Unless there's a signed "non-compete" document from Hurd, HP will just have to live with their mess up.
To the extent that that's true, a signed "non-compete" won't help. Non-compete agreements -- except those executed by owners at the time they sell their stake in a business to a new owner -- are prohibited and declared void by California state law (see CA Bus & Prof Code, Sec. 16600-16602.5)
Of course, if you raise the top marginal tax rate to 90%, total revenue from income tax will go down.
Not in the short term. In the long-term, as income distribution levelled, it might tend back toward the pre-hike level or below, but that would also correspond to less poverty and demand for poverty support programs, so you'd see lower spending to acheive the same policy goals as well as seeing lower revenues.
You apparently don't realize that the wealthiest 1% of income earners pay a larger share of Federal revenues today than the same group did in the 50s.
Yes, because the degree to which the wealthiest 1% out-earn the rest of the population has gone up since then more than the tax rates they pay have gone down. Its not surprising that that has occurred, since lower tax rates on the richest have a compounded effect over time on their wealth.
I can't think of a reason that one would expose a "set" method that just sets a private variable to the passed value.
It future-proofs the API a bit (this may be overengineering, but its a common enough evolution that assignment would develop side effects in the future that it may be worth considering.)
Also, while this isn't what the GP was doing (since it also had a getter) its the only way to create a write-only attribute, which is sometimes useful (since if you expose the variable, its a read/write attribute.)
Which is? If it's along the lines of "It's easier and cheaper to fix a blueprint than a building", then it does not apply to software.
I've seen studies which claimed that there was, IIRC, approximately an order of magnitude increase in cost to repair design defects in software identified once code has been built around it rather than in design, another order of magnitude for design defects identified in acceptance testing, and another order of magnitude for design defects caught once a system is in production.
So I'd like to see something more than an assertion that this saying does not apply to software.
Our election system does tend to narrow the range of ideas that government includes, but that has the benefit of (weakly) discouraging representation of nutcases.
Actually, it encourages representation of nutcases. Specifically, it naturally tends to a two-party system, and encourages both parties to seek an unshakeable irrational ideological core and to seek to use negative tactics to reduce the voting probability of the non-core parts of the electorate that might lean toward the other party; in a system that tends to more than two viable parties on a long-term basis, negative techniques don't work as well (and negative techniques that disaffect voters on both sides, but your main opponents slightly more than your own are counterproductive rather than beneficial as they are in a two-party system).
It may make more people unhappy with the government, but it tends to make it less likely that really damaging ideas become law.
No, it makes people more unhappy with government because it makes more damaging ideas become law, and puts more people who are damaged by those ideas in a position of having no effective recourse than would be the case in a system with more effective representation.
Those policies are there for a valid reason and it is refusing to admit it that keeps real immigration reform from happening.
No, what keeps real immigration reform from happening is: 1. Many of the people who claim to want some kind of real immigration reform (esp. politicians) actually don't, they just like to say they do. Having a broken system that everyone hates to keep blaming things on and using as an excuse to sell policies that aren't actually aimed at fixing the things that cause the problems that people experience, but that advance other interests, is politically useful.
2. Of the people that do want "real immigration reform", they don't actually share the same values, or perceive the same problems, and the "real immigration reform" that any one subgroup wants isn't just different policy than other subgroups want, its directed at diametrically opposed goals.
There is a subgroup which is fundamentally anti-immigrant, and sees the fundamental problem as the level of immigration, and therefore isn't interested in reforming the system of legal immigration, they are interested only in trying to find a magic bullet that will make the existing legal limits more practical to enforce.
There are a number of overlapping subgroups that are fundametally motivated by racial, ethnic, cultural, or religious hatred and see the fundamental problem as being too much of the "wrong kind" of people immigrating. Tactically these groups often align with each other and the preceding group, but they are distinct.
There's a group that views the basic principles underlying the family-relationship-based categories at the core of the main part of our legal immigration system as fundamentally sound, but finds that various elements of our current immigration policy undercut the purpose of that policy (which is in the recognition of the strength of family bonds and leveraging of them to anchor American identity in immigrants.)
There's a group that sees our current immigration policy as problematic because of the cost of enforcement compared to the perceived benefit, whether or not they share any of the other groups preferences on the overall principles.
There's a group that sees the main problem in our immigration policy is that it doesn't do enough to serve business owner's interest in cheap labor.
All of these groups want "real immigration reform" of one kind or another, and all of them are fertile targets for politicians who want to blame problems on the lack of such reform. But once you get into the specifics of a proposed reform, or even the goals it should seek, almost anything you propose is going to be seen as at best irrelevant and at worst counterproductive by more people than will support it, unless you are clever and vague and manage to let everyone read their own preferences into your vague pronouncements.
f you want to open up legal immigration then open it up but offering lifetime visas that must be renewed with language proficiency tests, criminal background reviews, and non-delinquent taxes every 10 years or so. These initial immigrants would never be given political power to vote in any level of city, county, state, or federal government but otherwise would enjoy the same status and rights as citizens.
If you reread GP, you will note that I never anywhere in it advocated increasing the overall level of legal immigration, I advocated maintaining the current overall limits but balancing the country-by-country limits so that the number of immigrants allowed in each limited category by country would be aligned with the number of applicants from each country, so that you don't cause really long delays in the countries that have the most otherwise-qualified prospective immigrants. A (for instance) 2 year wait everywhere would create much less pressure to evade the system than having negligible waits from most of the world, but decade plus waiting lists from a handful of countries
Sure... and then you get Obama, who's about as middle of the road as they come, hence why no one is happy.
I think its more likely that people are unhappy with our government, in proximate terms, because the economy sucks and the government at the time, rightly or wrongly, tends to disproportionately get the blame or credit for the economy. In more general, long term, terms, I think people are unhappy in the US with government because we have an electoral system which greatly limits the scope of ideas represented in the government and produces a government that does not effectively represent the preferences of the populace, which is borne out by the research on popular opinion of government across stable democratic states and the relation it bears to the electoral systems and the number of competitive parties.
Maybe if the government wasn't five years backlogged on immigration requests, more people would come here legally.
This somewhat critically misstates the problem.
It is not that the government is "backlogged" on immigration requests, which suggests that there is a pile of paperwork that just isn't getting processed because of insufficient resources, and if the government would just process the requests it has and execute the existing policy efficiently, there wouldn't be a problem.
The issue is that the US has an immigration system that involves per country limits on immigration (as I recall, part of that is a limit that no one country may have more than 5% of the people admitted in each year in any of the numerically-limited, relationship-based immigration categories -- note that H1Bs aren't part of this) which results in a surplus of requests from prospective immigrants who are legally qualified to immigrate resulting in long (more than a decade in some cases) wait for admission in certain categories from certain countries (this is true, IIRC, in all of the numerically-limited relationship-based categories for Mexico, as well as certain categories from India and some other countries.)
You could reduce, though not eliminate, the problem of new illegal immigration overnight without changing general immigration policy or overall legal immigration limits, simply by using overall limits without limited per-country allocation.
but by the early 80s he was asserting that having a concept for "null" was a requirement of a system that was "really relational".
And by the 1990, he was arguing that a relational system needed to have at least two different kinds of Nulls to operate correctly.
(Apparently, the beginning of the start of something in the db world which culminated with the proposal of requiring support for up to 128 types of distinguished NULLs in the early part of the standardization process of SQL-99.)
So, outside of the 1980s, Codd would seem to view SQL's use of NULL as a problem. Not that I'm find of arguments to authority.
The reason was that while you can have a nice, self-consistent mathematical model of relations without NULL, a practical system absolutely needs some kind of NULL.
Practical systems need to be able to admit of records that a short of ideal because some data is either missing, unknown, or otherwise different from the basic assumptions.
Each of those possibilities -- missing, unknown, and each other possibility of difference from the basic assumptions -- can be identified with an actual value without resort to special non-values that are dealt with in a manner fundamentally different from values.
So NULLs aren't fundamentally necessary, and they tend to obscure information.
They're often initially easier than doing the analysis of the specific need, but then so is not using the relational model in the first place and just storing data in a haphazard manner.
Granted, SQL's concept of NULL is problematic, but you are... let's say *seriously misguided*... if you think a programmer can work with a SQL based relational database (which is the only kind of RDBMS there is) without understanding SQL's concept of NULL and how it affects things like existential predicates or aggregate functions.
First, SQL isn't the only "relational" query language in use, though it is by far the most common.
Second, I never said that understanding NULLs wasn't important to using SQL. I said that understanding the meaning of NULL in the relational model isn't important because NULL isn't part of the relational model.
In any case, I'd like to see a serious argument that NULL violates the fundamental underpinnings of the relational model.
I think C.J. Date & Hugh Darwen have handled this admirably; the first 13 slides of this presentation are as good of an overview of the issues as I've seen (slides 4 & 13 are the most directly relevant to the problem of NULL with regard to the underpinnings of the model.)
I went through a patch a few years ago where I was interviewing programming candidates who had XML coming out of their ears but hadn't the foggiest idea of what "NULL" means in the relational model.
Why should they? NULL doesn't mean anything in the relational model; NULL is an SQL construct that violates the fundamental underpinnings of the relational model.
Seriously Google. This sounds like a.1, or even a.0.1 release.
Scheduled releases with feature upgrades are major version numbers in the Chrome versioning scheme. This is such a release. Consequently its a major version bump.
Google scheme seems to me to be less arbitrary that what their competitors use, where a feature release may bump the major version by 1 or the minor version by 1 or more.
Chrome's versioning scheme seems to have always been that major version numbers are general feature releases, and almost everything else is bug-fix (third-number releases). Their versioning is pretty rational, the only thing is that the second number seems pretty superfluous, since they don't ever seem to have any releases that qualify for whatever standard they have for that (I can't remember every seeing a Chrome version that wasn't x.0.y.z [z being the build number))
Also, they don't actually mean you can't do aborts. A transaction can abort. It is just that all replica's must do the same thing with the transaction - commit or abort. Given you can achieve that, the rest of the paper follows easily. The magical bit is the pre-processor that guarantees all nodes (which might have different data) guarantee that without being a bottle neck or a central point of failure. Now that is magical, and I don't know how they do it.
I don't know how they do it, but the fairly well-established mechanism of doing that without a central point of failure is use of consensus algorithms (e.g., the Paxos algorithm and its various descendants.)
For this reason I suggest that app language designers work on better fitting RDBMS and SQL rather than the other way around
I'd suggest instead that both app and query language designers should work on better fitting generalized handling of data using the relational model; SQL has a lot of limitations that have nothing to do with the essentials of the relational model, and much of the evolution of SQL has been toward shoehorning non-relational features into it rather than addressing the limitations that prevent it from solving some problems using the tools naturally available within the set theoretic model notionally underlying "relational" databases.
Certainly, the relational model has a lot to recommend it, and its good that many application languages (enabled largely by the trend to move features from functional languages into popular OO languages) are developing convenient and elegant means of expressing some relational operations within the application language without resort to external languages. But there is no particular reason that conforming closely to the quirks of SQL should be the goal.
Given all else being equal, supporting ACID is going to require more hardware than not supporting ACID
If all else could be equal, that might be true.
A lot of the motivation for NoSQL systems is that most of them provide strong availability guarantees, at the expense of strict consistency. In a distributed system implemented over an unreliable network, that's an unavoidable tradeoff -- you can have availability or consistency, but not both. That's the upshot of the CAP theorem.
A bigger issue may be the cost of ACID even if it can in theory scale. Supporting ACID is not free. A free web service may be able to afford losing say 1 out of 10,000 web transactions. Banks cannot do it, but Google Experiments can. The extra expense of big-iron ACID
Its not clear to me that guaranteeing availability with only eventual consistency (the usual non-ACID "NoSQL" approach) is any cheaper than guaranteeing consistency in a distributed system and sacrificing availability, and its even less clear that it is unavoidably more expensive.
The unavoidable cost (in a distributed system implemented in multiple nodes on an unreliable network) of consistency is availability, and vice versa. That's the essential meaning of the CAP Theorem.
You said it yourself, lawyers want jurors who will be amenable to their arguments -- often that means both sides have in interest in removing intelligent, critically thinking jurors in favor of those who can be more easily persuaded by misdirection and logical fallacies.
No, what I said doesn't mean that at all, unless you assume that each attorney thinks that their argument will be less convincing to intelligent, critically-thinking jurors than their opponent's argument.
While this combination of perceptions is possible, its only possible if at least one attorney underestimates the soundness of their own arguments or overestimates the soundness of the arguments likely to be used against them, which is exactly the opposite of what both are likely to do.
That's fine until someone asks you an unstructured question for which a two-dimensional array cannot contain the answer.
Like, for example, 'Here's an ordered DOM tree of nodes each containing tags, subtrees and/or chunks of CDATA'.
That's neither unstructured nor a question, and, since it is not a question it has no answer. But if you can certainly represent an ordered tree of nodes, each of which may either be a tag, or a subtree, or a chunk of text in a relational schema. And answer pretty much any answerable question about it using standard SQL.
Or 'Here is a set of objects each of which contain their own custom properties not found in others.'
Again, not a question. The relational model can handle that kind of data, though, and SQL can be used to answer most reasonable questions about it, though to be sure its not what most RDBMS's (or the SQL language) are optimized for. But that's more a problem with the particular products and SQL as a language than with the relational model.
Most of these companies are undoubtedly committing these violations unknowingly
When the patent was awarded, it had an expiration date, and the company either knew or reasonably should have known that their legal privilege of citing the patent expired with the patent. The fact that they chose not to make adequate plans to assure that they would not continue to claim the patent on the product once it expires is the companies fault.
Sure, by the time the patent expires, no one at the company may be aware of the problem, but if that's the case its only because the company, sometime before, didn't bother doing what was necessary. The way you get people not to cut corners like that is to assure that there is a cost to doing so.
Don't why the gov't has to split the money with random third parties, though, that's just asking for abuse.
It doesn't have to. It can sue the violators itself, and keep all the money. The provisions allowing third parties to sue and split the money with the government are there to promote efficiency in government by giving private parties an incentive to find and address violations without the government spending money (except in the portion of the fines it lets the private party keep.)
She thought he was guilty. Big deal. Isn't that what jurors are supposed to do?
Juries aren't supposed to make that determination until the presentation of evidence is complete, aren't supposed to express an opinion on it to other jurors until deliberations begin, and aren't supposed to express an opinion on it to anyone else until the verdict is announced. All of these are important -- the importance of the not expressing opinions parts is largely to avoid jurors feeling committed to an idea that they have expressed so as to avoid having to reverse themselves in light of (if before the end of trial) the evidence presented or (if before the end of jury deliberations) the deliberations in the jury room.
So, no, what she did -- which goes beyond thinking he was guilty -- is very much not what jurors are supposed to do.
The one criminal jury I was on had a pretty good cross-section in terms of education and experience. As I recall, only one potential jurors was removed: one who had previously served on a jury that failed to reach a verdict.
They want dumb jurors and it is what we get.
Each side in a jury trial wants jurors that are as well disposed as possible to their side, and has an incentive to remove jurors that they feel will be unreceptive to the arguments that they plan to use. Additionally, either side (and often both sides) have an interest in the jury actually reaching a verdict. Aside from the strong evidence of bias needed for a removal for cause, they have a limited number of peremptory challenges that they may use to accomplish this, which limits their ability to shape juries.
Aside from the potential shared interest in reaching a verdict, these interests are generally directly opposed, so where one party has an interest in removing intelligent potential jurors, the other party would generally have an interest in removing less-intelligent members of the jury pool, though I don't think its really that common where that would be a factor, anyway.
You're living in a dream world if you believe they would have been reimbursed attorney's fees. The system is of the lawyers, by the lawyers, for the lawyers. They lawyers get paid.
Reimbursement for attorney's fees to the defendant does not prevent the lawyer from getting paid. In fact, the availability of such reimbursement increases, all other things being equal, the amoun the lawyer can charge (and get paid) for services in such a case, and therefore supports the idea of "the lawyers get paid", so even if your cynical view was completely accurate, it would not be a basis for arguing against the point at issue.
Their best possibility is that the plaintiff lawyer amends the complaint to be against the correct website. But if he didn't neither the plaintiff, nor her lawyer would have to pay defendants lawyer just because they sued the wrong person.
As I recall, a fairly common standard for the award of attorney's fees against a plaintiff is approximately that the complaint filed was such that, with reasonable diligence, the plaintiff would have known the essential facts alleged necessary to support the defendant's liability were unsupported by the evidence. That would seem to apply, in spades, in this case.
This has to be the best evidence that I have seen that our "Justice" system is broken. Did the judge not even bother looking at the evidence?
Probably not, since IIRC the legal standard for a default judgement is generally that the facts alleged in the complaint state a case that would justify the damages sought. The presentation of evidence only happens after the defendant responds to the complaint, and then the evidence that is reviewed is only that evidence relevant to questions of fact that are disputed between the plaintiff and the defendant.
Until and unless the defendant responds, there are no questions of fact in dispute, and no reason to review evidence.
And what were they thinking proceeding with a verdict without the defendant having a voice in the proceedings?
I think they were thinking that the defendant surrendered their right to contest the complaint by failing to respond to it as required by law.
I would agree with that, which is why I said *SQL basics*. You won't change the basic reduction down to working with sets no matter what you do with the language or abstractions.
On that level, sure. But a lot of things that one might consider fairly basic to SQL aren't essential to leverage the generality of the relational model.
For instance, you could have a datastore where you assert tuples without reference to a particular relvar (or table), and then the datastore assures that the asserted tuple satisfies the constraints for all defined base relvars whose headers it satisfies. The body of a relvar is simply, then, the set of all asserted tuples that satisfy the header of the relvar.
If you do this, you'll want to have an independent namespace scheme for attribute (column) names rather than using relvars/tables as column namespaces as well.
This produces a database that is going to be very different from an SQL-based database on a fairly fundamental level, but still can leverage mathematical set theory and relational algebra in much the same way.
It also could conceivably support asserting (and querying) facts/tuples before they fit into defined relvars/tables, while still assuring that any defined integrity constraints on defined relvars were satisfied.
Slashdot Mirror
To the extent that that's true, a signed "non-compete" won't help. Non-compete agreements -- except those executed by owners at the time they sell their stake in a business to a new owner -- are prohibited and declared void by California state law (see CA Bus & Prof Code, Sec. 16600-16602.5)
Not in the short term. In the long-term, as income distribution levelled, it might tend back toward the pre-hike level or below, but that would also correspond to less poverty and demand for poverty support programs, so you'd see lower spending to acheive the same policy goals as well as seeing lower revenues.
Yes, because the degree to which the wealthiest 1% out-earn the rest of the population has gone up since then more than the tax rates they pay have gone down. Its not surprising that that has occurred, since lower tax rates on the richest have a compounded effect over time on their wealth.
It future-proofs the API a bit (this may be overengineering, but its a common enough evolution that assignment would develop side effects in the future that it may be worth considering.)
Also, while this isn't what the GP was doing (since it also had a getter) its the only way to create a write-only attribute, which is sometimes useful (since if you expose the variable, its a read/write attribute.)
I've seen studies which claimed that there was, IIRC, approximately an order of magnitude increase in cost to repair design defects in software identified once code has been built around it rather than in design, another order of magnitude for design defects identified in acceptance testing, and another order of magnitude for design defects caught once a system is in production.
So I'd like to see something more than an assertion that this saying does not apply to software.
Actually, it encourages representation of nutcases. Specifically, it naturally tends to a two-party system, and encourages both parties to seek an unshakeable irrational ideological core and to seek to use negative tactics to reduce the voting probability of the non-core parts of the electorate that might lean toward the other party; in a system that tends to more than two viable parties on a long-term basis, negative techniques don't work as well (and negative techniques that disaffect voters on both sides, but your main opponents slightly more than your own are counterproductive rather than beneficial as they are in a two-party system).
No, it makes people more unhappy with government because it makes more damaging ideas become law, and puts more people who are damaged by those ideas in a position of having no effective recourse than would be the case in a system with more effective representation.
No, what keeps real immigration reform from happening is:
1. Many of the people who claim to want some kind of real immigration reform (esp. politicians) actually don't, they just like to say they do. Having a broken system that everyone hates to keep blaming things on and using as an excuse to sell policies that aren't actually aimed at fixing the things that cause the problems that people experience, but that advance other interests, is politically useful.
2. Of the people that do want "real immigration reform", they don't actually share the same values, or perceive the same problems, and the "real immigration reform" that any one subgroup wants isn't just different policy than other subgroups want, its directed at diametrically opposed goals.
There is a subgroup which is fundamentally anti-immigrant, and sees the fundamental problem as the level of immigration, and therefore isn't interested in reforming the system of legal immigration, they are interested only in trying to find a magic bullet that will make the existing legal limits more practical to enforce.
There are a number of overlapping subgroups that are fundametally motivated by racial, ethnic, cultural, or religious hatred and see the fundamental problem as being too much of the "wrong kind" of people immigrating. Tactically these groups often align with each other and the preceding group, but they are distinct.
There's a group that views the basic principles underlying the family-relationship-based categories at the core of the main part of our legal immigration system as fundamentally sound, but finds that various elements of our current immigration policy undercut the purpose of that policy (which is in the recognition of the strength of family bonds and leveraging of them to anchor American identity in immigrants.)
There's a group that sees our current immigration policy as problematic because of the cost of enforcement compared to the perceived benefit, whether or not they share any of the other groups preferences on the overall principles.
There's a group that sees the main problem in our immigration policy is that it doesn't do enough to serve business owner's interest in cheap labor.
All of these groups want "real immigration reform" of one kind or another, and all of them are fertile targets for politicians who want to blame problems on the lack of such reform. But once you get into the specifics of a proposed reform, or even the goals it should seek, almost anything you propose is going to be seen as at best irrelevant and at worst counterproductive by more people than will support it, unless you are clever and vague and manage to let everyone read their own preferences into your vague pronouncements.
If you reread GP, you will note that I never anywhere in it advocated increasing the overall level of legal immigration, I advocated maintaining the current overall limits but balancing the country-by-country limits so that the number of immigrants allowed in each limited category by country would be aligned with the number of applicants from each country, so that you don't cause really long delays in the countries that have the most otherwise-qualified prospective immigrants. A (for instance) 2 year wait everywhere would create much less pressure to evade the system than having negligible waits from most of the world, but decade plus waiting lists from a handful of countries
I think its more likely that people are unhappy with our government, in proximate terms, because the economy sucks and the government at the time, rightly or wrongly, tends to disproportionately get the blame or credit for the economy. In more general, long term, terms, I think people are unhappy in the US with government because we have an electoral system which greatly limits the scope of ideas represented in the government and produces a government that does not effectively represent the preferences of the populace, which is borne out by the research on popular opinion of government across stable democratic states and the relation it bears to the electoral systems and the number of competitive parties.
This somewhat critically misstates the problem.
It is not that the government is "backlogged" on immigration requests, which suggests that there is a pile of paperwork that just isn't getting processed because of insufficient resources, and if the government would just process the requests it has and execute the existing policy efficiently, there wouldn't be a problem.
The issue is that the US has an immigration system that involves per country limits on immigration (as I recall, part of that is a limit that no one country may have more than 5% of the people admitted in each year in any of the numerically-limited, relationship-based immigration categories -- note that H1Bs aren't part of this) which results in a surplus of requests from prospective immigrants who are legally qualified to immigrate resulting in long (more than a decade in some cases) wait for admission in certain categories from certain countries (this is true, IIRC, in all of the numerically-limited relationship-based categories for Mexico, as well as certain categories from India and some other countries.)
You could reduce, though not eliminate, the problem of new illegal immigration overnight without changing general immigration policy or overall legal immigration limits, simply by using overall limits without limited per-country allocation.
That depends when you asked him, I suppose.
And by the 1990, he was arguing that a relational system needed to have at least two different kinds of Nulls to operate correctly.
(Apparently, the beginning of the start of something in the db world which culminated with the proposal of requiring support for up to 128 types of distinguished NULLs in the early part of the standardization process of SQL-99.)
So, outside of the 1980s, Codd would seem to view SQL's use of NULL as a problem. Not that I'm find of arguments to authority.
Practical systems need to be able to admit of records that a short of ideal because some data is either missing, unknown, or otherwise different from the basic assumptions.
Each of those possibilities -- missing, unknown, and each other possibility of difference from the basic assumptions -- can be identified with an actual value without resort to special non-values that are dealt with in a manner fundamentally different from values.
So NULLs aren't fundamentally necessary, and they tend to obscure information.
They're often initially easier than doing the analysis of the specific need, but then so is not using the relational model in the first place and just storing data in a haphazard manner.
First, SQL isn't the only "relational" query language in use, though it is by far the most common.
Second, I never said that understanding NULLs wasn't important to using SQL. I said that understanding the meaning of NULL in the relational model isn't important because NULL isn't part of the relational model.
I think C.J. Date & Hugh Darwen have handled this admirably; the first 13 slides of this presentation are as good of an overview of the issues as I've seen (slides 4 & 13 are the most directly relevant to the problem of NULL with regard to the underpinnings of the model.)
Why should they? NULL doesn't mean anything in the relational model; NULL is an SQL construct that violates the fundamental underpinnings of the relational model.
SQL, as such, is declarative. Many RDBMSs include, in addition to SQL, an SQL-derived procedural scripting language (Oracle's PL/SQL, and so on.)
Scheduled releases with feature upgrades are major version numbers in the Chrome versioning scheme. This is such a release. Consequently its a major version bump.
Google scheme seems to me to be less arbitrary that what their competitors use, where a feature release may bump the major version by 1 or the minor version by 1 or more.
Chrome's versioning scheme seems to have always been that major version numbers are general feature releases, and almost everything else is bug-fix (third-number releases). Their versioning is pretty rational, the only thing is that the second number seems pretty superfluous, since they don't ever seem to have any releases that qualify for whatever standard they have for that (I can't remember every seeing a Chrome version that wasn't x.0.y.z [z being the build number))
I don't know how they do it, but the fairly well-established mechanism of doing that without a central point of failure is use of consensus algorithms (e.g., the Paxos algorithm and its various descendants.)
I'd suggest instead that both app and query language designers should work on better fitting generalized handling of data using the relational model; SQL has a lot of limitations that have nothing to do with the essentials of the relational model, and much of the evolution of SQL has been toward shoehorning non-relational features into it rather than addressing the limitations that prevent it from solving some problems using the tools naturally available within the set theoretic model notionally underlying "relational" databases.
Certainly, the relational model has a lot to recommend it, and its good that many application languages (enabled largely by the trend to move features from functional languages into popular OO languages) are developing convenient and elegant means of expressing some relational operations within the application language without resort to external languages. But there is no particular reason that conforming closely to the quirks of SQL should be the goal.
If all else could be equal, that might be true.
A lot of the motivation for NoSQL systems is that most of them provide strong availability guarantees, at the expense of strict consistency. In a distributed system implemented over an unreliable network, that's an unavoidable tradeoff -- you can have availability or consistency, but not both. That's the upshot of the CAP theorem.
Its not clear to me that guaranteeing availability with only eventual consistency (the usual non-ACID "NoSQL" approach) is any cheaper than guaranteeing consistency in a distributed system and sacrificing availability, and its even less clear that it is unavoidably more expensive.
The unavoidable cost (in a distributed system implemented in multiple nodes on an unreliable network) of consistency is availability, and vice versa. That's the essential meaning of the CAP Theorem.
No, what I said doesn't mean that at all, unless you assume that each attorney thinks that their argument will be less convincing to intelligent, critically-thinking jurors than their opponent's argument.
While this combination of perceptions is possible, its only possible if at least one attorney underestimates the soundness of their own arguments or overestimates the soundness of the arguments likely to be used against them, which is exactly the opposite of what both are likely to do.
That's neither unstructured nor a question, and, since it is not a question it has no answer. But if you can certainly represent an ordered tree of nodes, each of which may either be a tag, or a subtree, or a chunk of text in a relational schema. And answer pretty much any answerable question about it using standard SQL.
Again, not a question. The relational model can handle that kind of data, though, and SQL can be used to answer most reasonable questions about it, though to be sure its not what most RDBMS's (or the SQL language) are optimized for. But that's more a problem with the particular products and SQL as a language than with the relational model.
When the patent was awarded, it had an expiration date, and the company either knew or reasonably should have known that their legal privilege of citing the patent expired with the patent. The fact that they chose not to make adequate plans to assure that they would not continue to claim the patent on the product once it expires is the companies fault.
Sure, by the time the patent expires, no one at the company may be aware of the problem, but if that's the case its only because the company, sometime before, didn't bother doing what was necessary. The way you get people not to cut corners like that is to assure that there is a cost to doing so.
It doesn't have to. It can sue the violators itself, and keep all the money. The provisions allowing third parties to sue and split the money with the government are there to promote efficiency in government by giving private parties an incentive to find and address violations without the government spending money (except in the portion of the fines it lets the private party keep.)
How is anything here "abuse"?
Juries aren't supposed to make that determination until the presentation of evidence is complete, aren't supposed to express an opinion on it to other jurors until deliberations begin, and aren't supposed to express an opinion on it to anyone else until the verdict is announced. All of these are important -- the importance of the not expressing opinions parts is largely to avoid jurors feeling committed to an idea that they have expressed so as to avoid having to reverse themselves in light of (if before the end of trial) the evidence presented or (if before the end of jury deliberations) the deliberations in the jury room.
So, no, what she did -- which goes beyond thinking he was guilty -- is very much not what jurors are supposed to do.
The one criminal jury I was on had a pretty good cross-section in terms of education and experience. As I recall, only one potential jurors was removed: one who had previously served on a jury that failed to reach a verdict.
Each side in a jury trial wants jurors that are as well disposed as possible to their side, and has an incentive to remove jurors that they feel will be unreceptive to the arguments that they plan to use. Additionally, either side (and often both sides) have an interest in the jury actually reaching a verdict. Aside from the strong evidence of bias needed for a removal for cause, they have a limited number of peremptory challenges that they may use to accomplish this, which limits their ability to shape juries.
Aside from the potential shared interest in reaching a verdict, these interests are generally directly opposed, so where one party has an interest in removing intelligent potential jurors, the other party would generally have an interest in removing less-intelligent members of the jury pool, though I don't think its really that common where that would be a factor, anyway.
Reimbursement for attorney's fees to the defendant does not prevent the lawyer from getting paid. In fact, the availability of such reimbursement increases, all other things being equal, the amoun the lawyer can charge (and get paid) for services in such a case, and therefore supports the idea of "the lawyers get paid", so even if your cynical view was completely accurate, it would not be a basis for arguing against the point at issue.
As I recall, a fairly common standard for the award of attorney's fees against a plaintiff is approximately that the complaint filed was such that, with reasonable diligence, the plaintiff would have known the essential facts alleged necessary to support the defendant's liability were unsupported by the evidence. That would seem to apply, in spades, in this case.
Probably not, since IIRC the legal standard for a default judgement is generally that the facts alleged in the complaint state a case that would justify the damages sought. The presentation of evidence only happens after the defendant responds to the complaint, and then the evidence that is reviewed is only that evidence relevant to questions of fact that are disputed between the plaintiff and the defendant.
Until and unless the defendant responds, there are no questions of fact in dispute, and no reason to review evidence.
I think they were thinking that the defendant surrendered their right to contest the complaint by failing to respond to it as required by law.
On that level, sure. But a lot of things that one might consider fairly basic to SQL aren't essential to leverage the generality of the relational model.
For instance, you could have a datastore where you assert tuples without reference to a particular relvar (or table), and then the datastore assures that the asserted tuple satisfies the constraints for all defined base relvars whose headers it satisfies. The body of a relvar is simply, then, the set of all asserted tuples that satisfy the header of the relvar.
If you do this, you'll want to have an independent namespace scheme for attribute (column) names rather than using relvars/tables as column namespaces as well.
This produces a database that is going to be very different from an SQL-based database on a fairly fundamental level, but still can leverage mathematical set theory and relational algebra in much the same way.
It also could conceivably support asserting (and querying) facts/tuples before they fit into defined relvars/tables, while still assuring that any defined integrity constraints on defined relvars were satisfied.