You keep refering to "your solution" as if I'd posed one. All I said was that you need to have a proliferation of TLDs to mix things up, but that there's no right solution
Short domains aren't the problem either. I can get a REALLY short domain if I want (well, I already have one, but I could easily get another). The problem is that marketing droids have this illogical fear of someone buying up foo.com when they hold foo.pl even though I don't know of anyone who has ever been hurt by such an arangement.
[The manufacturer defended themselves...] Of course the manufacturer said that.
Ah, you must be from outside the United States. It's easy to misunderstand US law with respect to burden of proof, since TV and movies that we export so often make it seem as if the burden is on the accused. It's not.
This should have happened this way:
* Complained filed with H.S. * H.S. refers it to either FBI or the state's A.G. * FBI or A.G. issues a warning or: * FBI or A.G. takes retailer to court or: * FBI or A.G. refers the matter to the owner of the trademark as a civil matter.
If you are not in one of these states, vote third party
That's fine advice as long as you happen to know how many people are going to do what you suggest. Keep in mind that a "landslide" election is less than 10 points. If a candidate is trailing in a state by, say, 5 points he considers it "lost", and moves on to where he can affect the vote.... if those 5 percent then decide to vote third party you have a bit of a problem.
The correct solution is to change the system. Lobby your state to implement IRV. As more states do this, the Federal government will come under more and more pressure to write an ammendment and submit it for state ratification that changes the way we vote. Replacing the electoral college with a straight tally of state voting records, and doing the tally by IRV makes a lot of sense once you explain the advantages to people.
Is IRV perfect? Not at all, but it's a great step in the right direction, and actually allows candidates to work issues instead of demographics.
In replying to you, I started falling into all of the traps I was trying to describe.... sigh... let me start over:
There's no problem with.us until you start to need sales.eu.ibm.ny.us which makes absolutely no sense.
Geographical location is a terrible way to locate things on the Net, and political location is only a hair better in most circumstances, and MUCH worse in some (if I'm in Kashmir, do I get an Indian or Pakistani domain? Depends on who I want to have hate me, I guess)!
There's no right solution. The only way you can approach correct is to have as many TLDs as there are ways of managing the process of giving domains out.
Don't believe me? Well, here's another way to look at it. Each domain is a hash key. That's really all it is. Go look up the research that's been done on hashing over the years and you'll find that when you're trying to hash complex data, the best way to do it is to have multiple sets of indicies (you'll recognize this from SQL databases). All we're doing here is indexing based on different sets of keys. You can index by (country, state, entity_name) or by (profit_status, entity_name) or by any combinaiton of identifiers you want. There's even meta-identifiers that are managed by search engines (Google's "ad words" for example), but they don't work so well for more programatic access like email.
Don't make the mistake of limiting yourself to one indexing scheme... anyone who has worked on a large, complex database can tell you that that ends up convoluting your data beyond all recognition (as you attempt to optimize data to match access modes instead of the other way around).
What is not mentioned is that the forking problem exists in closed source as well.
Re-read the summary line that I posted in the grandparent. They DO mention that much of the same issues exist in proprietary software. There are unique issues with regards to FOSS, but this is not one of them, and since the FDIC has been doing risk management studies of proprietary software for longer than I've been in this business, I'd be shocked if they don't realize that.
SenderID isn't at issue here, so please drop it. I was discussing SPF. SenderID is an SPF/CallerID hybrid which was DOA.
It relies on DNS, which is a fundamentally broken authentication mechanism
Works for me and tens of thousands of others.
Lots of sites never got their RDNS entries right.
That's fine. They'll learn as more an more of their mail gets rejected. Sites that have no reverse mapping at all are already being rejected in droves anyway. Oh, and SPF doesn't require that your RDNS mapping is correct (though most implementations do check that it exists). It only requires that the domain that you claim to be coming from the envelope is one that lists your host as an authorized sender.
DNS is unreliable.
DNS is unreliable?! Where do you get this kind of thing? DNS Is one of the most reliable distributed databases (arguably THE MOST reliable distributed database) in the world. Are you refering to the fact that DNS is done over UDP? If so, you should know that the redundancy in the domain resolution protocol provides at least as many guarantees as TCP. That is, your answer will be the same regardless of duplication of packets, and packet loss will be detected and dealt with appropriately. DNS is quite reliable, though its guarantees do not include your question ever being answered, neither do the guarantees of ANY IP-based protocol. Obviously, they cannot.
DNS resolution is usually not parallelizable.
I do it all the time.
denial of service attack[...]By overloading DNS servers of small companies.
That's highly unlikely. There is a 1:1 ratio of outgoing requests via SMTP to SPF requests generated. Yes, you could use a network of zombies to issue those requests, but that network of zombies could just as easily send DNS requests directly to the small companies servers, causing the same problem. Dangerous denial of service conditions stem from many:1 relationships, and SPF has none that I'm aware of.
[...]By using bogus domain names that cause 30 second stalls in your inbound traffic.
I would far, far rather have incoming SMTP delayed 30 seconds waiting for a response (though I would probably trim that time anyway) then start recieving the message. I ALREADY do this, in fact, for reverse DNS and Spamhaus SBL/XBL checking, and it's not a problem.
That's a GREAT article, and I think someone should submit a story to Slashdot with just that story.
Real-world risk management is absolutely key with any OS or software suite, and that essay from the FDIC is spot-on. I might disagree with a few points, but overall it's quite accurate.
I also enjoy their summary which starts:
"The use of FOSS by financial institutions does not pose risks that are fundamentally different from those presented by the use of proprietary or self-developed software.However, FOSS adoption and usage necessitates some distinctive risk management practices with which institutions must be familiar."
Yep, that's exaclty what Ballmer is trying to convince you isn't so... so who do we trust on risk assessment, federal bank insurers or Microsoft? Heh.
There's nothing wrong with SMTP... The problem lies with the lack of consensus on authentication, authorization and reputation systems for electronic mail.
For example, using a combination of SPF and SMTP/AUTH you can easily prevent anyone who uses SPF from accepting invalid mail "from" your domain(s) while continuing to use the world's most pervasive mail transfer protocol.
Problem is that people aren't willing to apply the time and effort required to do this globally.
The next step is reputation, and as soon as you can be sure that the person claiming to be joe@example.com is in fact from example.com, you can begin assigning example.com a reputation. You'll see dozens of distributed reputation databases, just like IP-based blacklists, overnight.
Want to move the process along? Add an SPF record for your domain and add an SPF milter (or equivalent for your MTA technology) to your mail server. The sooner forgeries stop, the sooner we can start building reputation and end this.
Why would anyone ever trust a developer release? Seriously, I download something from the developers' site once in a blue moon when I'm working on the code. Any other time, I wait for a system integrator to worry about all of the issues that I don't have time for (does it play nice with the other 750 packages I have installed? is it a substantial change that is going to break compatibility with other systems? Did important bug-fixes get picked up or do they need to be re-integrated with this version? Are there any new security issues? etc...)
Why on earth anyone would want to take on all of that work just to get some features a few months early I can't imagine. I have better things to do with my time.
Your inability to see a difference between Darth Vader and Hitler or a slave trader is particularly offensive, I might add.
Vader is a great example, and I don't see any difference at all between him and, say, Hitler other that two things:
1. Vader was fictional 2. Vader killed orders of magnitude more people
Since we're already talking about fictional circumstances, not real-world sexual discrimination we can toss the first one. That simply makes Vader a BETTER comparison, and renders complaints about the on-line game weaker.
The second requires a moment of thought... we embrace Vader as a "bad ass" villain. People play the Imperial March at social functions. "Luke, I am your father," has become a catch-phrase used by young and old alike. And yet, we're talking about a character who is best known for killing billions of people at a time and strangling those who reported to him.
In that light, I ask you if it's so out of line that a fictional character in an online game would be sexist.... personally I can't see how one would expect NOT to run socially unacceptable if nor outright immoral characters in these games.
If you're likely to be offended by a character that is sexist or racist or mean or violent, then I cannot imagine a multi-player game I would recommend to you other than hearts... and even then...
The Japanese name, "go-ji-ra", is being re-translated into english for future films as "Zilla" instead of "Godzilla" since, for some reason the target audience has been pegged as being the sort to be offended by this type of thing. Personally I can't see how those two groups would ever cross paths, but whatever.
I'm also at a bit of a loss for why we can't simply call it "Gojila"
I won't justify the parent flamer's choice of language by quoting directly, but I will take issue with the point that views are not sub-selects. Views most certainly are sub-selects, and as evidence for that look at the fact that you can either implement sub-selects using temporary views or (more naturally) that you can implement views using sub-selects trivially. Logically, a stored query that acts as a table *is* a sub-select.
That does not mean that an implementation is tied to directly translating any use of a view to the equivalent sub-select query. Far from it: there are many optimizations to be had in terms of caching and applying indicies to the view itself. But again, even optimizations of the implementation don't change the basic nature of the construct, which is a stored sub-select.
As far as sub-selects being implemented as temporary tables... they most certainly are, though there are relational optimizations to be had and the user-concept of a "temporary table" might not map to the internal implementation at all. Again, optimization does not change the fundamental nature of the beast. Please feel free to respond using something other than curses and vague claims. I'll respond in kind.
"Oh, no. I wasn't talking about Google, I was talking about Slashdot. You didn't really read my post, did you?"
Yeah, I did. That's not what you said. It might well be what you meant, but I'm not a mind-reader, and when I ask a question, I generally do at least browse the context for my question first. Thanks for asking, though.
"Thanks for validating the meaning contained in the sig."
I'm not sure how you got that out of what I said, but my basic premise stands: check your logic against your desire to draw parallels, and you'll find that people listen to what you have to say.
If you continue to (as you did in your post and your sig) attempt to draw equality between inequal sets, then... well, why bother?
For example, if you sig read:
"Microsloth"? That's hilarious. "Open sores" is funny, too. Oh, it isn't?
then I and many others would agree. You instead specifically chose a term that isn't funny at all, but a rather dry irony ("M$"). Why? It hurts your entire point.
Same deal with your post. You draw comparisons, but fail to enlighten your audience as to who the comparisons are drawn between and why.
Ok, nuff said. If you still think I'm just sounding off, then ignore me and I'm sure we can go our seperate ways without having to interact further.
I realize that they're talking about the lawsuit with AT&T (I still wear my "Free The BSD 4.4" shirt with pride, though I gave away my "4.4 > 5.4" poster). What I didn't understand was the implication that that suit began anything in particular. It was a controversy raised long after the code had been used to create 386BSD and at least a decade after BSD first became popular, so I didn't see it as a "beginning".
Yeah, sub-selects are mostly just temporary tables.
That is to say, if you:
create temporary table foo select f.id from fiz f, fam m where f.name = m.name; select b.stuff from bar b, foo f where b.id = f.id;
Then that's the same thing as:
select b.stuff from bar b where b.id in (select f.id from fiz f, fam m where f.name = m.name)
The only difference being that for sub-selects, since it's done server-side, you can optimize by evlauating the sub-select lazily. Some databases that support sub-selects do not do this optimization, but most do and in some cases it's a big win. It's just a rare enough case that it's not that big a deal for the general discussion of the value of sub-selects vs other features that people want.
MySQL's SQL is quite standard, and while there are differences, they are either the lack of a small number of features (like sub-selects, which you go into below) or very baroque trade-offs, which most users will not encounter.
I suspect you're instead running into assumptions about what is and is not "standard" based on what you've used in the past. MySQL follows the ANSI SQL standard as closely as all of the other databases I've used, having its own small quirks and LOTS of extensions, but basically doing as good a job of standards compliance as, say, gcc does with the ANSI C standard.
I understand your frustration, though. I'm an old Sybase user, and I miss the Sybase extensions in many places.
No subselects [...] That is huge
It is?! Why? All a sub-select does is instantiate a temporary table on the fly. You can, of course, instantiate a temporary table on the fly, so what is a sub-select buying you?
The only possible advantage would be the possibility of lazily evaluating a very large subselect, but I've come across that need in perhaps 2 or 3 situations in the course of 12 years of working with databases in production environments... how is this "huge"?
In addition, the version that I used didn't support views.
You're showing your lack of depth here. By definition a system that lacks sub-selects lacks views (as views are just named sub-selects).
All that said, sub-selects are being added to MySQL, but people who pay for work on the core server don't ask for these features as much as certain others, and that's how the folks at MySQL decide what to put effort into.
I've written my thoughts on Stephenson's endings here before, but let me re-state: he's a very bright man, and like most of his ilk, is obviously very easily distracted. This leads to the sorts of endings that make it feel like he's left the room.
That said, you get more out of the first 90% of a Stephenson book than you do with almost all modern fiction (there are exceptions, and they're ALL worth reading, and many suffer from the same problem). Personally, I find his insights on topics ranging from nanotechnology to pipe organs useful enough to warrant suffering his endings. I even hear that the Baroque Cycle marks his first set of good endings, and I look forward to getting my copy of Quicksilver back to find out;-)
YMMV, but I find that true insight into maters of modern technology and society are rare. This is why I grasp onto authors like Vernor Vinge and Neil Stephenson. They're the bards of our age, and we should listen and learn what we have taught them.
And about this word... I think describing problems with e-voting as "SNAFU" is the most accurate assessment I've heard to date. Bravo to these people for coming clean!
I disagree. Regardless of the particular institution, I think we're all aware of the fact that providing sensitive information to a univerisity without tight controls is amazingly dangerous. I don't blame Berkeley here, nor would I blame any school. I do think they should take this as a warning, and improve their security, but the state has to ask themselves what they were thinking, and restrict future access to this kind of information.
The process for releasing sensitive data should be:
1. Establish the exact scope of the release (who gets it) 2. Perform a reasonable amount of screening of those people based on the sensitivity of the information (ranging from simple ID checks all the way up to full security clearance procedures depending on how much risk you are exposing yourself to). 3. Perform the above two steps for any computers that will have access to the data. This means either performing a security audit, or requiring a third-party ceritification. Again, you tier the requirements based on the risk.
The state did not do that. The state got burned. If the state had handed this info to some random guy on the street and that guy had turned around and sold it to the highest bidder, I wouldn't be blaming the guy.
I don't find the character motion to be choppy, but in the event of bit-droppage over the wire, you can always do another "take". As for mouths... yeah, that would be the biggest problem, but how does the Sims version deal with that?
You could always edit in the mouth motion or just use characters in gear that covers the mouth (e.g. troll face-guards or full-warrior helms). I don't think it's insurmountable at all, and I've seen a few videos of raids that have come out so nicely that I've really wanted to see some fiction (e.g. scripted interaction vs. role playing) done with it.
I was using BSD 4.2 in college over 15 years ago, and the litigation didn't happen for quite some time.
BSD's roots are in the early 80s when they were working closely with Bell Labs, and both versions of Unix were quasi-official.
Obviously, the big break for the modern BSD was 386BSD, which brought the OS to the personal computer a little over a decade ago.
Today, I think it is the rich set of userland capabilities that distinguish the BSDs to the point that occasionally Linux distributions pop up that emulate their functionality (e.g. Gentoo's use of a BSD-like ports system).
BSD is a rich OS with a long history, and I'm glad that it's still around and growing into niches that need it. Today, I'm mostly a Linux user, but I remember my roots and the joy that life was when BSD gained popularity over the proprietary OSes of the day back in the 80s.
It seems to me that EverQuest is ripe for this sort of thing. Advantages over Sims or Halo:
Hundreds of very unique locations which, in turn, have dozens of nooks and crannies to shoot in. Some areas will be too difficult to get to yourself for "shooting", but if you've ever played EQ, you know that the places with nobody around are in the majority, even in "popular zones"
Wide varitety of looks (easily numbering in the thousands of easily distinguished combinations) from race to gender to hair/eyes to armor (including user-selected dye) to weapons.
Ability to use a practically unlimited number of "puppeteers" (players) at once
Spell and particle effects are very nice for all sorts of "special effects".
A rich existing mythology if you wish to draw on it
Thousands of available "extras", many of whom would be happy to help stage events for the show
Availablity of scripting for actors (though such "macro programs" are technically a bannable offense in-game).
I would think this would be the place to do such a movie (though any popular MMORPG would have the same advantages).
The down side is having to pay a monthly fee per simultaneous actor.
Just thought of a different way to respond to this.
You say, "I happen to have creative control on several websites, two of which get a substantial amount of traffic. According to your reasoning, it's perfectly OK for me to put a George Kerry bumpersticker on my car, but it should be a federal offense for me to put a John Bush ad on my website?"
That's right. When you say, "I happen to have... several websites," you could just as easily have said, "I happen to have $1,000,000."
However, I'm not too thrilled with the idea of one contributor having that much control over an election. If we allow contributions like that, then we create a system where a small group of big-spenders can "buy" the election.
I'm sure you don't want that any more than I do, but it's hard to see that that's the right choice when YOU are the one being restricted. I understand and empahtize, but that doesn't change my stand on this.
In fact, I'm rather a radical on this point. I'd rather see candidates cut off from the campaign process entirely. I want to see a non-governmental association of interested parties (including some governmental parties like the FEC and states representation) run the promotional aspects of an election (for all federal offices). Produce a biographical film about the candidates. Publish print, Internet and television ads which describe the candidates' stands on major issues. Schedule debates among ALL of the candidates spanning the entire summer leading up to the election, not just the two months before.
If there is only one pool of money on which to draw, and all candidates draw evenly, then we'll be able to hold an election based on the issues and the capabilities of those who wish to lead.
Slashdot Mirror
You keep refering to "your solution" as if I'd posed one. All I said was that you need to have a proliferation of TLDs to mix things up, but that there's no right solution
Short domains aren't the problem either. I can get a REALLY short domain if I want (well, I already have one, but I could easily get another). The problem is that marketing droids have this illogical fear of someone buying up foo.com when they hold foo.pl even though I don't know of anyone who has ever been hurt by such an arangement.
We'll learn to grow up and get over it.
[The manufacturer defended themselves ...] Of course the manufacturer said that.
Ah, you must be from outside the United States. It's easy to misunderstand US law with respect to burden of proof, since TV and movies that we export so often make it seem as if the burden is on the accused. It's not.
This should have happened this way:
* Complained filed with H.S.
* H.S. refers it to either FBI or the state's A.G.
* FBI or A.G. issues a warning or:
* FBI or A.G. takes retailer to court or:
* FBI or A.G. refers the matter to the owner of the trademark as a civil matter.
There's no excuse for the way this played out.
HP sells a SuSE pre-installed laptop. It was announced a few months ago (possibly even showed up here on Slashdot).
If you are not in one of these states, vote third party
That's fine advice as long as you happen to know how many people are going to do what you suggest. Keep in mind that a "landslide" election is less than 10 points. If a candidate is trailing in a state by, say, 5 points he considers it "lost", and moves on to where he can affect the vote.... if those 5 percent then decide to vote third party you have a bit of a problem.
The correct solution is to change the system. Lobby your state to implement IRV. As more states do this, the Federal government will come under more and more pressure to write an ammendment and submit it for state ratification that changes the way we vote. Replacing the electoral college with a straight tally of state voting records, and doing the tally by IRV makes a lot of sense once you explain the advantages to people.
Is IRV perfect? Not at all, but it's a great step in the right direction, and actually allows candidates to work issues instead of demographics.
In replying to you, I started falling into all of the traps I was trying to describe.... sigh... let me start over:
.us until you start to need sales.eu.ibm.ny.us which makes absolutely no sense.
There's no problem with
Geographical location is a terrible way to locate things on the Net, and political location is only a hair better in most circumstances, and MUCH worse in some (if I'm in Kashmir, do I get an Indian or Pakistani domain? Depends on who I want to have hate me, I guess)!
There's no right solution. The only way you can approach correct is to have as many TLDs as there are ways of managing the process of giving domains out.
Don't believe me? Well, here's another way to look at it. Each domain is a hash key. That's really all it is. Go look up the research that's been done on hashing over the years and you'll find that when you're trying to hash complex data, the best way to do it is to have multiple sets of indicies (you'll recognize this from SQL databases). All we're doing here is indexing based on different sets of keys. You can index by (country, state, entity_name) or by (profit_status, entity_name) or by any combinaiton of identifiers you want. There's even meta-identifiers that are managed by search engines (Google's "ad words" for example), but they don't work so well for more programatic access like email.
Don't make the mistake of limiting yourself to one indexing scheme... anyone who has worked on a large, complex database can tell you that that ends up convoluting your data beyond all recognition (as you attempt to optimize data to match access modes instead of the other way around).
What is not mentioned is that the forking problem exists in closed source as well.
Re-read the summary line that I posted in the grandparent. They DO mention that much of the same issues exist in proprietary software. There are unique issues with regards to FOSS, but this is not one of them, and since the FDIC has been doing risk management studies of proprietary software for longer than I've been in this business, I'd be shocked if they don't realize that.
SenderID isn't an acceptable solution
SenderID isn't at issue here, so please drop it. I was discussing SPF. SenderID is an SPF/CallerID hybrid which was DOA.
It relies on DNS, which is a fundamentally broken authentication mechanism
Works for me and tens of thousands of others.
Lots of sites never got their RDNS entries right.
That's fine. They'll learn as more an more of their mail gets rejected. Sites that have no reverse mapping at all are already being rejected in droves anyway. Oh, and SPF doesn't require that your RDNS mapping is correct (though most implementations do check that it exists). It only requires that the domain that you claim to be coming from the envelope is one that lists your host as an authorized sender.
DNS is unreliable.
DNS is unreliable?! Where do you get this kind of thing? DNS Is one of the most reliable distributed databases (arguably THE MOST reliable distributed database) in the world. Are you refering to the fact that DNS is done over UDP? If so, you should know that the redundancy in the domain resolution protocol provides at least as many guarantees as TCP. That is, your answer will be the same regardless of duplication of packets, and packet loss will be detected and dealt with appropriately. DNS is quite reliable, though its guarantees do not include your question ever being answered, neither do the guarantees of ANY IP-based protocol. Obviously, they cannot.
DNS resolution is usually not parallelizable.
I do it all the time.
denial of service attack[...]By overloading DNS servers of small companies.
That's highly unlikely. There is a 1:1 ratio of outgoing requests via SMTP to SPF requests generated. Yes, you could use a network of zombies to issue those requests, but that network of zombies could just as easily send DNS requests directly to the small companies servers, causing the same problem. Dangerous denial of service conditions stem from many:1 relationships, and SPF has none that I'm aware of.
[...]By using bogus domain names that cause 30 second stalls in your inbound traffic.
I would far, far rather have incoming SMTP delayed 30 seconds waiting for a response (though I would probably trim that time anyway) then start recieving the message. I ALREADY do this, in fact, for reverse DNS and Spamhaus SBL/XBL checking, and it's not a problem.
Real-world risk management is absolutely key with any OS or software suite, and that essay from the FDIC is spot-on. I might disagree with a few points, but overall it's quite accurate.
I also enjoy their summary which starts:Yep, that's exaclty what Ballmer is trying to convince you isn't so... so who do we trust on risk assessment, federal bank insurers or Microsoft? Heh.
There's nothing wrong with SMTP... The problem lies with the lack of consensus on authentication, authorization and reputation systems for electronic mail.
For example, using a combination of SPF and SMTP/AUTH you can easily prevent anyone who uses SPF from accepting invalid mail "from" your domain(s) while continuing to use the world's most pervasive mail transfer protocol.
Problem is that people aren't willing to apply the time and effort required to do this globally.
The next step is reputation, and as soon as you can be sure that the person claiming to be joe@example.com is in fact from example.com, you can begin assigning example.com a reputation. You'll see dozens of distributed reputation databases, just like IP-based blacklists, overnight.
Want to move the process along? Add an SPF record for your domain and add an SPF milter (or equivalent for your MTA technology) to your mail server. The sooner forgeries stop, the sooner we can start building reputation and end this.
Why would anyone ever trust a developer release? Seriously, I download something from the developers' site once in a blue moon when I'm working on the code. Any other time, I wait for a system integrator to worry about all of the issues that I don't have time for (does it play nice with the other 750 packages I have installed? is it a substantial change that is going to break compatibility with other systems? Did important bug-fixes get picked up or do they need to be re-integrated with this version? Are there any new security issues? etc...)
Why on earth anyone would want to take on all of that work just to get some features a few months early I can't imagine. I have better things to do with my time.
Your inability to see a difference between Darth Vader and Hitler or a slave trader is particularly offensive, I might add.
Vader is a great example, and I don't see any difference at all between him and, say, Hitler other that two things:
1. Vader was fictional
2. Vader killed orders of magnitude more people
Since we're already talking about fictional circumstances, not real-world sexual discrimination we can toss the first one. That simply makes Vader a BETTER comparison, and renders complaints about the on-line game weaker.
The second requires a moment of thought... we embrace Vader as a "bad ass" villain. People play the Imperial March at social functions. "Luke, I am your father," has become a catch-phrase used by young and old alike. And yet, we're talking about a character who is best known for killing billions of people at a time and strangling those who reported to him.
In that light, I ask you if it's so out of line that a fictional character in an online game would be sexist.... personally I can't see how one would expect NOT to run socially unacceptable if nor outright immoral characters in these games.
If you're likely to be offended by a character that is sexist or racist or mean or violent, then I cannot imagine a multi-player game I would recommend to you other than hearts... and even then...
Interesting. Thanks for the clarification.
The Japanese name, "go-ji-ra", is being re-translated into english for future films as "Zilla" instead of "Godzilla" since, for some reason the target audience has been pegged as being the sort to be offended by this type of thing. Personally I can't see how those two groups would ever cross paths, but whatever.
I'm also at a bit of a loss for why we can't simply call it "Gojila"
I won't justify the parent flamer's choice of language by quoting directly, but I will take issue with the point that views are not sub-selects. Views most certainly are sub-selects, and as evidence for that look at the fact that you can either implement sub-selects using temporary views or (more naturally) that you can implement views using sub-selects trivially. Logically, a stored query that acts as a table *is* a sub-select.
That does not mean that an implementation is tied to directly translating any use of a view to the equivalent sub-select query. Far from it: there are many optimizations to be had in terms of caching and applying indicies to the view itself. But again, even optimizations of the implementation don't change the basic nature of the construct, which is a stored sub-select.
As far as sub-selects being implemented as temporary tables... they most certainly are, though there are relational optimizations to be had and the user-concept of a "temporary table" might not map to the internal implementation at all. Again, optimization does not change the fundamental nature of the beast. Please feel free to respond using something other than curses and vague claims. I'll respond in kind.
Yeah, I did. That's not what you said. It might well be what you meant, but I'm not a mind-reader, and when I ask a question, I generally do at least browse the context for my question first. Thanks for asking, though.
"Thanks for validating the meaning contained in the sig."
I'm not sure how you got that out of what I said, but my basic premise stands: check your logic against your desire to draw parallels, and you'll find that people listen to what you have to say.
If you continue to (as you did in your post and your sig) attempt to draw equality between inequal sets, then... well, why bother?
For example, if you sig read:then I and many others would agree. You instead specifically chose a term that isn't funny at all, but a rather dry irony ("M$"). Why? It hurts your entire point.
Same deal with your post. You draw comparisons, but fail to enlighten your audience as to who the comparisons are drawn between and why.
Ok, nuff said. If you still think I'm just sounding off, then ignore me and I'm sure we can go our seperate ways without having to interact further.
I realize that they're talking about the lawsuit with AT&T (I still wear my "Free The BSD 4.4" shirt with pride, though I gave away my "4.4 > 5.4" poster). What I didn't understand was the implication that that suit began anything in particular. It was a controversy raised long after the code had been used to create 386BSD and at least a decade after BSD first became popular, so I didn't see it as a "beginning".
That is to say, if you:Then that's the same thing as:The only difference being that for sub-selects, since it's done server-side, you can optimize by evlauating the sub-select lazily. Some databases that support sub-selects do not do this optimization, but most do and in some cases it's a big win. It's just a rare enough case that it's not that big a deal for the general discussion of the value of sub-selects vs other features that people want.
The non-standard SQL isn't too bad
MySQL's SQL is quite standard, and while there are differences, they are either the lack of a small number of features (like sub-selects, which you go into below) or very baroque trade-offs, which most users will not encounter.
I suspect you're instead running into assumptions about what is and is not "standard" based on what you've used in the past. MySQL follows the ANSI SQL standard as closely as all of the other databases I've used, having its own small quirks and LOTS of extensions, but basically doing as good a job of standards compliance as, say, gcc does with the ANSI C standard.
I understand your frustration, though. I'm an old Sybase user, and I miss the Sybase extensions in many places.
No subselects [...] That is huge
It is?! Why? All a sub-select does is instantiate a temporary table on the fly. You can, of course, instantiate a temporary table on the fly, so what is a sub-select buying you?
The only possible advantage would be the possibility of lazily evaluating a very large subselect, but I've come across that need in perhaps 2 or 3 situations in the course of 12 years of working with databases in production environments... how is this "huge"?
In addition, the version that I used didn't support views.
You're showing your lack of depth here. By definition a system that lacks sub-selects lacks views (as views are just named sub-selects).
All that said, sub-selects are being added to MySQL, but people who pay for work on the core server don't ask for these features as much as certain others, and that's how the folks at MySQL decide what to put effort into.
except for the lame ending.
;-)
I've written my thoughts on Stephenson's endings here before, but let me re-state: he's a very bright man, and like most of his ilk, is obviously very easily distracted. This leads to the sorts of endings that make it feel like he's left the room.
That said, you get more out of the first 90% of a Stephenson book than you do with almost all modern fiction (there are exceptions, and they're ALL worth reading, and many suffer from the same problem). Personally, I find his insights on topics ranging from nanotechnology to pipe organs useful enough to warrant suffering his endings. I even hear that the Baroque Cycle marks his first set of good endings, and I look forward to getting my copy of Quicksilver back to find out
YMMV, but I find that true insight into maters of modern technology and society are rare. This is why I grasp onto authors like Vernor Vinge and Neil Stephenson. They're the bards of our age, and we should listen and learn what we have taught them.
And about this word... I think describing problems with e-voting as "SNAFU" is the most accurate assessment I've heard to date. Bravo to these people for coming clean!
I disagree. Regardless of the particular institution, I think we're all aware of the fact that providing sensitive information to a univerisity without tight controls is amazingly dangerous. I don't blame Berkeley here, nor would I blame any school. I do think they should take this as a warning, and improve their security, but the state has to ask themselves what they were thinking, and restrict future access to this kind of information.
The process for releasing sensitive data should be:
1. Establish the exact scope of the release (who gets it)
2. Perform a reasonable amount of screening of those people based on the sensitivity of the information (ranging from simple ID checks all the way up to full security clearance procedures depending on how much risk you are exposing yourself to).
3. Perform the above two steps for any computers that will have access to the data. This means either performing a security audit, or requiring a third-party ceritification. Again, you tier the requirements based on the risk.
The state did not do that. The state got burned. If the state had handed this info to some random guy on the street and that guy had turned around and sold it to the highest bidder, I wouldn't be blaming the guy.
I don't find the character motion to be choppy, but in the event of bit-droppage over the wire, you can always do another "take". As for mouths... yeah, that would be the biggest problem, but how does the Sims version deal with that?
You could always edit in the mouth motion or just use characters in gear that covers the mouth (e.g. troll face-guards or full-warrior helms). I don't think it's insurmountable at all, and I've seen a few videos of raids that have come out so nicely that I've really wanted to see some fiction (e.g. scripted interaction vs. role playing) done with it.
I was using BSD 4.2 in college over 15 years ago, and the litigation didn't happen for quite some time.
BSD's roots are in the early 80s when they were working closely with Bell Labs, and both versions of Unix were quasi-official.
Obviously, the big break for the modern BSD was 386BSD, which brought the OS to the personal computer a little over a decade ago.
Today, I think it is the rich set of userland capabilities that distinguish the BSDs to the point that occasionally Linux distributions pop up that emulate their functionality (e.g. Gentoo's use of a BSD-like ports system).
BSD is a rich OS with a long history, and I'm glad that it's still around and growing into niches that need it. Today, I'm mostly a Linux user, but I remember my roots and the joy that life was when BSD gained popularity over the proprietary OSes of the day back in the 80s.
I would think this would be the place to do such a movie (though any popular MMORPG would have the same advantages).
The down side is having to pay a monthly fee per simultaneous actor.
Just thought of a different way to respond to this.
... several websites," you could just as easily have said, "I happen to have $1,000,000."
You say, "I happen to have creative control on several websites, two of which get a substantial amount of traffic. According to your reasoning, it's perfectly OK for me to put a George Kerry bumpersticker on my car, but it should be a federal offense for me to put a John Bush ad on my website?"
That's right. When you say, "I happen to have
However, I'm not too thrilled with the idea of one contributor having that much control over an election. If we allow contributions like that, then we create a system where a small group of big-spenders can "buy" the election.
I'm sure you don't want that any more than I do, but it's hard to see that that's the right choice when YOU are the one being restricted. I understand and empahtize, but that doesn't change my stand on this.
In fact, I'm rather a radical on this point. I'd rather see candidates cut off from the campaign process entirely. I want to see a non-governmental association of interested parties (including some governmental parties like the FEC and states representation) run the promotional aspects of an election (for all federal offices). Produce a biographical film about the candidates. Publish print, Internet and television ads which describe the candidates' stands on major issues. Schedule debates among ALL of the candidates spanning the entire summer leading up to the election, not just the two months before.
If there is only one pool of money on which to draw, and all candidates draw evenly, then we'll be able to hold an election based on the issues and the capabilities of those who wish to lead.