Or at least, they traditionally did, and we haven't really figured out as a whole what we want to transition to.
Historically, only a small percentage of people got university degrees. Professors were (and in many places still are) first and foremost researchers; their real job is not teaching, but advancing their field and publishing their results. A secondary job is research mentorship: they advise and supervise graduate students as the next generation of researchers. A tertiary job is teaching of undergraduate material, to historically only a small percentage of the population that had a need to learn advanced-level stuff from an expert in the field. Those people were generally expected, furthermore, to be interested in and to benefit from a well-rounded education rather than only training in their specific area, e.g. to become scientists who also had an understanding of ethics and history.
Today, it's more or less expected in many areas that you have a college degree. As a result, a lot of people go to university mainly as a sort of certification that qualifies them for jobs. They don't necessarily want the traditional liberal education, even the science version of it; they want vocational training. But universities were not really set up to provide that, and their staff are entirely the wrong ones to provide it: the people publishing CS research papers and the people who would be good at teaching a vocational programming class that prepares one for a role as C++ programmer in industry are only occasionally the same.
That's why we historically had separate trade schools and vocational schools, which did focus on practical skills, and had teachers who were focused on teaching such skills. But there's been a sort of prestige treadmill so companies want you to have a University Degree for a job that actually need vocational training, not a well-rounded liberal-arts education with mentorship from a PhD researcher.
There's a lot of possible solutions, of course. One is to go back to the old model, where universities do research and teach a small percentage of the population, and vocationally focused institutions teach most people. The most likely, though, is probably a gradual morphing of universities into a superset of the two kinds of institutions. Already it's becoming common to hire lecturers to teach introductory classes, and some schools are offering variations on degrees to let students opt between more traditional university majors or more applied vocational majors; often this also leads to a parallel split between staff who are mainly "teaching faculty" versus "research faculty".
Sure, $50-60k out of college isn't bad, but there are plenty of liberal-artsy jobs that start at $40-50k, so it's not an amazing differential. If you want to actually attract people into the field who wouldn't have otherwise come, as tech companies claim they're interested in doing, you have to offer a bigger differential than that. Basically, people have other options, and are taking them. Tech companies don't want to do the capitalist thing to entice new employees into the field (if supply is low at a given price, the price is too low), so instead they whine to Congress about H-1Bs.
Although academia was never really the ivory tower of caricature, where people sat around and smoked pipes and "just thought", it does feel from what I've read that it's less that than ever. Doing things other than The Project With A Deadline is really, I think, necessary for significant progress other than churning out incremental results. Everyone's big on "interdisciplinary" work these days, for example, but that seems to usually mean frantically cribbing something from another field that you can shallowly map to your problem at hand. I think a deeper interdisciplinarity really requires academics to have free time to pursue other interests, where they can run into possible areas of cross-pollination that would never happen in the shallow "I need something from biology to support this grant proposal" style of interdisciplinarity.
But that really ties into the "tons of publications is necessary" culture being bad for all sorts of reasons. It's bad for people raising kids, which is bad for retaining a whole class of people. It's bad for interdisciplinary work and fostering long-term research. It's bad for readers, who have to wade through papers that would never have been published if the author didn't feel compelled to churn them out (sometimes somewhat guiltily).
Perhaps I'm being unreasonable, but I don't want someone doing a life-threatening job who's worked more than about 14-18 hours straight at most. Truck drivers have a maximum shift of 14 hours, of which no more than 11 may be driving. Is the stuff residents do really less error-prone for fatigued people than driving a truck is?
You seem to have basically responded to "residents work ridiculous 36-hour shifts" with "no no it's all better now, nobody works more than a perfectly reasonable 30-hour shift".
Not as true as it used to be, but on average men earn about 50% more than women ($39k median versus $26k median), so are doing considerably more of the monetary providing in aggregate.
I'm not sure what alternate reality you came from, but in the actual United States of the 1970s and 80s, the majority of people who did well for themselves (well enough to send their kids to college, anyway) did not work 70-80-hour weeks. If you want to run you own business, sure, but the salaried office job was not invented yesterday. I know plenty of people who pulled in excellent salaries working 40-hour weeks at companies like Boeing, IBM, and Amoco in the 1970s.
I mostly agree, but I find college culture does vary, especially in more techie fields, and if you meet the right people to take advantage of it. There really is a difference (imo) between a CS degree from a school that has a strong hacker/geek culture and one that doesn't. I'm not going to say everything about hacker/geek culture is optimal, but I think someone who has an exposure to that gains a useful way of looking at technology.
Of course, that doesn't correlate all that well with money; there are plenty of affordable state schools with strong tech cultures.
I have no aversion to mathematical models: I'm an AI researcher with a large focus on machine learning, so it's of course in my interest that lots and lots of people use such models.
But as someone who develops them, it also shocks me what people actually use them for. The hardest part of a complex problem is not usually applying a statistical algorithm, of which there are a lot, many of whose properties are pretty well understood. Rather, it's figuring out how to pose your problem as a statistical one in a way that is remotely valid. Also, getting good data is a big stumbling block.
It's remarkably hard to do this for complex problems; the level of complexity at which modeling is really well understood is not all that high. And your results are worse than useless if you don't even have a data set to begin with whose properties you understand.
In this case, I'd say lack of any data to start with would be the first stumbling block. Is it actually understood what properties of workers are good for the long-term interests of an organization? How do you even define "good", what is your notion of "long-term", and does this vary by organization? Sure, represent worker X as a 20-dimensional vector of numbers; I don't inherently have a problem with that. But how are you predicting the response variable likelihood_of_excellence(X) from that? Do you have some magical labeled data set of 10,000 workers and their long-term excellence values?
I know plenty of people who didn't really give up the mixture of idealism, narcissism, and aversion to being a cog in someone else's machine, even after graduating and entering the "real world". The trick, though, is that you can only really do it in an uncompromising way if you always have an out, so the moment you don't want to be that cog, you really have an alternative and can leave.
I may have a rather warped view of this, since a disproportionate number of my friends and acquaintances are Silicon Valley techies. It's not a free pass by any means--- the easiest way to pull in a good salary is still to work for some large tech company. But it's surprisingly easy to make enough off ad revenue to support a modest lifestyle without bosses or a "real job".
Pretty much any mainstream economist will tell you that productivity has increased enormously over the past century, due to a combination of factors, technology probably being the biggest. Productivity increases decrease the level of scarcity for any fixed basket of goods to which they apply, because more stuff is produced than previously without an increase in resources.
Of course, you can take that "productivity dividend" in various forms. One way to maintain the illusion of scarcity is to increase your baseline of what you "need", so you always need the things that have just barely become affordable. Then scarcity is definitionally constant, because what you're really doing is holding scarcity fixed and varying your basket of goods accordingly.
The netbook trend shows the opposite way you can take the productivity dividend: hold fixed the things you "need", and enjoy the ever-decreasing scarcity by having to give fewer resources (i.e. hours of work) to get those same goods. Applied to other areas, it's quite possible to reduce the amount of work people have to do on average, as long as you increase the "need" baseline slower than the gains from better productivity decrease scarcity. Typically people haven't done that: do Americans use the productivity increases of the past 50 years to work fewer hours? No, they generally use them to increase material consumption; e.g. the average house size has nearly doubled. But that isn't entirely necessary.
Of course, Bertrand Russell went over all of this in 1932, so it's not particularly novel.
I'd say very few companies have a sense of loyalty to their employees, so it seems a bit unreasonable to expect the reverse without the quid pro quo. Maybe in a mom-and-pop shop, but not in a mega-corporation.
I'd say events of the past year or so have if anything validated the idea of "management as top-heavy, who do little and skim money from everything". Dozens of companies didn't fail because their employees were insufficiently invested in their companies; they failed because management fucked up, sometimes due to incompetence, and sometimes due to fraud.
The reason schools are making their engineering programs do more interesting-sounding things are because engineering and CS enrollments, especially among US citizens, are dropping rapidly. So schools are trying to find creative ways to interest people in majoring in those areas; "training for boring cubicle job", funnily enough, doesn't entice people.
The only other solution, really, is the capitalist one: offer so much money that people will go into the field even if it does sound boring. But you need to offer a lot more than current going rates for that.
I've never seen a cable-ISP contract that provided service at a specified rate in Mbps. You can get those contracts as a business user, but they're not the standard ones home users have. Usually home contracts say something along the lines of "up to xx Mbps; actual speeds may vary and are not guaranteed".
A widely disliked, tiny minority that the vast majority of the social conservatives in California don't even consider Christian, and they're somehow responsible for what a bit over 50% of California's voters voted for?
Maybe I'm old-fashioned, but I blame the people who voted for the fucking thing, not Teh Mormons. If I were to single them out by religion, the Catholics and Evangelical Protestants would be the two main blocs to blame.
These sorts of ugly scapegoating reactions have been oddly too common, though; the other group that gets blamed, besides the Mormons, are blacks (Andrew Sullivan started that one, I think). Yeah, blacks, who have virtually no political power in California, and constitute a small minority of the state's population are responsible for an election outcome.
Next time, how about blaming the real culprits: white and Hispanic Catholics and white evangelical Protestants.
That was the Democrats circa LBJ's "Great Society" in the 60s. By the 1990s, the pro-business faction was solidly in charge, culminating in 1992 when the head of the Democratic Leadership Council (the pro-business / anti-lefty caucus) won the Democratic nomination. Sort of analogous to New Labour displacing Old Labour in the UK.
There is an English meaning of the word "libel" that's generally understood though; libel laws are passed to encode that in statute, not the other way around. The argument here seems to be that, as a conceptual matter, public disclosure of private facts is qualitatively different from the concept that traditionally is called "libel", which mostly deals with false and harmful statements, and therefore it would be useful if the law covered them under separate statutes instead of muddying the waters.
It appears most states do actually do that, but Massachusetts has a crappy law in this case.
I can see a useful principle of "employee confidentiality", appropriately delineated. But that seems like a separate issue from libel, and should be covered by a separate statute, with violations being prosecuted as "violation of employee confidentiality", not as "libel".
Most commonwealth countries, as you point out, do not have the U.S. principle that truth is an absolute defense against libel. The UK has historically gone the farthest in allowing such suits.
The U.S. actually has a stronger principle in the other direction for public figures: a statement is not necessarily libel even if false and harmful, unless the publisher either knew it to be false, or published with "reckless disregard" for its truth.
I should add that as far as LaTeX diffs go, you can get a result almost as good as Word's "track changes" mode via texdiff. The main reason it's only "almost as good" is that since Word actually saves the changes instead of trying to reconstruct them from the before/after documents, it doesn't screw up quite as much (e.g. misidentifying insertions/deletions versus block moves, especially of shorter fragments of text). But texdiff is good enough for most uses, just not well integrated into SVN's diff system, which itself doesn't even support disconnected operation.
The social-science-oriented areas like HCI tend to use Word. The engineering- and math-oriented areas like machine learning and theory use latex. The other areas seem to have a mix, to the extent that most conferences feel required to offer both Word and latex templates.
The line-oriented diffs of svn are particularly useless for a language in which newlines are often not semantically meaningful. To get anything useful out of this approach, people end up constraining their use of paragraph reflowing, so you end up with crazy hard-to-read.tex files, which is definitely something that should be handled by a revision-control system better.
It also interacts badly with other synchronization methods between multiple machines. I use a laptop and a desktop, and synchronize them with unison. But svn changes their local file formats across platforms and versions, so when the.svn directories get synced between the current Debian and current fink (OS X) versions, stuff gets corrupted. The workaround there is to *not* sync on my side with unison, and instead use "check into svn and check back out" as a synchronization mechanism. But that leads to checking in half-done stuff that gets in other people's way.
Large corporations tend to have a problem that nobody is: 1) responsible for; or 2) even aware of; what's going on in many parts of the organization. This is especially true for long-term responsibility and awareness.
The lower-level managers ideally have the best idea of what's going on in their particular groups, and should be aware of any issues their group is having communicating with other groups in the company. This is especially true of day-to-day annoyances like "we always have problems opening these goddamn files people email us", which tend not to be on the agenda of higher-level management very often.
But this would work best if lower-level managers stayed with the groups for a while; at many large corporations, lower-level management positions are just stepping stones to higher-level management positions, and often get rotated through every 2-3 years. They often never really do figure out what their group is doing in detail; they just want to make sure nothing problematic happens and the figures look good (according to whatever metric the particular company happens to use).
Even in the cases where they do, the same problem continues as you percolate up the chain. Does the middle-level manager overseeing, say, 10 of the small groups stick around long enough to develop a working relationship with the 10 lower-level manager, and a working model of what is going right or wrong in the groups? Often, no.
Basically, a huge percentage of everyone's time is spent just trying to figure out who is in charge of what this month, bringing them up to speed on very basic things, and so on. Forget "rationalize office software"; you're lucky if the meetings aren't about "wait a minute, who are you again? I didn't even know we had a group doing that!"
They do retain more of that old New-England-puritan hangup about alcohol, on which California's a bit more liberal. You can only buy liquor at state-run liquor stores in NH, and even licenses for on-premises consumption of beer and wine have more regulations---in California it's not uncommon to find coffee shops selling alcohol in the evenings, because the license is easy to get and cheap; good luck finding that in NH.
Slashdot Mirror
Or at least, they traditionally did, and we haven't really figured out as a whole what we want to transition to.
Historically, only a small percentage of people got university degrees. Professors were (and in many places still are) first and foremost researchers; their real job is not teaching, but advancing their field and publishing their results. A secondary job is research mentorship: they advise and supervise graduate students as the next generation of researchers. A tertiary job is teaching of undergraduate material, to historically only a small percentage of the population that had a need to learn advanced-level stuff from an expert in the field. Those people were generally expected, furthermore, to be interested in and to benefit from a well-rounded education rather than only training in their specific area, e.g. to become scientists who also had an understanding of ethics and history.
Today, it's more or less expected in many areas that you have a college degree. As a result, a lot of people go to university mainly as a sort of certification that qualifies them for jobs. They don't necessarily want the traditional liberal education, even the science version of it; they want vocational training. But universities were not really set up to provide that, and their staff are entirely the wrong ones to provide it: the people publishing CS research papers and the people who would be good at teaching a vocational programming class that prepares one for a role as C++ programmer in industry are only occasionally the same.
That's why we historically had separate trade schools and vocational schools, which did focus on practical skills, and had teachers who were focused on teaching such skills. But there's been a sort of prestige treadmill so companies want you to have a University Degree for a job that actually need vocational training, not a well-rounded liberal-arts education with mentorship from a PhD researcher.
There's a lot of possible solutions, of course. One is to go back to the old model, where universities do research and teach a small percentage of the population, and vocationally focused institutions teach most people. The most likely, though, is probably a gradual morphing of universities into a superset of the two kinds of institutions. Already it's becoming common to hire lecturers to teach introductory classes, and some schools are offering variations on degrees to let students opt between more traditional university majors or more applied vocational majors; often this also leads to a parallel split between staff who are mainly "teaching faculty" versus "research faculty".
Sure, $50-60k out of college isn't bad, but there are plenty of liberal-artsy jobs that start at $40-50k, so it's not an amazing differential. If you want to actually attract people into the field who wouldn't have otherwise come, as tech companies claim they're interested in doing, you have to offer a bigger differential than that. Basically, people have other options, and are taking them. Tech companies don't want to do the capitalist thing to entice new employees into the field (if supply is low at a given price, the price is too low), so instead they whine to Congress about H-1Bs.
Although academia was never really the ivory tower of caricature, where people sat around and smoked pipes and "just thought", it does feel from what I've read that it's less that than ever. Doing things other than The Project With A Deadline is really, I think, necessary for significant progress other than churning out incremental results. Everyone's big on "interdisciplinary" work these days, for example, but that seems to usually mean frantically cribbing something from another field that you can shallowly map to your problem at hand. I think a deeper interdisciplinarity really requires academics to have free time to pursue other interests, where they can run into possible areas of cross-pollination that would never happen in the shallow "I need something from biology to support this grant proposal" style of interdisciplinarity.
But that really ties into the "tons of publications is necessary" culture being bad for all sorts of reasons. It's bad for people raising kids, which is bad for retaining a whole class of people. It's bad for interdisciplinary work and fostering long-term research. It's bad for readers, who have to wade through papers that would never have been published if the author didn't feel compelled to churn them out (sometimes somewhat guiltily).
Perhaps I'm being unreasonable, but I don't want someone doing a life-threatening job who's worked more than about 14-18 hours straight at most. Truck drivers have a maximum shift of 14 hours, of which no more than 11 may be driving. Is the stuff residents do really less error-prone for fatigued people than driving a truck is?
You seem to have basically responded to "residents work ridiculous 36-hour shifts" with "no no it's all better now, nobody works more than a perfectly reasonable 30-hour shift".
Not as true as it used to be, but on average men earn about 50% more than women ($39k median versus $26k median), so are doing considerably more of the monetary providing in aggregate.
I'm not sure what alternate reality you came from, but in the actual United States of the 1970s and 80s, the majority of people who did well for themselves (well enough to send their kids to college, anyway) did not work 70-80-hour weeks. If you want to run you own business, sure, but the salaried office job was not invented yesterday. I know plenty of people who pulled in excellent salaries working 40-hour weeks at companies like Boeing, IBM, and Amoco in the 1970s.
I mostly agree, but I find college culture does vary, especially in more techie fields, and if you meet the right people to take advantage of it. There really is a difference (imo) between a CS degree from a school that has a strong hacker/geek culture and one that doesn't. I'm not going to say everything about hacker/geek culture is optimal, but I think someone who has an exposure to that gains a useful way of looking at technology.
Of course, that doesn't correlate all that well with money; there are plenty of affordable state schools with strong tech cultures.
I have no aversion to mathematical models: I'm an AI researcher with a large focus on machine learning, so it's of course in my interest that lots and lots of people use such models.
But as someone who develops them, it also shocks me what people actually use them for. The hardest part of a complex problem is not usually applying a statistical algorithm, of which there are a lot, many of whose properties are pretty well understood. Rather, it's figuring out how to pose your problem as a statistical one in a way that is remotely valid. Also, getting good data is a big stumbling block.
It's remarkably hard to do this for complex problems; the level of complexity at which modeling is really well understood is not all that high. And your results are worse than useless if you don't even have a data set to begin with whose properties you understand.
In this case, I'd say lack of any data to start with would be the first stumbling block. Is it actually understood what properties of workers are good for the long-term interests of an organization? How do you even define "good", what is your notion of "long-term", and does this vary by organization? Sure, represent worker X as a 20-dimensional vector of numbers; I don't inherently have a problem with that. But how are you predicting the response variable likelihood_of_excellence(X) from that? Do you have some magical labeled data set of 10,000 workers and their long-term excellence values?
I know plenty of people who didn't really give up the mixture of idealism, narcissism, and aversion to being a cog in someone else's machine, even after graduating and entering the "real world". The trick, though, is that you can only really do it in an uncompromising way if you always have an out, so the moment you don't want to be that cog, you really have an alternative and can leave.
I may have a rather warped view of this, since a disproportionate number of my friends and acquaintances are Silicon Valley techies. It's not a free pass by any means--- the easiest way to pull in a good salary is still to work for some large tech company. But it's surprisingly easy to make enough off ad revenue to support a modest lifestyle without bosses or a "real job".
Pretty much any mainstream economist will tell you that productivity has increased enormously over the past century, due to a combination of factors, technology probably being the biggest. Productivity increases decrease the level of scarcity for any fixed basket of goods to which they apply, because more stuff is produced than previously without an increase in resources.
Of course, you can take that "productivity dividend" in various forms. One way to maintain the illusion of scarcity is to increase your baseline of what you "need", so you always need the things that have just barely become affordable. Then scarcity is definitionally constant, because what you're really doing is holding scarcity fixed and varying your basket of goods accordingly.
The netbook trend shows the opposite way you can take the productivity dividend: hold fixed the things you "need", and enjoy the ever-decreasing scarcity by having to give fewer resources (i.e. hours of work) to get those same goods. Applied to other areas, it's quite possible to reduce the amount of work people have to do on average, as long as you increase the "need" baseline slower than the gains from better productivity decrease scarcity. Typically people haven't done that: do Americans use the productivity increases of the past 50 years to work fewer hours? No, they generally use them to increase material consumption; e.g. the average house size has nearly doubled. But that isn't entirely necessary.
Of course, Bertrand Russell went over all of this in 1932, so it's not particularly novel.
I'd say very few companies have a sense of loyalty to their employees, so it seems a bit unreasonable to expect the reverse without the quid pro quo. Maybe in a mom-and-pop shop, but not in a mega-corporation.
I'd say events of the past year or so have if anything validated the idea of "management as top-heavy, who do little and skim money from everything". Dozens of companies didn't fail because their employees were insufficiently invested in their companies; they failed because management fucked up, sometimes due to incompetence, and sometimes due to fraud.
In tech, see, you don't need to "build a business" or "launch a product", just "convince some VCs".
The reason schools are making their engineering programs do more interesting-sounding things are because engineering and CS enrollments, especially among US citizens, are dropping rapidly. So schools are trying to find creative ways to interest people in majoring in those areas; "training for boring cubicle job", funnily enough, doesn't entice people.
The only other solution, really, is the capitalist one: offer so much money that people will go into the field even if it does sound boring. But you need to offer a lot more than current going rates for that.
I've never seen a cable-ISP contract that provided service at a specified rate in Mbps. You can get those contracts as a business user, but they're not the standard ones home users have. Usually home contracts say something along the lines of "up to xx Mbps; actual speeds may vary and are not guaranteed".
A widely disliked, tiny minority that the vast majority of the social conservatives in California don't even consider Christian, and they're somehow responsible for what a bit over 50% of California's voters voted for?
Maybe I'm old-fashioned, but I blame the people who voted for the fucking thing, not Teh Mormons. If I were to single them out by religion, the Catholics and Evangelical Protestants would be the two main blocs to blame.
These sorts of ugly scapegoating reactions have been oddly too common, though; the other group that gets blamed, besides the Mormons, are blacks (Andrew Sullivan started that one, I think). Yeah, blacks, who have virtually no political power in California, and constitute a small minority of the state's population are responsible for an election outcome.
Next time, how about blaming the real culprits: white and Hispanic Catholics and white evangelical Protestants.
That was the Democrats circa LBJ's "Great Society" in the 60s. By the 1990s, the pro-business faction was solidly in charge, culminating in 1992 when the head of the Democratic Leadership Council (the pro-business / anti-lefty caucus) won the Democratic nomination. Sort of analogous to New Labour displacing Old Labour in the UK.
There is an English meaning of the word "libel" that's generally understood though; libel laws are passed to encode that in statute, not the other way around. The argument here seems to be that, as a conceptual matter, public disclosure of private facts is qualitatively different from the concept that traditionally is called "libel", which mostly deals with false and harmful statements, and therefore it would be useful if the law covered them under separate statutes instead of muddying the waters.
It appears most states do actually do that, but Massachusetts has a crappy law in this case.
I can see a useful principle of "employee confidentiality", appropriately delineated. But that seems like a separate issue from libel, and should be covered by a separate statute, with violations being prosecuted as "violation of employee confidentiality", not as "libel".
Most commonwealth countries, as you point out, do not have the U.S. principle that truth is an absolute defense against libel. The UK has historically gone the farthest in allowing such suits.
The U.S. actually has a stronger principle in the other direction for public figures: a statement is not necessarily libel even if false and harmful, unless the publisher either knew it to be false, or published with "reckless disregard" for its truth.
I should add that as far as LaTeX diffs go, you can get a result almost as good as Word's "track changes" mode via texdiff. The main reason it's only "almost as good" is that since Word actually saves the changes instead of trying to reconstruct them from the before/after documents, it doesn't screw up quite as much (e.g. misidentifying insertions/deletions versus block moves, especially of shorter fragments of text). But texdiff is good enough for most uses, just not well integrated into SVN's diff system, which itself doesn't even support disconnected operation.
The social-science-oriented areas like HCI tend to use Word. The engineering- and math-oriented areas like machine learning and theory use latex. The other areas seem to have a mix, to the extent that most conferences feel required to offer both Word and latex templates.
The line-oriented diffs of svn are particularly useless for a language in which newlines are often not semantically meaningful. To get anything useful out of this approach, people end up constraining their use of paragraph reflowing, so you end up with crazy hard-to-read .tex files, which is definitely something that should be handled by a revision-control system better.
It also interacts badly with other synchronization methods between multiple machines. I use a laptop and a desktop, and synchronize them with unison. But svn changes their local file formats across platforms and versions, so when the .svn directories get synced between the current Debian and current fink (OS X) versions, stuff gets corrupted. The workaround there is to *not* sync on my side with unison, and instead use "check into svn and check back out" as a synchronization mechanism. But that leads to checking in half-done stuff that gets in other people's way.
Large corporations tend to have a problem that nobody is: 1) responsible for; or 2) even aware of; what's going on in many parts of the organization. This is especially true for long-term responsibility and awareness.
The lower-level managers ideally have the best idea of what's going on in their particular groups, and should be aware of any issues their group is having communicating with other groups in the company. This is especially true of day-to-day annoyances like "we always have problems opening these goddamn files people email us", which tend not to be on the agenda of higher-level management very often.
But this would work best if lower-level managers stayed with the groups for a while; at many large corporations, lower-level management positions are just stepping stones to higher-level management positions, and often get rotated through every 2-3 years. They often never really do figure out what their group is doing in detail; they just want to make sure nothing problematic happens and the figures look good (according to whatever metric the particular company happens to use).
Even in the cases where they do, the same problem continues as you percolate up the chain. Does the middle-level manager overseeing, say, 10 of the small groups stick around long enough to develop a working relationship with the 10 lower-level manager, and a working model of what is going right or wrong in the groups? Often, no.
Basically, a huge percentage of everyone's time is spent just trying to figure out who is in charge of what this month, bringing them up to speed on very basic things, and so on. Forget "rationalize office software"; you're lucky if the meetings aren't about "wait a minute, who are you again? I didn't even know we had a group doing that!"
Because energy is a useless fiat commodity, while you can eat cold, hard dollar bills.
They do retain more of that old New-England-puritan hangup about alcohol, on which California's a bit more liberal. You can only buy liquor at state-run liquor stores in NH, and even licenses for on-premises consumption of beer and wine have more regulations---in California it's not uncommon to find coffee shops selling alcohol in the evenings, because the license is easy to get and cheap; good luck finding that in NH.