morning, doesn't put a condom on, she tells him not to do it without a condom, and he does it anyway.
Everyone is interpreting the claims of the prosecution that consent had been withdrawn to mean that the women actually said "No", "Stop", or "Don't". That is the interpretation the prosecution would like us to have. Indeed, that would be rape. But I've never seen the prosecution actually claimed the women ever said "No." The claims of the prosecution have been very vague, and its sounding more and more like BS.
Well, that's how prosecution claims tend to sound when your only source about them is what the defense says.
In any case it's clear that the women were initially pleased with Assange and only reported the events to the Police immediately after they discovered that Assange had been sleeping around with other women. That doesn't sound like rape, that sounds like promiscuity.
As I've said elsewhere, real-life rape victims often act very strangely, in a way that's superficially inconsistent with having been raped.
He did deal with the charges when he was still in Sweden. He offered on multiple occasions to come in for interrogation and asked if he was free to leave the country before he left.
And we know that's true because Assange's lawyers said it!
Consent in law is a very complicated subject in law. Depending on circumstances the withdrawal of consent for a activity that is in-progress may be void. For example, if you consent to be operated on but your anaesthetic wore off and woke up, you cannot suddenly withdraw consent.
It takes many daring leaps of logic, however, to get from there to "'take your dick out of my pussy' means I can continue."
The story certainly is more complicated than that. Especially with the women continuing a positive relation with him for several days, and the alleged tweets that happened after the fact.
You know, people are strange. That they minimize rape is one of the less strange facts, but the fact that many minimize their own rape is much stranger, but it does happen.
However, instead of rape, let's use a different example: domestic violence. It is very, very well known that chronically battered wives tend to blame themselves for the violence, not their husband. They will routinely cover up for their husbands, oppose charges against them, or defend him when charges are brought up. Yeah, it's very messed up, but that's pretty standard battered wife behavior.
Well, guess what, rape victim behavior can be very similar. Even in cases where the man clearly used force to obtain compliance from a woman who didn't want it or enjoy it, you'll find that many women will think and insist that it was their own fault.
So, whether the alleged victime "continued a positive relation" with the guy after the alleged rape is no more exculpatory evidence of rape than it would be of domestic abuse. By that standard, lots of clear cases of either would fail.
According to accounts the women gave to the police and friends, they each had consensual sexual encounters with Mr. Assange that became nonconsensual. One woman said that Mr. Assange had ignored her appeals to stop after a condom broke. The other woman said that she and Mr. Assange had begun a sexual encounter using a condom, but that Mr. Assange did not comply with her appeals to stop when it was no longer in use.
A better example would be a case where a guy committed purgery when, on the stand, when he testified that he was with his wife on the night that his wife's sister was accused of drunk driving. Having told his wife that he was at work, he pisses off his wife so much that she pulls out a gun and shoots the Judge, all members of the Jury, and the husband's lawyer. Obviously, he should be charged with murder.
Um, criminal codes typically specify that the felony murder rule applies to specific felonies. A typical list is: robbery, rape or forcible deviant sexual intercourse, arson, burglary, felonious escape, terrorism, kidnapping and carjacking. Perjury ain't on any of these lists that I'm aware of.
There is also a body of legal precedent that lays out guidelines for when the rule applies.
Eh, I don't see any evidence in anything I've read that the Armed Forces are considering to incorporate a gas piston system into the M16 or M4. Rather, all I've read indicates that they've repeatedly shot down the idea of modifications to the rifles just to enhance reliability. They'd rather switch to a new weapon that provides superior capabilities—hence the XM25 system being discussed in this story.
Can't say I don't see their point. The present rifles have less reliability that other designs, sure, but rifles play a relatively minor role in warfare anyway, so better save money and headaches unless something bigger comes along with the new rifle.
In Vietnam, American troops were armed with the recently-developed M-16, early versions of which frequently jammed. They jammed because the rifle was prototyped using ammunition packed with pellet-shaped nitrocellulose gunpowder (which worked fine in bad conditions), but mass-produced using stick-type nitrocellulose/nitrogylcerin gunpowder (which fouled the barrel if the weapon was not cleaned regularly). The lack of cleaning supplies and instructions for troops didn't help matters either.
Once this design flaw was identified, the powder was changed, the barrel was lined with chrome, and troops were given instructions and tools to clean the weapons. Afterward, they became much more reliable in jungle conditions.
This apology for the M-16 just misses the forest for the trees. The reason the M-16 is so sensitive to the type of gunpowder used is because it uses direct impingement gas operation. Note that most other common military rifle families don't use this design. Why don't they? Because it's less reliable!
This is actually one of the reasons those superzoom cameras work as well as they do.
Interesting. I was actually speaking of superzoom lenses for SLRs, not cameras with integral superzooms. In the latter case I can see that they could just correct it in software. In the former case, the camera you put the lens on may well have been made before the lens was designed, so they can't rely on software.
The gating factors here are: (a) if you have an optical viewfinder, you can't perform software corrections that would change the framing; (b) if the lens is supposed to also work on film cameras, you have to deliver a decent quality image to the film. So what you see is that, e.g., Nikon DSLRs do automatic correction of chromatic aberration but not distortion, while Micro Four Thirds cameras can do both since they have a 100% electronic viewing system, can do both.
That, plus the smaller sensor (2x crop) and the shorter mount-to-sensor distance in the mirrorless camera, is why Panasonic's ultra wide angle 7-14mm f/4 lens is smaller than Nikon's basic 18-55mm kit lenses, and gives you photos with very little distortion.
But the quality of the optics is no better than it used to be.
Actually, that's not true. The field of optical engineering is seeing very rapid advancement, in part due to new lens coatings and in part due to the ability to model complex series of lenses in computers, rather than having to build and test them experimentally.
Strictly speaking, that's been the case since the 80s.
The result is that today's top-quality lenses are sharper and have less distortion (chromatic abberation, barrel and pincushion distortion, etc.) than older lenses. And the new technology is making some things possible that simply weren't before, like the new crop of superzooms that actually have reasonable performance across very wide zoom ranges.
And actually, one of the newest techs is the use of software instead of optics to correct chromatic aberration and curvilinear distortion. I.e., if your camera's viewing and capture systems are all-electronic, your lens design can be relaxed when it comes to those, as long as you can profile it accurately and correct it in software with a reasonable loss in quality.
This is actually one of the reasons those superzoom cameras work as well as they do.
The other thing I'll say is that my impression is that the high end of at least the T2i is probably more noisy "natively" than it was on the XT (which maxed out at 1600), so effectively I'd say that the T2i is between one and two stops better than the XT on that count. However, it's a little hard to say; they've put more effort into noise reduction too, so I don't think it's quite a fair comparison.
That said, for amateur shots especially, many shots taken with ISO 6400 and with Lightroom 3's noise reduction (which pretty much rocks) still come out quite well. I'd say you can't quite rely on it, and you do lose some detail, but you could almost certainly shoot at 1600 with the T2i and have consistently good photos.
(I did see an article talking about camera manufacturers turning ISO into "the new megapixels" and cranking it up just because it sounds good, but they are slowly getting better. Personally, I'd love to see the T3i or whatever shed some megapixels and decrease noise.)
The relationship between pixel count and noise is something that the amateur common wisdom has managed to completely misconceive over the past few years. This "fewer megapixels" comment of yours is, frankly, one example.
The misconception is based around one thing that is a fact: smaller pixels are more prone to noise than larger pixels. However, it ignores the fact that, at a given sensor size, smaller pixels means more pixels, which compensates for the increased noise of the individual pixels. Since noise is random, downscaling an image reduces per-pixel noise.
As a general rule, more, smaller pixels are better as long as the individual pixels are not disproportionately worse than fewer, larger ones. Rule of thumb: if you have twice as many pixels, and those pixels capture half as much information as larger ones would, you've broken even on the noise angle, and you've also increased spatial resolution, so you're ahead.
This means that if you take two cameras with the same sensor size, shoot the same photo with each, and display them at the same physical size (which is critical), one of the following will hold (assuming other factors don't mess it up):
At smaller display sizes, where the image with more pixels needs to be downscaled, neither image will be noisier than the other. The image from the higher-pixels camera might have more fine detail, especially in strongly colored areas.
At larger display sizes, where the image with fewer pixels needs to be upscaled, the image from the higher-pixels camera will be noisier, but it will have more detail. Unless you apply some form of noise reduction (e.g. by downscaling and then upscaling), in which case you can get the same image from both.
The reason people keep convincing themselves that cameras with more pixels are "worse" is because they view the images from both cameras at 1:1 pixel ratio, which means that the image with more pixels is enlarged more, and will therefore look noisier.
* For anyone who doesn't know, the "boost" means that the work is being done in software. My impression of "boost" ISO modes is that they are roughly equivalent to underexposing the image a corresponding amount then correcting that in postprocessing, except done in-camera. So on a T2i, setting the ISO to 12800 will do the same about thing as setting it to 6400, setting the exposure compensation down one stop, opening the photo in Photoshop/Lightroom/Aperture/whatever, and setting the exposure correction to +1. This could be totally off though.
It's been shown that many cameras implement various non-"boost" ISO settings in software. I.e., the actual exposure level used to take the shot is a function of both sensor gain and output processing curve, so ISO settings are implemented by some combination of these two controls.
I really suspect that these separate "boost" modes are just arbitrarily segregated by the UI from the regular ones, because the camera maker wants to lower people's expectations of their image quality.
The other downside is that DSL does not imply interchangeable lenses and would technically be an accurate name for some smaller form cameras.
"SLR" doesn't imply that either. No difference there. In fact, the traditional camera design names don't say whether the camera has interchangeable names. View camera, rangefinder, TLR (Twin Lens Reflex), SLR (single lens reflex): all of those have existed in fixed-lens versions.
Someone should build a P&S that takes interchangeable lenses and then offer a kit with a 500mm mirror lens and a few 2X teleconverters, just to piss off the Kuwaiti heads of Ministries of this and that and the other thing.
It's much easier to focus on exactly what you want with an SLR, even a digital one.
Hell no, not in general. My all-electronic Panasonic G1 is far easier to focus than my old Nikon D70. The Panasonic's contrast-based autofocus is more accurate than the Nikon's phase-detection system. If I want to focus manually, the G1 can magnify the live view to pixel level. The electronic viewfinder on the G1 is larger and brighter than the D70's. Also, the EVF has a shimmering color motion artifact (moiré?) that often shows up on subject detail that's in sharp focus.
Or what about interchangeable lens cameras with an LED-based "viewfinder" that do not actually use a reflex mirror? I think they're called by some "bridge cameras."
A bridge camera is something different: a non-interchangeable small-sensor camera with a larger body style and a very large zoom range, often with RAW output.
What you are talking about are the newer crop of all-digital large-sensor interchangeable lenses cameras by Panasonic/Olympus (Micro Four Thirds), Samsung (NX system) and Sony (NEX system). There still isn't a consensus name for this type of camera, but here's the common names you hear:
EVIL: Electronic Viefinder Interchangeable Lens
Mirrorless
MILC: Mirrorless Interchangeable Lens Camera
DSL: Digital Single Lens (term really only used by Panasonic, but it's my fave)
If you are in the really cheap seats, you should probably spend whatever audio money you have on speakers or headphones that don't utterly suck.
No, I have to disagree on this excessively general prescription (even though you did stick the "probably" in there). I've heard on-board audio that was so bad that clearly it was more important to replace the audio output than to upgrade the headphones or speakers. I had a ThinkPad once whose output was very hissy and had interference from other components in the system, and it was all clearly audible with crappy headphones.
Basically, the most general prescription that can be made is to identify the weak link and upgrade that first. If you hear hiss or interference, get a reasonable sound card. It doesn't have to be fancy; it just has to be competent.
Actually, it is not necessary to use an additional "s" to form a possesive with words that end with an "s" sound. Several sources say that it is prefered, however I find that it is more readable without the additional "s".
For the non-grammar-Nazis out there, I think it's worth tossing into this discussion a brief descriptive account of how possessive marking works in English (i.e., in real, spoken English, not in the crappy conventional orthographic renderings thereof).
In English, a noun consists of an obligatory stem and an optional inflectional ending. So, for example, in the word dogs, dog- is the stem, and -s is the inflectional ending—which in this case serves as a plural marker. Other examples:
In the word dog, we have just the stem and no ending.
In the words man and men, we also have just a stem and no ending. Man is an irregular noun that forms its plural by using a different stem, not by adding an ending.
In the words lease, wuss and Asus, we also have no ending, but with one interesting fact that'll become relevant in a bit: these words have stems that end on an s or z sound.
So, some rules now:
Nouns with irregular plurals have their own rules for plural formation, which are not listed here.
You form the (regular) plural or possessive of a noun with a stem that ends in a s or z sound by adding an ending that sounds like -uz: leases, the lease's end, wusses, the wuss's friend, the analyst predicted that Acme and Beta will be the next Asuses, Asus's advantages over its competitors. (Don't put too much stock on the orthography; actually pronounce these and listen to how the words are pronounced.)
For words that end on a voiceless consonant, you add an -s sound at the end: cats, the cat's whiskers.
For other words, you add a -z ending at the end (written as s): dogs, Joe's.
You can only have one inflectional ending marker on a noun. A regular nouns that's simultaneously marked for plural and possessive gets only one marker: the wusses' friends.
Irregular nouns that don't use the -s ending for the plural can still use it for possessive: all of those men's possessions.
So, the orthographic rule that renders as wuss's the possessive-marked singular noun is the one that most closely accords with the rules of English grammar.
I worked on some code done by someone else, where on massive records, they were always selecting "*" instead of the needed or anticipated values. Big waste when one needs (by ID#) last and first name and selects a whole row instead - then wonders why it's not scaling upwards.
Eh, I wonder if you're overstating the performance implications of that. Those are all row-oriented databases. Unless all of the columns your query needs are found in an index, it's going to have to read the whole row from disk anyway; the extra costs from the * then become (a) memory and CPU usage and (b) network bandwidth. In my experience, network bandwidth is usually not a big problem; memory and CPU usage can be an issue, but the big performance killers tend to be inefficient joins (because they don't scale linearly), while scalar stuff (the which the * would fall into) are usually cheap.
Nope, not overstating anything.
Let's say it's an ambulance company database that's used to calculate their LOSAP points for the year... that requires calculations from EVERY data parameter input, since everything a member does goes towards their LOSAP points. In that data are things like their PCRs (Patient Care Reports). Each PCR may have.1MB (not counting scans) of data associated with it. Let's say there are 200 members and 24,000 PCRs. Now... let's assume the server has 2GB of RAM. That's 2.4GB of data to read just from the PCR tables alone if one loads all columns. Or a twenty minute report. Even stepping through record by record (due to the overhead of 24,000 individual read requests JUST for the PCR data - even when using the same DB connection).
Your explanation doesn't narrow down at all the cause of the speedup that you're experiencing. If that 2.4 GB case refers to the data in one table, the database, unless the query and schema fit some narrow conditions (e.g., all of the 8 columns your query wants are stored in the same index), is still reading the 2.4 gig of data. This is because all of the data for each row is stored together in disk; you can't (normally) just read the 8 columns you want.
There is at least one more factor missing from your explanation (which, to be frank, I find hopelessly vague) that's just as essential to explain the speedup you're seeing. One (wild) guess: you're doing some large joins, the database needs to materialize intermediate join results, and losing the stars means those intermediate result sets become at lot smaller. Another (also wild) guess: your application is using the database as a dumb data store, pulling rows one by one from the DB, processing them individually in the application server. That's inevitably going to be very slow.
If the schools realized that it's 2010, not 1810, and if teachers actually were a bit more passionated about learning than a corpse i'm certain cheating would drop a fair bit.
I don't normally criticise people for language and grammar, since it is beside the point, but I think since you are criticising university teaching quality and seem to imply that you are a student on one, it is fair in this case. So, don't you mean to say something like "If teachers were a bit more passionate (note the form of the word) about teaching (teacher may learn, but they are supposed to teach)"?
Um, notice the placement of the "e" and the "d" keys in the keyboard. And, have you considered the possibility that GP's problem really is teachers who aren't passionate about learning? Why do you assume that GP must have meant "teaching"?
It would lend more credibility to your arguments if you didn't commit such sloppy errors.
Why? How does it lend credibility to an argument to write it in a way that pleases somebody who's looking for excuses not to evaluate the freaking argument anyway?
Slashdot Mirror
I thought that not needing to worry about memory management was supposed to be one of the advantages of Java. If you need to take care about which structures you load into memory when, then you are better off writing the whole thing in C, where you have a more finely tuned way to control it.
And, where as a bonus, you will fuck it up over and over in all sorts of random places and you'll never be able to stamp it out. Don't optimize prematurely, bro.
To be fair, they issued a red notice, which is odd given that the crime he is charged with is actually not a crime in most other states.
He's been accused of rape. That's a crime everywhere. The whole idea that he's not been charged with rape, and instead for "consensual sex without a condom" is bullshit that his lawyers have made up and managed to get the press to repeat.
And we know that's true because Assange's lawyer told us!
There is nothing in the NYTimes article that justifies this elaborate portrayal of the events. Is there another source that justifies this?
Oh, yeah, I forgot that one. It comes from The Daily Mail, which claims copies of the women's police statements as the source for most of its reporting.
Though not for this one tidbit, which was the source for my description: "One source close to the investigation said the woman had insisted he wear a condom, but the following morning he made love to her without one."
When exactly did either woman say "Stop"? Where did you get your information?
Sweden Issues Warrant for WikiLeaks Founder By JOHN F. BURNS and RAVI SOMAIYA Published: November 18, 2010 "According to accounts the women gave to the police and friends, they each had consensual sexual encounters with Mr. Assange that became nonconsensual. One woman said that Mr. Assange had ignored her appeals to stop after a condom broke. The other woman said that she and Mr. Assange had begun a sexual encounter using a condom, but that Mr. Assange did not comply with her appeals to stop when it was no longer in use."
I ask because your interpretation doesn't square with the factual record.
And the problem is that your "factual record" probably is sourced from Assange's lawyers, who are going around and bullshitting about this case.
For example, Assange wasn't wearing a condom when he fucked Jessica, so how could Jessica withdraw consent because of a broken condom?
That's not how the allegations go. It's more like this: (a) she sleeps with him first on the night, he uses a condom; (b) he is going to sleep with her again on the morning, doesn't put a condom on, she tells him not to do it without a condom, and he does it anyway.
Everyone is interpreting the claims of the prosecution that consent had been withdrawn to mean that the women actually said "No", "Stop", or "Don't". That is the interpretation the prosecution would like us to have. Indeed, that would be rape. But I've never seen the prosecution actually claimed the women ever said "No." The claims of the prosecution have been very vague, and its sounding more and more like BS.
Well, that's how prosecution claims tend to sound when your only source about them is what the defense says.
In any case it's clear that the women were initially pleased with Assange and only reported the events to the Police immediately after they discovered that Assange had been sleeping around with other women. That doesn't sound like rape, that sounds like promiscuity.
As I've said elsewhere, real-life rape victims often act very strangely, in a way that's superficially inconsistent with having been raped.
He did deal with the charges when he was still in Sweden. He offered on multiple occasions to come in for interrogation and asked if he was free to leave the country before he left.
And we know that's true because Assange's lawyers said it!
It takes many daring leaps of logic, however, to get from there to "'take your dick out of my pussy' means I can continue."
The story certainly is more complicated than that. Especially with the women continuing a positive relation with him for several days, and the alleged tweets that happened after the fact.
You know, people are strange. That they minimize rape is one of the less strange facts, but the fact that many minimize their own rape is much stranger, but it does happen.
However, instead of rape, let's use a different example: domestic violence. It is very, very well known that chronically battered wives tend to blame themselves for the violence, not their husband. They will routinely cover up for their husbands, oppose charges against them, or defend him when charges are brought up. Yeah, it's very messed up, but that's pretty standard battered wife behavior.
Well, guess what, rape victim behavior can be very similar. Even in cases where the man clearly used force to obtain compliance from a woman who didn't want it or enjoy it, you'll find that many women will think and insist that it was their own fault.
So, whether the alleged victime "continued a positive relation" with the guy after the alleged rape is no more exculpatory evidence of rape than it would be of domestic abuse. By that standard, lots of clear cases of either would fail.
That certainly isn't in the referenced article - where do you find that she appealed to him to stop and he did not?
Sweden Issues Warrant for WikiLeaks Founder
By JOHN F. BURNS and RAVI SOMAIYA
Published: November 18, 2010
The money quote:
According to accounts the women gave to the police and friends, they each had consensual sexual encounters with Mr. Assange that became nonconsensual. One woman said that Mr. Assange had ignored her appeals to stop after a condom broke. The other woman said that she and Mr. Assange had begun a sexual encounter using a condom, but that Mr. Assange did not comply with her appeals to stop when it was no longer in use.
The big problem that I see is that there's some media right now whose "reporting" is basically repeating Assange's lawyers' statements at length.
A better example would be a case where a guy committed purgery when, on the stand, when he testified that he was with his wife on the night that his wife's sister was accused of drunk driving. Having told his wife that he was at work, he pisses off his wife so much that she pulls out a gun and shoots the Judge, all members of the Jury, and the husband's lawyer. Obviously, he should be charged with murder.
Um, criminal codes typically specify that the felony murder rule applies to specific felonies. A typical list is: robbery, rape or forcible deviant sexual intercourse, arson, burglary, felonious escape, terrorism, kidnapping and carjacking. Perjury ain't on any of these lists that I'm aware of.
There is also a body of legal precedent that lays out guidelines for when the rule applies.
I misspoke a bit. No, the XM25 isn't a replacement for the M16, but it's the product of the OICW program that's considering a replacement for the M16.
Eh, I don't see any evidence in anything I've read that the Armed Forces are considering to incorporate a gas piston system into the M16 or M4. Rather, all I've read indicates that they've repeatedly shot down the idea of modifications to the rifles just to enhance reliability. They'd rather switch to a new weapon that provides superior capabilities—hence the XM25 system being discussed in this story.
Can't say I don't see their point. The present rifles have less reliability that other designs, sure, but rifles play a relatively minor role in warfare anyway, so better save money and headaches unless something bigger comes along with the new rifle.
In Vietnam, American troops were armed with the recently-developed M-16, early versions of which frequently jammed. They jammed because the rifle was prototyped using ammunition packed with pellet-shaped nitrocellulose gunpowder (which worked fine in bad conditions), but mass-produced using stick-type nitrocellulose/nitrogylcerin gunpowder (which fouled the barrel if the weapon was not cleaned regularly). The lack of cleaning supplies and instructions for troops didn't help matters either.
Once this design flaw was identified, the powder was changed, the barrel was lined with chrome, and troops were given instructions and tools to clean the weapons. Afterward, they became much more reliable in jungle conditions.
This apology for the M-16 just misses the forest for the trees. The reason the M-16 is so sensitive to the type of gunpowder used is because it uses direct impingement gas operation. Note that most other common military rifle families don't use this design. Why don't they? Because it's less reliable!
This is actually one of the reasons those superzoom cameras work as well as they do.
Interesting. I was actually speaking of superzoom lenses for SLRs, not cameras with integral superzooms. In the latter case I can see that they could just correct it in software. In the former case, the camera you put the lens on may well have been made before the lens was designed, so they can't rely on software.
The gating factors here are: (a) if you have an optical viewfinder, you can't perform software corrections that would change the framing; (b) if the lens is supposed to also work on film cameras, you have to deliver a decent quality image to the film. So what you see is that, e.g., Nikon DSLRs do automatic correction of chromatic aberration but not distortion, while Micro Four Thirds cameras can do both since they have a 100% electronic viewing system, can do both.
That, plus the smaller sensor (2x crop) and the shorter mount-to-sensor distance in the mirrorless camera, is why Panasonic's ultra wide angle 7-14mm f/4 lens is smaller than Nikon's basic 18-55mm kit lenses, and gives you photos with very little distortion.
But the quality of the optics is no better than it used to be.
Actually, that's not true. The field of optical engineering is seeing very rapid advancement, in part due to new lens coatings and in part due to the ability to model complex series of lenses in computers, rather than having to build and test them experimentally.
Strictly speaking, that's been the case since the 80s.
The result is that today's top-quality lenses are sharper and have less distortion (chromatic abberation, barrel and pincushion distortion, etc.) than older lenses. And the new technology is making some things possible that simply weren't before, like the new crop of superzooms that actually have reasonable performance across very wide zoom ranges.
And actually, one of the newest techs is the use of software instead of optics to correct chromatic aberration and curvilinear distortion. I.e., if your camera's viewing and capture systems are all-electronic, your lens design can be relaxed when it comes to those, as long as you can profile it accurately and correct it in software with a reasonable loss in quality.
This is actually one of the reasons those superzoom cameras work as well as they do.
The other thing I'll say is that my impression is that the high end of at least the T2i is probably more noisy "natively" than it was on the XT (which maxed out at 1600), so effectively I'd say that the T2i is between one and two stops better than the XT on that count. However, it's a little hard to say; they've put more effort into noise reduction too, so I don't think it's quite a fair comparison.
That said, for amateur shots especially, many shots taken with ISO 6400 and with Lightroom 3's noise reduction (which pretty much rocks) still come out quite well. I'd say you can't quite rely on it, and you do lose some detail, but you could almost certainly shoot at 1600 with the T2i and have consistently good photos.
(I did see an article talking about camera manufacturers turning ISO into "the new megapixels" and cranking it up just because it sounds good, but they are slowly getting better. Personally, I'd love to see the T3i or whatever shed some megapixels and decrease noise.)
The relationship between pixel count and noise is something that the amateur common wisdom has managed to completely misconceive over the past few years. This "fewer megapixels" comment of yours is, frankly, one example.
The misconception is based around one thing that is a fact: smaller pixels are more prone to noise than larger pixels. However, it ignores the fact that, at a given sensor size, smaller pixels means more pixels, which compensates for the increased noise of the individual pixels. Since noise is random, downscaling an image reduces per-pixel noise.
As a general rule, more, smaller pixels are better as long as the individual pixels are not disproportionately worse than fewer, larger ones. Rule of thumb: if you have twice as many pixels, and those pixels capture half as much information as larger ones would, you've broken even on the noise angle, and you've also increased spatial resolution, so you're ahead.
This means that if you take two cameras with the same sensor size, shoot the same photo with each, and display them at the same physical size (which is critical), one of the following will hold (assuming other factors don't mess it up):
The reason people keep convincing themselves that cameras with more pixels are "worse" is because they view the images from both cameras at 1:1 pixel ratio, which means that the image with more pixels is enlarged more, and will therefore look noisier.
* For anyone who doesn't know, the "boost" means that the work is being done in software. My impression of "boost" ISO modes is that they are roughly equivalent to underexposing the image a corresponding amount then correcting that in postprocessing, except done in-camera. So on a T2i, setting the ISO to 12800 will do the same about thing as setting it to 6400, setting the exposure compensation down one stop, opening the photo in Photoshop/Lightroom/Aperture/whatever, and setting the exposure correction to +1. This could be totally off though.
It's been shown that many cameras implement various non-"boost" ISO settings in software. I.e., the actual exposure level used to take the shot is a function of both sensor gain and output processing curve, so ISO settings are implemented by some combination of these two controls.
I really suspect that these separate "boost" modes are just arbitrarily segregated by the UI from the regular ones, because the camera maker wants to lower people's expectations of their image quality.
The other downside is that DSL does not imply interchangeable lenses and would technically be an accurate name for some smaller form cameras.
"SLR" doesn't imply that either. No difference there. In fact, the traditional camera design names don't say whether the camera has interchangeable names. View camera, rangefinder, TLR (Twin Lens Reflex), SLR (single lens reflex): all of those have existed in fixed-lens versions.
Someone should build a P&S that takes interchangeable lenses and then offer a kit with a 500mm mirror lens and a few 2X teleconverters, just to piss off the Kuwaiti heads of Ministries of this and that and the other thing.
You mean something like a Panasonic G2 with a 100-300mm lens?
It's much easier to focus on exactly what you want with an SLR, even a digital one.
Hell no, not in general. My all-electronic Panasonic G1 is far easier to focus than my old Nikon D70. The Panasonic's contrast-based autofocus is more accurate than the Nikon's phase-detection system. If I want to focus manually, the G1 can magnify the live view to pixel level. The electronic viewfinder on the G1 is larger and brighter than the D70's. Also, the EVF has a shimmering color motion artifact (moiré?) that often shows up on subject detail that's in sharp focus.
A bridge camera is something different: a non-interchangeable small-sensor camera with a larger body style and a very large zoom range, often with RAW output.
What you are talking about are the newer crop of all-digital large-sensor interchangeable lenses cameras by Panasonic/Olympus (Micro Four Thirds), Samsung (NX system) and Sony (NEX system). There still isn't a consensus name for this type of camera, but here's the common names you hear:
If you are in the really cheap seats, you should probably spend whatever audio money you have on speakers or headphones that don't utterly suck.
No, I have to disagree on this excessively general prescription (even though you did stick the "probably" in there). I've heard on-board audio that was so bad that clearly it was more important to replace the audio output than to upgrade the headphones or speakers. I had a ThinkPad once whose output was very hissy and had interference from other components in the system, and it was all clearly audible with crappy headphones.
Basically, the most general prescription that can be made is to identify the weak link and upgrade that first. If you hear hiss or interference, get a reasonable sound card. It doesn't have to be fancy; it just has to be competent.
Actually, it is not necessary to use an additional "s" to form a possesive with words that end with an "s" sound. Several sources say that it is prefered, however I find that it is more readable without the additional "s".
For the non-grammar-Nazis out there, I think it's worth tossing into this discussion a brief descriptive account of how possessive marking works in English (i.e., in real, spoken English, not in the crappy conventional orthographic renderings thereof).
In English, a noun consists of an obligatory stem and an optional inflectional ending. So, for example, in the word dogs, dog- is the stem, and -s is the inflectional ending—which in this case serves as a plural marker. Other examples:
So, some rules now:
So, the orthographic rule that renders as wuss's the possessive-marked singular noun is the one that most closely accords with the rules of English grammar.
I worked on some code done by someone else, where on massive records, they were always selecting "*" instead of the needed or anticipated values. Big waste when one needs (by ID#) last and first name and selects a whole row instead - then wonders why it's not scaling upwards.
Eh, I wonder if you're overstating the performance implications of that. Those are all row-oriented databases. Unless all of the columns your query needs are found in an index, it's going to have to read the whole row from disk anyway; the extra costs from the * then become (a) memory and CPU usage and (b) network bandwidth. In my experience, network bandwidth is usually not a big problem; memory and CPU usage can be an issue, but the big performance killers tend to be inefficient joins (because they don't scale linearly), while scalar stuff (the which the * would fall into) are usually cheap.
Nope, not overstating anything.
Let's say it's an ambulance company database that's used to calculate their LOSAP points for the year... that requires calculations from EVERY data parameter input, since everything a member does goes towards their LOSAP points. In that data are things like their PCRs (Patient Care Reports). Each PCR may have .1MB (not counting scans) of data associated with it. Let's say there are 200 members and 24,000 PCRs. Now... let's assume the server has 2GB of RAM. That's 2.4GB of data to read just from the PCR tables alone if one loads all columns. Or a twenty minute report. Even stepping through record by record (due to the overhead of 24,000 individual read requests JUST for the PCR data - even when using the same DB connection).
Your explanation doesn't narrow down at all the cause of the speedup that you're experiencing. If that 2.4 GB case refers to the data in one table, the database, unless the query and schema fit some narrow conditions (e.g., all of the 8 columns your query wants are stored in the same index), is still reading the 2.4 gig of data. This is because all of the data for each row is stored together in disk; you can't (normally) just read the 8 columns you want.
There is at least one more factor missing from your explanation (which, to be frank, I find hopelessly vague) that's just as essential to explain the speedup you're seeing. One (wild) guess: you're doing some large joins, the database needs to materialize intermediate join results, and losing the stars means those intermediate result sets become at lot smaller. Another (also wild) guess: your application is using the database as a dumb data store, pulling rows one by one from the DB, processing them individually in the application server. That's inevitably going to be very slow.
If the schools realized that it's 2010, not 1810, and if teachers actually were a bit more passionated about learning than a corpse i'm certain cheating would drop a fair bit.
I don't normally criticise people for language and grammar, since it is beside the point, but I think since you are criticising university teaching quality and seem to imply that you are a student on one, it is fair in this case. So, don't you mean to say something like "If teachers were a bit more passionate (note the form of the word) about teaching (teacher may learn, but they are supposed to teach)"?
Um, notice the placement of the "e" and the "d" keys in the keyboard. And, have you considered the possibility that GP's problem really is teachers who aren't passionate about learning? Why do you assume that GP must have meant "teaching"?
It would lend more credibility to your arguments if you didn't commit such sloppy errors.
Why? How does it lend credibility to an argument to write it in a way that pleases somebody who's looking for excuses not to evaluate the freaking argument anyway?