My colleagues look at me with utter horror, as if I were some nut follower of David Duke, what I most surely am not. Then I say: "See why this is silly? I used your exact phrase, only switching the subjects."
The problem with your logic here is that "white" people don't have barriers to entry and advancement in the fields you're talking about. That's the whole point you and so many others are missing; it's not the relative proportions of ethnicities or the genders, it's the barriers. Overrepresentation of one group that faces no barriers to entry is OK; underrepresentation of a group that faces such barriers is bad.
Well, duh, because they need to get the other side of town, and their circumstances are such that the expected value of walking or cycling is way lower than that of taking the bus.
In any case, overloaded, unsafe vans driving on dangerous roads are one of the world's major forms of transportation. Which just goes to show GP's point--what good is all that market fundamentalism if the predictions are contradicted by, oh, most public transport in the world?
Zero's are only needed in a system such as our where digit value is context specific (i.e. the "1" in "100" means something different than the "1" in "10") - the roman numeric system doesn't work this way.
Actually, they're not needed. Imagine that instead of a 0 digit, we had an "X" digit whose value was ten. We'd get something like this:
I don't really understand why it should be hard to Zoom - as you describe it - right. You as programmer know about the content within the window, so you have all the information. Interestingly enough I never really saw this behavior happening. So far I really always see maximize till the first dimension hits the border.
Open Safari, resize the window to be pretty big, go to the Google front page, and hit the Zoom button. In Tiger, what I observe is that the zoom button reduces the size of the my window to match the content. The reason you see Zoom so often setting your window to maximum vertical size is that most pages are vertically much longer than the screen. However, for applications where open documents have more modest sizes relative to the screen, zoom is very useful, especially on very large screens.
Though I must say that I've always hated Preview's behavior when Zooming a PDF--it never resizes the window to the size that I'd pick to read it onscreen, and it often makes the PDFs way too small. (I think it's because of an attempt of showing the document at life size, but life size often is something like 4" square for some PDFs.)
"RAW" photos are a lossless capture, which means they are larger files (bad) but with few of the artifacts produced by JPEG compression, and thus your editing options are greatly increased (good).
That, while strictly true, is not the correct answer. RAW images are the raw data from the camera sensor, with little if any processing. That raw data must be processed to generate a raster image like a jpeg, tiff or png.
Nearly all cameras come set up so that the RAW -> JPEG conversion happens on the camera. This conversion, in a good camera, results in two things:
Loss of information--JPEG only supports 8 bits per color channel per pixel; camera sensors often have more dynamic range than that.
RAW -> JPEG conversion performs interpolation, which adds information not present in the raw data.
Setting your camera to output RAW and processing the RAW files in your computer allows you to control the raw conversion. A RAW converter in your computer can sometimes get more detail, and most usefully, allows you to compensate for under or overexposure to some degree.
I bet you they're not talking about the system stack pointer. Remember, yacc is a parser generator; parsing algorithms always use some sort of stack data structure. So, the "stack pointer" in question is just a plain old pointer, pointing into a stack that yacc's generated code uses.
Of course other people should be able to control how others make use of their property. Nobody's denying that. The question is: where can you exercise your right to free speech in the Internet, without being subject to others' right to control how you make use of their property?
In real life, there exist spaces that are clearly public. In the Internet, there aren't any obvious ones. Even if you try to set up your own site, the various providers may censor you if they choose to do so.
It never ceases to amaze me why people think that the Portuguese situation is somehow "different" from the myriads of different English, French and Spanish variations. The structural differences are actually almost non existing [...]
The Portuguese situation is different from the Spanish and English situations, for sure. Brazilians tend to have a hard time understanding European Portuguese, because they don't get to hear a lot of it. Whereas the Portuguese get to hear plenty of Brazilian music and TV. Seriously, they dub Portuguese TV shows into Brazilian Portuguese. The same thing, but in the opposite colonizer-colony direction, happens with Parisian French and Quebec French--the Quebecois hear plenty of Parisian all the time and understand it fine, while the Parisians hardly ever hear any Quebecois, and can't understand it very well.
Both of these examples definitely have to do with the fact that Brazil and France have bigger populations and more cultural impact than Portugal and Quebec; Paris is the center of the French-speaking world, and while Brazil can't claim to be the center of the Portuguese-speaking world, it has an order of magnitude more lusophones than Portugal. There's no such situation for English and Spanish; none of the urban dialects is isolated from any others.
Structural similarities (which you overstate; the phonology of European and Brazilian Portuguese is substantially different) have less of an effect on mutual intelligibility than you seem to think. You can quite easily fail to understand a language or dialect that's extremely similar to yours if you have very little exposure to it. Where the similarity helps is that you can pick it up in weeks instead of years.
You could argue that the reverse isn't true - which is true - which would actually mirror the experiences of every other European language: the "original" speakers tend to pick up the New World variations a lot better than the opposite, mainly due to the fact that they are a lot more "closed" in terms of used sounds.
Sorry, that's one hella messy sentence that I can't follow. I'll reiterate one of my examples from above: the French in general don't understand Quebecois (nevermind Cajun), but the Quebecois understand the French.
Crosslinguistically, intransitive verbs are classified into two categories: unaccusative and unergative. The correct definition of these terms is a very long (and controversial) topic, but a decent rule of thumb is the following: unergative verbs are verbs where the subject must be an agent (i.e., the doer of the action), while unaccusative verbs are those where the subject might not be an agent. This can get tricky: descendre 'descend' is unaccusative, even though a person may deliberately and agentively descend, say, some stairs. The thing is that being the subject of descendre is compatible with not being the agent behind the action described in the sentence.
The French verbs that use être are unaccusatives (in French, at least; the English counterparts might not be!). Again, this is tricky: partir 'leave, go away' sure sounds like it would be unergative if you had to guess from knowing English, but there are French sentences like La chance de gagner la guerre est partie 'The opportunity to win the war is gone'.
Of course, I won't miss the chance to point out that the people who're always going on about how we should teach grammar in schools have likely never ever heard the terms "unaccusative" and "unergative." Don't overestimate the value of knowing "grammar," because most likely, the people you think do really don't know much.
I largely agree with your attitude. The impression I got from GP's description of "whole-language" English teaching is that he's portraying in many ways the opposite mistake as what traditional grammar teaching; too little attention to grammar. (Note that I've been careful to say it's GP's portrayal; I don't know what the actual proponents of the method in question would say, or whether they would agree with GP, but GP's perception of the method as a front-line teacher is undoubtedly of interest.)
Still, I think you are missing an important component here, which is how an educational system should deal with language variation found in its students. In fact, I think the best arguments for bringing grammar into first-language instruction have to do with the fact that unless you do so, you're in danger of assuming that all your students speak the same standard dialect. Students who don't speak the standard dialect will certainly need to be taught explicit grammatical differences between their dialect and the standard. Students who do speak the standard dialect can still benefit from this knowledge; a non-judgmental lesson on grammatical differences is arguably a more valuable lesson about grammar than what normally passes as "teaching grammar."
The problem, of course, is that the folks who normally rant about how we should teach "proper grammar" to children, apart from not actually knowing grammar well at all, don't actually want our educational system to positively address language variation. They wish to bury their heads in the sand about the problems, and blame them on the students or their parents.
(For the record: I have a graduate degree in linguistics, and I'm even more obnoxious than most linguists. In my mind, for example, you don't really understand what "noun" and "verb" means unless you can explain to me the controversy over whether certain native languages of the Pacific Northwest have nouns or not.)
The distinction that's relevant in the case of the intransitive verbs is unaccusative vs. unergative. (IIRC, it is often argued that the use of être in reflexives is related to this.)
I am skeptical that analyzing the content of the books can lead to good recommendations, let alone "infallible". Two books can be very similar in subject matter and writing style and yet one can be great and the other one awful.
Who uses the algorithm makes a big difference. From the way you frame the problem, you're looking for "good" books, i.e., books you'll enjoy reading. But think of somebody doing academic research or looking for patent prior art--for them, one important task is to find all relevant references on a topic, good or bad, and sifting through them.
1. Invariable variables. This appears to have been done for no reason other than the designer's preference. In fact, it's not strictly true -- variables can be unbound, and later bound. They just can't be re-bound once bound.
Um, have you never heard of functional programming? Erlang's a pure functional language, a design choice that opens up a number of optimization possibilities. The most obvious one is that the fact that since programs can't mutate memory, there is no memory synchronization or locking required. There are other advantages to it, too: operations on pure functional data structures don't destroy older states of the data structure, meaning that no thread ever sees a shared data structure in an inconsistent state.
To tell you the truth, I'd prefer a language where mutable memory is optional (e.g., O'Caml), but I don't think that Erlang's design choice is by any means stupid. I do think the syntax is funny--it was largely copied from Prolog and I don't like Prolog's syntax--but come on, this is also a minor point. The only good point you've brought up is the awful support for Unicode.
Um, I would like to point out that, the depreciating state of our economy has nothing to do with the war in Iraq.
Point out? How about argue? That claim can't be let pass without argument.
The government turned to debt to finance the war at the same time it reduced taxes. You think that's got nothing to do with the economy?
Also, it is extremely difficult to destabilize something that is already unstable. If anything, we have made things more stable.
Oh really? By dismantling the state in one of the largest countries in the region, and failing to put a working successor in place? By giving Turkey reason to intervene militarily in the Kurdish areas of Iraq? By creating a really, really good chance that Iraq will align itself with Iran after we leave?
I would add a third possibility as to why the company failed. The product didn't provide the functionality the company intended to deliver to their targeted market.
That one and the previous ones can be subsumed under a more general heading: sales, marketing and development not being driven by the same goals. It can fail if any of the three components fails: marketing doesn't identify the correct market, development doesn't develop a product correctly tailored to the target market, or sales sells it to the wrong market. And the three kinds of failures can happen at the same time.
Ever heard of an axiom? It's the basis of all logic, mathematics and scientific discourse. This is an axiom, it's objectively untestable.
Pretty much all you've just said there is wrong.
Axioms are no more the basis of logic than rules of inference are. An axiomatic theory consists of a language, axioms, and rules of inference that establish what the logical consequences of the axioms are.
The idea that axiomatic theories are the basis of all logic, mathematics and scientific discourse is, at best, arguable. The alternative position is that axiomatic systems serve as a tool to organize our knowledge and provide a method to expand upon it. Historically, axiomatization of a branch of mathematics comes after people have acquired a body of knowledge about it.
Euclid didn't wake up one day and become struck by 5 self-evident truths of geometry; rather, he systematized geometrical results that were obtained by his forebears. Organizing geometry as the system of logical consequences of 5 axioms means that you can now group theorems of geometry into 32 logically equivalent sets, depending on which of the 5 axioms they depend on.
This entails that we have the ability to have mathematical knowledge without axiomatization. In fact, we couldn't possibly favor one set of axioms over another, contradictory one if we didn't have pre-axiomatic mathematical knowledge. There are many controversies in mathematics over whether a given axiomatization is "true"; witness alternative mathematics like relevant logic, or constructive analysis, which argue that mainstream logic and mathematics have picked axioms that are false.
Of course, all that I've just said makes short work of the idea that an axiom must be "objectively untestable." An axiomatic system must organize and help expand our knowledge of the mathematics in question, no more and no less.
"There are ten people in there. None, some, or all of them may have legal firearms. I have a gun. If I rob that place, what are the odds that I am outgunned? even worse, what are the odds that I am LOOKING at the ONE guy drawing a gun on me out of ten in the place?"
"Well, if I can't be certain that the odds are stacked in my favor when robbing this place, then I'll just keep looking for a time and place when they are." If the noontime store with 10 people in it is too risky to rob, the criminal will rob the store when only the shopkeeper is there.
The uncertainty about possession of firearms has a strong deterrent effect - I'd rather rob someone who is unarmed than someone who IS armed. The possibility that ANYONE may be armed is enough to convince me to take up a new hobby, like identity theft, instead of armed robbery.
The criminal always has the initiative, remember. He gets to choose the time and place that the crime happens. This advantage, if used with just a little bit of skill, negates the victims' weapons. You'll only be able to use a weapon to defend yourself if the criminal makes an error. Counting on that is not a good plan.
FYI, although Mac OS X probably could replace 95% of the OS without rebooting, it doesn't.
That's true, but the GP started this thread down a bad direction that you're following: assuming that the scenario is OS updates. Gates' email's scenario is about installing an application for the first time.
Most OS X apps don't even have installers in the sense that Windows apps do. You download a disk image file, open it, drag the app to your applications folder, and presto, you're done. Updating an app? Same process. No reboots. (There are some apps that break this paradigm, true: MS Office, and, IIRC, Acrobat Reader.)
In contrast, all sorts of app installations and upgrades in Windows just demand that you reboot. That is annoying as hell.
I suppose you could start where he, again, tries to present the argument that correlation really is "good enough" - causation be damned.
Not to defend TFA (which I believe I dislike at least as much as you), but the problem with that argument isn't the "causation be damned" part. Causation is a very hard philosophical topic.
No, DNS was good for the early days of the Internet, and succeeded at supporting that. It still also succeeds at providing an official layer of indirection between symbolic host names and IP addresses, to allow the relationship to change or be many-to-many.
We have better solutions nowadays to the "intent" you allude to: search engines and web directories. The real problem isn't that DNS has broken down, it's that we're still clinging on to it as the way to organize the Internet's content.
Try this experiment: turn off the address bar in your browser, and navigate the web with nothing more than your search engine of choice as your home page, and bookmarks to useful sites. That, and its future improvements, are what should "replace" DNS. (That doesn't mean that DNS and URLs go away, but rather, that they become low-level protocols invisible to most users.)
Type in "dell"... which Dell are you looking for? Personally I'm looking for my homepage, not the computer company (same last name as Michael).
So don't use DNS to solve the Internet search problem, use search engines and specialized directories.
Hell, you can try it now. Turn off your browser's address bar, and navigate the web with nothing but bookmarks, search engines and browser history. It works pretty well. Most disadvantages you find will either have better solutions than the address bar, or, in the absolute worst case, are only disadvantages for users who are technically savvy enough that they could turn on an optional address bar.
Slashdot Mirror
The problem with your logic here is that "white" people don't have barriers to entry and advancement in the fields you're talking about. That's the whole point you and so many others are missing; it's not the relative proportions of ethnicities or the genders, it's the barriers. Overrepresentation of one group that faces no barriers to entry is OK; underrepresentation of a group that faces such barriers is bad.
Well, duh, because they need to get the other side of town, and their circumstances are such that the expected value of walking or cycling is way lower than that of taking the bus.
In any case, overloaded, unsafe vans driving on dangerous roads are one of the world's major forms of transportation. Which just goes to show GP's point--what good is all that market fundamentalism if the predictions are contradicted by, oh, most public transport in the world?
Actually, they're not needed. Imagine that instead of a 0 digit, we had an "X" digit whose value was ten. We'd get something like this:
1 2 3 4 5 6 7 8 9 X 11 12 13 14 15 16 17 18 19 1X 21 22 23 24 25 ...
The number 13X64 would be (1 * XXXX) + (3 * XXX) + (X * XX) + (6 * X) + 4. In decimal, (1 * 10,000) + (3 * 1,000) + (10 * 100) + (6 * 10) + 4 = 14064.
Subitization
Open Safari, resize the window to be pretty big, go to the Google front page, and hit the Zoom button. In Tiger, what I observe is that the zoom button reduces the size of the my window to match the content. The reason you see Zoom so often setting your window to maximum vertical size is that most pages are vertically much longer than the screen. However, for applications where open documents have more modest sizes relative to the screen, zoom is very useful, especially on very large screens.
Though I must say that I've always hated Preview's behavior when Zooming a PDF--it never resizes the window to the size that I'd pick to read it onscreen, and it often makes the PDFs way too small. (I think it's because of an attempt of showing the document at life size, but life size often is something like 4" square for some PDFs.)
That, while strictly true, is not the correct answer. RAW images are the raw data from the camera sensor, with little if any processing. That raw data must be processed to generate a raster image like a jpeg, tiff or png.
Nearly all cameras come set up so that the RAW -> JPEG conversion happens on the camera. This conversion, in a good camera, results in two things:
Setting your camera to output RAW and processing the RAW files in your computer allows you to control the raw conversion. A RAW converter in your computer can sometimes get more detail, and most usefully, allows you to compensate for under or overexposure to some degree.
I bet you they're not talking about the system stack pointer. Remember, yacc is a parser generator; parsing algorithms always use some sort of stack data structure. So, the "stack pointer" in question is just a plain old pointer, pointing into a stack that yacc's generated code uses.
Of course other people should be able to control how others make use of their property. Nobody's denying that. The question is: where can you exercise your right to free speech in the Internet, without being subject to others' right to control how you make use of their property?
In real life, there exist spaces that are clearly public. In the Internet, there aren't any obvious ones. Even if you try to set up your own site, the various providers may censor you if they choose to do so.
The Portuguese situation is different from the Spanish and English situations, for sure. Brazilians tend to have a hard time understanding European Portuguese, because they don't get to hear a lot of it. Whereas the Portuguese get to hear plenty of Brazilian music and TV. Seriously, they dub Portuguese TV shows into Brazilian Portuguese. The same thing, but in the opposite colonizer-colony direction, happens with Parisian French and Quebec French--the Quebecois hear plenty of Parisian all the time and understand it fine, while the Parisians hardly ever hear any Quebecois, and can't understand it very well.
Both of these examples definitely have to do with the fact that Brazil and France have bigger populations and more cultural impact than Portugal and Quebec; Paris is the center of the French-speaking world, and while Brazil can't claim to be the center of the Portuguese-speaking world, it has an order of magnitude more lusophones than Portugal. There's no such situation for English and Spanish; none of the urban dialects is isolated from any others.
Structural similarities (which you overstate; the phonology of European and Brazilian Portuguese is substantially different) have less of an effect on mutual intelligibility than you seem to think. You can quite easily fail to understand a language or dialect that's extremely similar to yours if you have very little exposure to it. Where the similarity helps is that you can pick it up in weeks instead of years.
Sorry, that's one hella messy sentence that I can't follow. I'll reiterate one of my examples from above: the French in general don't understand Quebecois (nevermind Cajun), but the Quebecois understand the French.
So good luck changing the pronunciation of a language to match its orthography.
Crosslinguistically, intransitive verbs are classified into two categories: unaccusative and unergative. The correct definition of these terms is a very long (and controversial) topic, but a decent rule of thumb is the following: unergative verbs are verbs where the subject must be an agent (i.e., the doer of the action), while unaccusative verbs are those where the subject might not be an agent. This can get tricky: descendre 'descend' is unaccusative, even though a person may deliberately and agentively descend, say, some stairs. The thing is that being the subject of descendre is compatible with not being the agent behind the action described in the sentence.
The French verbs that use être are unaccusatives (in French, at least; the English counterparts might not be!). Again, this is tricky: partir 'leave, go away' sure sounds like it would be unergative if you had to guess from knowing English, but there are French sentences like La chance de gagner la guerre est partie 'The opportunity to win the war is gone'.
Of course, I won't miss the chance to point out that the people who're always going on about how we should teach grammar in schools have likely never ever heard the terms "unaccusative" and "unergative." Don't overestimate the value of knowing "grammar," because most likely, the people you think do really don't know much.
I largely agree with your attitude. The impression I got from GP's description of "whole-language" English teaching is that he's portraying in many ways the opposite mistake as what traditional grammar teaching; too little attention to grammar. (Note that I've been careful to say it's GP's portrayal; I don't know what the actual proponents of the method in question would say, or whether they would agree with GP, but GP's perception of the method as a front-line teacher is undoubtedly of interest.)
Still, I think you are missing an important component here, which is how an educational system should deal with language variation found in its students. In fact, I think the best arguments for bringing grammar into first-language instruction have to do with the fact that unless you do so, you're in danger of assuming that all your students speak the same standard dialect. Students who don't speak the standard dialect will certainly need to be taught explicit grammatical differences between their dialect and the standard. Students who do speak the standard dialect can still benefit from this knowledge; a non-judgmental lesson on grammatical differences is arguably a more valuable lesson about grammar than what normally passes as "teaching grammar."
The problem, of course, is that the folks who normally rant about how we should teach "proper grammar" to children, apart from not actually knowing grammar well at all, don't actually want our educational system to positively address language variation. They wish to bury their heads in the sand about the problems, and blame them on the students or their parents.
(For the record: I have a graduate degree in linguistics, and I'm even more obnoxious than most linguists. In my mind, for example, you don't really understand what "noun" and "verb" means unless you can explain to me the controversy over whether certain native languages of the Pacific Northwest have nouns or not.)
The distinction that's relevant in the case of the intransitive verbs is unaccusative vs. unergative. (IIRC, it is often argued that the use of être in reflexives is related to this.)
Then how come the vowel in the first syllable is pronounced different from the vowel in the second one, yet they're spelled with the same letter?
Who uses the algorithm makes a big difference. From the way you frame the problem, you're looking for "good" books, i.e., books you'll enjoy reading. But think of somebody doing academic research or looking for patent prior art--for them, one important task is to find all relevant references on a topic, good or bad, and sifting through them.
Um, have you never heard of functional programming? Erlang's a pure functional language, a design choice that opens up a number of optimization possibilities. The most obvious one is that the fact that since programs can't mutate memory, there is no memory synchronization or locking required. There are other advantages to it, too: operations on pure functional data structures don't destroy older states of the data structure, meaning that no thread ever sees a shared data structure in an inconsistent state.
To tell you the truth, I'd prefer a language where mutable memory is optional (e.g., O'Caml), but I don't think that Erlang's design choice is by any means stupid. I do think the syntax is funny--it was largely copied from Prolog and I don't like Prolog's syntax--but come on, this is also a minor point. The only good point you've brought up is the awful support for Unicode.
Point out? How about argue? That claim can't be let pass without argument.
The government turned to debt to finance the war at the same time it reduced taxes. You think that's got nothing to do with the economy?
Oh really? By dismantling the state in one of the largest countries in the region, and failing to put a working successor in place? By giving Turkey reason to intervene militarily in the Kurdish areas of Iraq? By creating a really, really good chance that Iraq will align itself with Iran after we leave?
That one and the previous ones can be subsumed under a more general heading: sales, marketing and development not being driven by the same goals. It can fail if any of the three components fails: marketing doesn't identify the correct market, development doesn't develop a product correctly tailored to the target market, or sales sells it to the wrong market. And the three kinds of failures can happen at the same time.
Pretty much all you've just said there is wrong.
Euclid didn't wake up one day and become struck by 5 self-evident truths of geometry; rather, he systematized geometrical results that were obtained by his forebears. Organizing geometry as the system of logical consequences of 5 axioms means that you can now group theorems of geometry into 32 logically equivalent sets, depending on which of the 5 axioms they depend on.
This entails that we have the ability to have mathematical knowledge without axiomatization. In fact, we couldn't possibly favor one set of axioms over another, contradictory one if we didn't have pre-axiomatic mathematical knowledge. There are many controversies in mathematics over whether a given axiomatization is "true"; witness alternative mathematics like relevant logic, or constructive analysis, which argue that mainstream logic and mathematics have picked axioms that are false.
"Well, if I can't be certain that the odds are stacked in my favor when robbing this place, then I'll just keep looking for a time and place when they are." If the noontime store with 10 people in it is too risky to rob, the criminal will rob the store when only the shopkeeper is there.
The criminal always has the initiative, remember. He gets to choose the time and place that the crime happens. This advantage, if used with just a little bit of skill, negates the victims' weapons. You'll only be able to use a weapon to defend yourself if the criminal makes an error. Counting on that is not a good plan.
That's true, but the GP started this thread down a bad direction that you're following: assuming that the scenario is OS updates. Gates' email's scenario is about installing an application for the first time.
Most OS X apps don't even have installers in the sense that Windows apps do. You download a disk image file, open it, drag the app to your applications folder, and presto, you're done. Updating an app? Same process. No reboots. (There are some apps that break this paradigm, true: MS Office, and, IIRC, Acrobat Reader.)
In contrast, all sorts of app installations and upgrades in Windows just demand that you reboot. That is annoying as hell.
Not to defend TFA (which I believe I dislike at least as much as you), but the problem with that argument isn't the "causation be damned" part. Causation is a very hard philosophical topic.
Javascript is not a Lisp. I thought this should be obvious enough, but let's point out some key differences:
Calling a language a "Lisp" on the basis of nothing more than first-class functions is silly.
No, DNS was good for the early days of the Internet, and succeeded at supporting that. It still also succeeds at providing an official layer of indirection between symbolic host names and IP addresses, to allow the relationship to change or be many-to-many.
We have better solutions nowadays to the "intent" you allude to: search engines and web directories. The real problem isn't that DNS has broken down, it's that we're still clinging on to it as the way to organize the Internet's content.
Try this experiment: turn off the address bar in your browser, and navigate the web with nothing more than your search engine of choice as your home page, and bookmarks to useful sites. That, and its future improvements, are what should "replace" DNS. (That doesn't mean that DNS and URLs go away, but rather, that they become low-level protocols invisible to most users.)
So don't use DNS to solve the Internet search problem, use search engines and specialized directories.
Hell, you can try it now. Turn off your browser's address bar, and navigate the web with nothing but bookmarks, search engines and browser history. It works pretty well. Most disadvantages you find will either have better solutions than the address bar, or, in the absolute worst case, are only disadvantages for users who are technically savvy enough that they could turn on an optional address bar.