Keep in mind that while it's a bogus claim in the US, it may not be in other countries, and given that MySQL AB is not a US company, that could easily explain their reasoning for including the clause.
There are two ways to look at it:
1) It's open source, which is great for the open source community.
2) It's only open source, which is bad for the proprietary software community.
I choose to adopt the former point of view, but the latter is just as valid.
Yes, really. Paul's problem is a subtle one, and one that I can appreciate, but it's collateral damge of the next order. Paul's problem is that Spamhaus tracks spam by IP, and SOME spam originates from his IPs because they are shared resources. This is an unfortunate consequence of the granularity of SMTP when used in an unauthenticated way. I'd be THRILLED with something more sophisticated (such as always accepting the connection, but dropping it if their IP was blacklisted and they issued "DATA" without authenticating with a cert listed in a "cert whitelist".
Until then, we lack the tools to help Paul without helping spammers.
This is quite different from SPEWS which had the technology at their disposal to distinguish IPs, but chose not use it because the pain cause would put pressure on ISPs. That is called terrorism (sorry if post-9/11 touchiness about that word is triggered here, but I would have used the same word for it in 2000), and I don't engage in that sort of offensive.
That said, I'm fairly certain that that particular problem was dealt with, and that Yahoo! has either better patroled use of their IPs or employees are now using a different set of IPs for corporate communication.
By "innocent bystanders," do you mean [...] that "normal" users will find that the ISP cannot reliably deliver e-mail. Those users will pressure the ISP
That's the essence of the innocent bystander problem, but it's worse than that because there's really no hope that the few who complain will have any impact, so you're just causing pain for the wrong people with no benefit for the people being injured (those who recieve the spam).
Spamhaus, on the other hand, lists only those IP addresses which originate spam, and that is entirely effective without pain. I feel confident pre-filtering with SBL/XBL rules, knowing that I will not be blocking any mail that I care about. THIS is what good internet citizenry is about. Live by the rules, and you're a peer. Break the rules and you don't exist.
Given that their 52-week range is 10.37 - 26.32, and they're currently at 26.30, I'd say that you missed the best time to make that call. However, given that RHAT just entered the Nasdaq 100, and their earnings continue to grow at stupid rates, you might do well to invest in them.
I am not a stock broker or otherwise deeply knowledgeable when it comes to such matters. Do your own DD.
That defends against access AFTER the machine has been turned off, but with physical access to a machine while it's up, that does you no good. You can simply attach a debugger to a process that has legitimate access to the encrypted information, and dump the information returned from read(2) (assuming POSIX semantics).
"How about if there were no ways for this attacker to gain root priveliges from a local login"
Given physical access, that's almost impossible to arrange. For example, you could boot from external media; boot the standard system inside of a virtual machine and corrupt security attributes for running processes (elevate someone's shell to root, for example); trigger a "suspend to disk" and edit the on-disk core image before resuming; etc.
If you have money to spend, then building devices which take over control of the bus or snapshot RAM are also doable, though quite expensive, and quickly outdated. Such efforts are only worth it if you have a specific target of high value in mind.
That's an over-generalization. Sociopaths find politics easier than most, so there are a disproportionately larger number of sociopaths in politics than in the general population, but not all politicians are sociopaths.
Other work that appeals to sociopaths: executive management, military, law enforcement, lawyer. Most of the negative sterotypes of people in these fields come from the (still small fraction) who are truly sociopathic.
"You can't just start being picky over where a person resigns or not compared to when they are fired."
Can and do.
"Are you going to take responsibility for a person's actions when they don't have any interest in doing the best job they can (or at least an adequate performance so they can keep their job)?"
If you are a manager, you take responsibility for everything that the employees in your charge do. When they come to you and say, "I'm movig on," you have several responsibilities. One is to provide a smooth transition for the company. The other is to avoid burning bridges. If you need that person's help in the future as a consultant, you don't want them being bitter about how you kicked them out the door with a 2-week paycheck and a reminder about their NDA. You want them to remember the party you threw them on their last day and the way you treated them with no more or less respect in those last couple of weeks than you did during the rest of their career with you.
"Are you going to take the chance that the person who resigned isn't stealing all your corporate data and selling it to the competitor they were just hired at? What if the person gets mad/insanely drunk and does something on a whim that sets you back?"
You don't hire people like that in the first place, do you? I certainly don't. I also rely on the fact that I hire intelligent people who understand that they WANT me as a reference in the future.
"Most companies don't have the time nor the resources to play babysitter to someone that just said "I don't want to be here"."
If all you can think to do with someone who has resigned is "babysitting" them, then I suggest that you have larger problems than what happens if they steal corporate secrets, and most likely your corporate secrets aren't worth very much.
I've never worked for anyone who thought like that. In fact, I'm pretty sure I WOULDN'T work for someone who seemed to think like that. It's just unprofessional. When I resign, I give 3-4 weeks notice, and I expect the company to make the most of my time. I usually have an equity stake in the company, and I want them to succeed just as much as they do. If your company is treating you this way it is for one of two reasons: 1) they don't trust you or 2) they feel compelled to behave in a detrimental manner because a manager with either too much or too little authority thinks it's their job.
Now FIRING SOMEONE... that's different. In that case, I compartmentalize them starting the day before, backing up anything that they can touch. I then shut off their machine after they leave, remove or lock accounts and remove their remote access if they had it.
This is all as much for their benefit as mine. If they had no means of access after they found out, no one can accuse them of anything.
I also ALWAYS offer to forward people's mail, though that's gotten harder in the last few years. Companies now feel that there's too much of a chance of mail being sent to their old account with proprietary information in it. Oh well.
I was under the impression they watched some re-runs of Star Blazers (Space Cruiser Yamato) before they came up with Crusade.
First off, keep in mind that "they" (that is TNT) didn't come up with Crusade. They just broke it.
Certainly SCY had done the "leave earth in a special ship to find a way to save it" thing, and the special "all or nothing" gun on the Excalibur seems like a knowing nod in that direction (remember that JMS is a long-time SF&F fan who does like to throw in references to the genre, such as the "rangers" in B5 who were clearly modeled on Tolkein or the Gaim whose environment suits were modeled on Neil Gaiman's Sandman), but that's really where the comparison ends. SCY was about an alien saving the human race from said alien's evil counterpart, and the goal was to get to the alien in the ship. Crusade had no such set goal, only a sort of researchish mandate to explore what the First Ones had left behind and find some tech that could help. The story was also much more about trust and responsibility.
We'll never know for sure how the series would have been, but what we can say for sure is that what we saw was not JMS' original intention.
Remove "Babylon 5" from your comment. No one had ever written that many scripts per season before in the U.S. In the U.K. Blake's 7 may have come close or beat B5, I'm not sure. In the U.S., you had Twin Peaks and The Twilight Zone, but 3 straight seasons with ONE episode written by another? That had never been done. See The West Wing for a later example, though I'm not sure if Sorkin beat Straczinski's record before they fired him for being late and not hiring additional writers.
That said, I'd like to back-pedal a bit. I'm a big fan of Straczynski, but he's no Sorkin. There's an elegance to his writing when he's "on", but Sorkin pens that kind of elegance on his off-days.
I've always wanted to see a show with Aaron Sorkin, Neil Gaiman, JMS, Joss Whedon, and Peter David writing. It'd be one hell of a show, but I'm not sure that they would get it done before they killed each other;)
Well, by the same token, Hill Street Blues was a non-SF Blake's 7... and of course there were other ongoing stories before that... BUT, for the most part they weren't multi-season arcs that were designed as a single story. That was a first, and so far, there are very, very few other examples of that style of storytelling. Firefly was designed around a loose arc, but was cancelled. Nothing else I can think of has used that model in the U.S.
HSB was a great show, and yes, everything that came after it drew from that greatness, or was the worse for not having learned its lessons. I don't think that changes the impact that B5 had, though.
It was a splash of cold water to the networks. They were SURE that the SF&F markets had no interest in stories. They wanted phasers and green babes as far as they could tell. When the B5 audiences swelled, the big question was: why? When it started to win awards, there was a sort of cautious optimism. When it went a full 5 seasons, and swithched to a real network, "non-episodic" became a hollywood buzzword, and "B5-like" was a phrase applied by the marketing teams behind quite a few shows that you heard about and would never think had any connection.
Granted, there was no real, general understanding of what they had on their hands. The sequel series (Crusade) was horribly broken from day one as a result of a torrent of "notes" from TNT that destroyed any sense of what the original concept was. The suits also misunderstood the nature of the structure. They thought that X-Files and B5 had the same structure, and any attempt to explain the book-like structure of B5 was met with blank stares (I'm generalizing, this was not a universal failure, of course, just the norm). Still, there was a real change in the way Hollywood made TV, and every SF show and MANY of the non-SF shows to air since have had B5 to thank for that change. Buffy, The West Wing, Farscape, Firefly, and many other shows would likely not have been possible without B5s influence on the BUSINESS as well as on the writers, costume designers, make-up, special effects (call it cheesy now, but NO ONE thought you could do computer-generated effects for a series on-budget before B5 came out, doing it initially on Amigas).
B5 changed so much that, in retrospect, it's almost impossible to understand. You just can't bring yourself to accept that one show pushed the envelope in so many areas. JMS is an obsessive, territorial, hard-ass who many people in the industry have come to dislike, but credit where credit is due: his show was a turning point.
"it is definitely not valid for anything worth having any credibility because you can just 'edit'."
Libraries are not a useful research tool, after all, you can just 'publish'.
Clearly both of the above statements are hyperbolic and argumentative. Wikipedia is valuable because it is a massive resource with more information on more topics than any single reference I have ever used. NO SOURCE should ever be used in a vacuum, and I could certainly find a great deal of bias in any library (most of which carry "Chariots of the Gods?" for example) The New York Times and Nature have both been spoofed in the past, so no source is sacred. Still, correct use with followup and verification from other sources makes Wikipedia invaluable, and if you're not using it for your own research needs, you are failing ot take advatage of the phenomenon of global communication, IMHO.
No, the correct response is that Wikipedia is not a traditional reference source, where you turn to the "Foo" article and find the singular word on "Foo". When you use Wikipedia to perform research, you are essentially dipping into a consentual view of reality, and as such you need to analyze how that view changes over time.
You MUST perform your research using article histories as a guide. Otherwise you are doing roughly the same thing as looking at a still photo in order to analyze fluid dynamics. Wikipedia changes over time, and that change comprises valuable information about the nature of the information in question.
Someone really needs to write a "How To Research Using Wikipedia" book that explains this and many of the other complex issues facing those who wish to use it for real research. I've turned to it many times myslef, but there are difficulties in the transition that you would be foolish to ignore.
That said, yes of course he should have just edited the article and moved on. The Wikipedia community is very open to people correcting its facts, and all you have to do is edit out huge swaths of the article with the comment, "No citation is given, and the facts are contradicted in sources A and B." where A and B are valid sources that do, in fact, contradict the facts in question. Anyone who reverts such an edit will have to justify their position, or someone else will revert the reversion... leading ultimately to administrator involvement which will always come down on the side of valid cited sources vs. hearsay.
Perl has a similar thing in the Class::Contract module, which makes it easy to design by contract in Perl. It's actually pretty slick, though the Class features in modern Perl OO programming can take some getting used to for those who are used to treating Perl as an OO-free or just as a roll-your-own-OO-model language. These features are no longer just bolted on inheritance, but a true object model, similar to Perl 6.
Of course, I'm pretty sure Ruby has such a mechanism as well.
Assertions have long been a part of C and C++ programming, of course, but that's not really designing by contract, so much as performing error checking. These are different things, though certainly the latter is a superset of the former.
For assertions, my favorite language these days is C (not C++) using glib's gassert mechanism. It's easy to use, integrates smoothly into GUI applications where dialog-based feedback is needed, and can even deal with external bug-reporting systems. It's also entirely optional, and massive performance gains can be had by disabling some or all assertions when releasing the code (I always re-compile Gnome without assertions, and the speedup is noticable).
Good for you! I'm actually very glad that this has not come up for you. I'm sure there are many types of business where controling the platform isn't neccessary, but they're in the vast minority, and testing should account for the majority, I would think.
Why does a KDE announcement have to be about Gnome? KDE is fine, and while I use a Gnome desktop, I often run some excellent KDE apps like K3B. The two actually have very few core differences, and the casual user usually ends up paying more attention to the default configuration differences, and never even knowing about the true divergence between them anyway.
"The problem is that encryption without authentication is really not secure as you'd be vulnerable to a man in the middle attack."
I assumed this was an understood element of the premise of my comments, but there are times (such as those that I mentioned) where you don't particularly care about that sort of scenario because either authentication is provided by an external element or because security concerns which render authentication unreliable exist anyway.
In those circumstances, doing needless authentication actually hurts security, since failures will be regarded as "noise", training users to ignore otherwise critical security.
If authentication were only done when and where it mattered most, it would be taken much more seriously, and would thus be far more reliable universally.
Encrypting traffic routinely, however, is something that, IMHO, should always happen unless you're transmitting data which is so bandwidth sensitive that the overhead of encryption would be prohibitive (this is rarely the case, but some sorts of streaming media would be impacted).
You may consider the constraint unfair, but it's a perfectly practical and realistic business constraint that Linux has to cope with on a daily basis.
In fact, that's one of my overriding complaints about Linux software. There's this sort of loose assumption that backward compatibility isn't required because you can just download the source for something new. But, when you work in an environment where you have 10 applications, each with its own realease cycle, you have to adopt a platform from hardware all the way up to OS and tools for those applications to target. You can't just upgrade at the drop of a hat without chaning the deadlines for half a dozen of your projects.
So when you discover that project A's new widget will require a security update to the software it depends on, and that will require a new version of libc, you're totally screwed. It's nice to live in the "it's my machine, and I'll upgrade when I like," world, but if you're going to compare OSes for the enterprise, you're talking about a very different ball of wax.
First off, I've provided server arrays to such organizations before, and honestly I've never seen a "big iron" server that was purchased with any more justification than the phrase "big iron" turned on some executive. I've seen Oracle installations on single (or few) large systems work as well and as poorly as their open source competition from MySQL and PostgreSQL in VERY large environments. Some also hybridize (I think Sabre announced on a Webcast at one point that they used MySQL front-ends with DB2 back-ends), and that works well too.
The question is: what are your costs, and what do you get for those costs? In my experience the costs are far higher in terms of systems and administration overhead with Oracle. If your experience differs, then we'll have to agree to disagree until one of us is exposed to a sample of the others' experience.
They did not just rebuild source RPMS because that would have violated business constraints, which were the basis for comparison.
He did comment that thre admins provided feedback saying that they would have considered a distribution upgrade over the glibc upgrade if they were allowed to. That would seem to me to be a more likely path for a business to have taken. Still, for the constraints posed, this was a fairly valid test (and remember that the constraints were posed on both sides).
Stu, you've clearly had different experiences from me. I've fought with Oracle and watched others do the same many, many times. If you've come away unscathed, then more power to you (drop me a line if you're interested in interviewing), but I've just seen so many horror stories now that I'm tired of trying. Every time I deploy any other database it's a 5 minute install and little or no tweeking. Every time I or a DBA installs Oracle it's a 1-2 day install that requires care and feeding for days or weeks before it's in a state that can be left unattended.
Mind you, claims that more extensive reading of documentation would help seem a tad off the mark, since other databases get the same cursory treatment that 99% of system tools get. Your database simply shouldn't be a major learning experience, any more than your filesystem is. You should be able to set it up with a few simple commands that are packaged appropriately for the operating system and applications that need it should then be able to use it.
"A little known fact (not described in the High School history books) is that the Pilgrims landed at Plymouth because they had run out of beer and needed to make more."
This is a common misrepresentation of the situation. They were running low on provisions in general, but the most urgent was drinkable liquid. Because beer keeps even better than water, but has enough nutrients to keep you alive for quite a while, it's a last-resort foodstuff. The fact that they were running low on ale probably indicates that they were dangerously close to dying.
I would not say that they stopped to make more beer, so much as they stopped to avoid dying because all they had left was beer, and not much of it. Those are very different situations.
Generally any view of history that makes the Plymouth pilgims sound cooler should be regarded with heavy suspicion.
I've watched dozens of projects in a handful of companies move from other databases to Oracle. These other databases have included Fame, Sybase, BDB, MSSQL, MS Access, MySQL, Postgress, PostgreSQL, Ingres, and a variety of lesser known databases.
NEVER, have I seen such a project which I would refer to as "painless". Oracle is a monolithic beast which requires constant care and feeding by experts who have been so steeped in its ways that they are prohibitively expensive. Oracle perpetuates this situation and, as best I can tell, deliberately obfuscates their product in order to continue to rake in huge fees for training and services. Mind you, I've done Oracle DBA work in the past, so I'm not talking through my hat, here.
I would never, ever rely on Oracle. I use it as a bargaining tool in project management ("if we hit a wall with PostgreSQL/MySQL/etc. we can always switch to Oracle"), but unless there is a practical reason such as customer requirements which cannot be altered by me, I dodge it, and I would suggest that others do the same.
This has nothing to do with petty in-fighting about what feature-set is better. This is about a company that makes its money by abusing its customers. Why would you stand for that, even if it means giving up some feature, or working harder to implement some feature with a different tool? I would rather hire 5 MySQL or PostgreSQL engineers that work on expanding their feature-sets full-time, than pay Oracle a licensing fee.
Slashdot Mirror
Keep in mind that while it's a bogus claim in the US, it may not be in other countries, and given that MySQL AB is not a US company, that could easily explain their reasoning for including the clause.
There are two ways to look at it:
1) It's open source, which is great for the open source community.
2) It's only open source, which is bad for the proprietary software community.
I choose to adopt the former point of view, but the latter is just as valid.
Yes, really. Paul's problem is a subtle one, and one that I can appreciate, but it's collateral damge of the next order. Paul's problem is that Spamhaus tracks spam by IP, and SOME spam originates from his IPs because they are shared resources. This is an unfortunate consequence of the granularity of SMTP when used in an unauthenticated way. I'd be THRILLED with something more sophisticated (such as always accepting the connection, but dropping it if their IP was blacklisted and they issued "DATA" without authenticating with a cert listed in a "cert whitelist".
Until then, we lack the tools to help Paul without helping spammers.
This is quite different from SPEWS which had the technology at their disposal to distinguish IPs, but chose not use it because the pain cause would put pressure on ISPs. That is called terrorism (sorry if post-9/11 touchiness about that word is triggered here, but I would have used the same word for it in 2000), and I don't engage in that sort of offensive.
That said, I'm fairly certain that that particular problem was dealt with, and that Yahoo! has either better patroled use of their IPs or employees are now using a different set of IPs for corporate communication.
Spamhaus, on the other hand, lists only those IP addresses which originate spam, and that is entirely effective without pain. I feel confident pre-filtering with SBL/XBL rules, knowing that I will not be blocking any mail that I care about. THIS is what good internet citizenry is about. Live by the rules, and you're a peer. Break the rules and you don't exist.
"Maybe it's time to invest in RHAT"
Given that their 52-week range is 10.37 - 26.32, and they're currently at 26.30, I'd say that you missed the best time to make that call. However, given that RHAT just entered the Nasdaq 100, and their earnings continue to grow at stupid rates, you might do well to invest in them.
I am not a stock broker or otherwise deeply knowledgeable when it comes to such matters. Do your own DD.
"How about an encrypted filesystem?"
That defends against access AFTER the machine has been turned off, but with physical access to a machine while it's up, that does you no good. You can simply attach a debugger to a process that has legitimate access to the encrypted information, and dump the information returned from read(2) (assuming POSIX semantics).
"How about if there were no ways for this attacker to gain root priveliges from a local login"
Given physical access, that's almost impossible to arrange. For example, you could boot from external media; boot the standard system inside of a virtual machine and corrupt security attributes for running processes (elevate someone's shell to root, for example); trigger a "suspend to disk" and edit the on-disk core image before resuming; etc.
If you have money to spend, then building devices which take over control of the bus or snapshot RAM are also doable, though quite expensive, and quickly outdated. Such efforts are only worth it if you have a specific target of high value in mind.
That's an over-generalization. Sociopaths find politics easier than most, so there are a disproportionately larger number of sociopaths in politics than in the general population, but not all politicians are sociopaths.
Other work that appeals to sociopaths: executive management, military, law enforcement, lawyer. Most of the negative sterotypes of people in these fields come from the (still small fraction) who are truly sociopathic.
"You can't just start being picky over where a person resigns or not compared to when they are fired."
Can and do.
"Are you going to take responsibility for a person's actions when they don't have any interest in doing the best job they can (or at least an adequate performance so they can keep their job)?"
If you are a manager, you take responsibility for everything that the employees in your charge do. When they come to you and say, "I'm movig on," you have several responsibilities. One is to provide a smooth transition for the company. The other is to avoid burning bridges. If you need that person's help in the future as a consultant, you don't want them being bitter about how you kicked them out the door with a 2-week paycheck and a reminder about their NDA. You want them to remember the party you threw them on their last day and the way you treated them with no more or less respect in those last couple of weeks than you did during the rest of their career with you.
"Are you going to take the chance that the person who resigned isn't stealing all your corporate data and selling it to the competitor they were just hired at? What if the person gets mad/insanely drunk and does something on a whim that sets you back?"
You don't hire people like that in the first place, do you? I certainly don't. I also rely on the fact that I hire intelligent people who understand that they WANT me as a reference in the future.
"Most companies don't have the time nor the resources to play babysitter to someone that just said "I don't want to be here"."
If all you can think to do with someone who has resigned is "babysitting" them, then I suggest that you have larger problems than what happens if they steal corporate secrets, and most likely your corporate secrets aren't worth very much.
I've never worked for anyone who thought like that. In fact, I'm pretty sure I WOULDN'T work for someone who seemed to think like that. It's just unprofessional. When I resign, I give 3-4 weeks notice, and I expect the company to make the most of my time. I usually have an equity stake in the company, and I want them to succeed just as much as they do. If your company is treating you this way it is for one of two reasons: 1) they don't trust you or 2) they feel compelled to behave in a detrimental manner because a manager with either too much or too little authority thinks it's their job.
Now FIRING SOMEONE... that's different. In that case, I compartmentalize them starting the day before, backing up anything that they can touch. I then shut off their machine after they leave, remove or lock accounts and remove their remote access if they had it.
This is all as much for their benefit as mine. If they had no means of access after they found out, no one can accuse them of anything.
I also ALWAYS offer to forward people's mail, though that's gotten harder in the last few years. Companies now feel that there's too much of a chance of mail being sent to their old account with proprietary information in it. Oh well.
I was under the impression they watched some re-runs of Star Blazers (Space Cruiser Yamato) before they came up with Crusade.
First off, keep in mind that "they" (that is TNT) didn't come up with Crusade. They just broke it.
Certainly SCY had done the "leave earth in a special ship to find a way to save it" thing, and the special "all or nothing" gun on the Excalibur seems like a knowing nod in that direction (remember that JMS is a long-time SF&F fan who does like to throw in references to the genre, such as the "rangers" in B5 who were clearly modeled on Tolkein or the Gaim whose environment suits were modeled on Neil Gaiman's Sandman), but that's really where the comparison ends. SCY was about an alien saving the human race from said alien's evil counterpart, and the goal was to get to the alien in the ship. Crusade had no such set goal, only a sort of researchish mandate to explore what the First Ones had left behind and find some tech that could help. The story was also much more about trust and responsibility.
We'll never know for sure how the series would have been, but what we can say for sure is that what we saw was not JMS' original intention.
Remove "Babylon 5" from your comment. No one had ever written that many scripts per season before in the U.S. In the U.K. Blake's 7 may have come close or beat B5, I'm not sure. In the U.S., you had Twin Peaks and The Twilight Zone, but 3 straight seasons with ONE episode written by another? That had never been done. See The West Wing for a later example, though I'm not sure if Sorkin beat Straczinski's record before they fired him for being late and not hiring additional writers.
;)
That said, I'd like to back-pedal a bit. I'm a big fan of Straczynski, but he's no Sorkin. There's an elegance to his writing when he's "on", but Sorkin pens that kind of elegance on his off-days.
I've always wanted to see a show with Aaron Sorkin, Neil Gaiman, JMS, Joss Whedon, and Peter David writing. It'd be one hell of a show, but I'm not sure that they would get it done before they killed each other
Well, by the same token, Hill Street Blues was a non-SF Blake's 7... and of course there were other ongoing stories before that... BUT, for the most part they weren't multi-season arcs that were designed as a single story. That was a first, and so far, there are very, very few other examples of that style of storytelling. Firefly was designed around a loose arc, but was cancelled. Nothing else I can think of has used that model in the U.S.
HSB was a great show, and yes, everything that came after it drew from that greatness, or was the worse for not having learned its lessons. I don't think that changes the impact that B5 had, though.
It was a splash of cold water to the networks. They were SURE that the SF&F markets had no interest in stories. They wanted phasers and green babes as far as they could tell. When the B5 audiences swelled, the big question was: why? When it started to win awards, there was a sort of cautious optimism. When it went a full 5 seasons, and swithched to a real network, "non-episodic" became a hollywood buzzword, and "B5-like" was a phrase applied by the marketing teams behind quite a few shows that you heard about and would never think had any connection.
Granted, there was no real, general understanding of what they had on their hands. The sequel series (Crusade) was horribly broken from day one as a result of a torrent of "notes" from TNT that destroyed any sense of what the original concept was. The suits also misunderstood the nature of the structure. They thought that X-Files and B5 had the same structure, and any attempt to explain the book-like structure of B5 was met with blank stares (I'm generalizing, this was not a universal failure, of course, just the norm). Still, there was a real change in the way Hollywood made TV, and every SF show and MANY of the non-SF shows to air since have had B5 to thank for that change. Buffy, The West Wing, Farscape, Firefly, and many other shows would likely not have been possible without B5s influence on the BUSINESS as well as on the writers, costume designers, make-up, special effects (call it cheesy now, but NO ONE thought you could do computer-generated effects for a series on-budget before B5 came out, doing it initially on Amigas).
B5 changed so much that, in retrospect, it's almost impossible to understand. You just can't bring yourself to accept that one show pushed the envelope in so many areas. JMS is an obsessive, territorial, hard-ass who many people in the industry have come to dislike, but credit where credit is due: his show was a turning point.
"you research using wikipedia?"
Of course.
"it is definitely not valid for anything worth having any credibility because you can just 'edit'."
Libraries are not a useful research tool, after all, you can just 'publish'.
Clearly both of the above statements are hyperbolic and argumentative. Wikipedia is valuable because it is a massive resource with more information on more topics than any single reference I have ever used. NO SOURCE should ever be used in a vacuum, and I could certainly find a great deal of bias in any library (most of which carry "Chariots of the Gods?" for example) The New York Times and Nature have both been spoofed in the past, so no source is sacred. Still, correct use with followup and verification from other sources makes Wikipedia invaluable, and if you're not using it for your own research needs, you are failing ot take advatage of the phenomenon of global communication, IMHO.
I had numbers, but they are VERY old. I'll have to perform a test with modern Gnome at some point.
No, the correct response is that Wikipedia is not a traditional reference source, where you turn to the "Foo" article and find the singular word on "Foo". When you use Wikipedia to perform research, you are essentially dipping into a consentual view of reality, and as such you need to analyze how that view changes over time.
You MUST perform your research using article histories as a guide. Otherwise you are doing roughly the same thing as looking at a still photo in order to analyze fluid dynamics. Wikipedia changes over time, and that change comprises valuable information about the nature of the information in question.
Someone really needs to write a "How To Research Using Wikipedia" book that explains this and many of the other complex issues facing those who wish to use it for real research. I've turned to it many times myslef, but there are difficulties in the transition that you would be foolish to ignore.
That said, yes of course he should have just edited the article and moved on. The Wikipedia community is very open to people correcting its facts, and all you have to do is edit out huge swaths of the article with the comment, "No citation is given, and the facts are contradicted in sources A and B." where A and B are valid sources that do, in fact, contradict the facts in question. Anyone who reverts such an edit will have to justify their position, or someone else will revert the reversion... leading ultimately to administrator involvement which will always come down on the side of valid cited sources vs. hearsay.
If you do, and I have mod points, I might just take the meta-mod hit and bump you up a notch ;)
Perl has a similar thing in the Class::Contract module, which makes it easy to design by contract in Perl. It's actually pretty slick, though the Class features in modern Perl OO programming can take some getting used to for those who are used to treating Perl as an OO-free or just as a roll-your-own-OO-model language. These features are no longer just bolted on inheritance, but a true object model, similar to Perl 6.
Of course, I'm pretty sure Ruby has such a mechanism as well.
Assertions have long been a part of C and C++ programming, of course, but that's not really designing by contract, so much as performing error checking. These are different things, though certainly the latter is a superset of the former.
For assertions, my favorite language these days is C (not C++) using glib's gassert mechanism. It's easy to use, integrates smoothly into GUI applications where dialog-based feedback is needed, and can even deal with external bug-reporting systems. It's also entirely optional, and massive performance gains can be had by disabling some or all assertions when releasing the code (I always re-compile Gnome without assertions, and the speedup is noticable).
Good for you! I'm actually very glad that this has not come up for you. I'm sure there are many types of business where controling the platform isn't neccessary, but they're in the vast minority, and testing should account for the majority, I would think.
Why does a KDE announcement have to be about Gnome? KDE is fine, and while I use a Gnome desktop, I often run some excellent KDE apps like K3B. The two actually have very few core differences, and the casual user usually ends up paying more attention to the default configuration differences, and never even knowing about the true divergence between them anyway.
"The problem is that encryption without authentication is really not secure as you'd be vulnerable to a man in the middle attack."
I assumed this was an understood element of the premise of my comments, but there are times (such as those that I mentioned) where you don't particularly care about that sort of scenario because either authentication is provided by an external element or because security concerns which render authentication unreliable exist anyway.
In those circumstances, doing needless authentication actually hurts security, since failures will be regarded as "noise", training users to ignore otherwise critical security.
If authentication were only done when and where it mattered most, it would be taken much more seriously, and would thus be far more reliable universally.
Encrypting traffic routinely, however, is something that, IMHO, should always happen unless you're transmitting data which is so bandwidth sensitive that the overhead of encryption would be prohibitive (this is rarely the case, but some sorts of streaming media would be impacted).
You may consider the constraint unfair, but it's a perfectly practical and realistic business constraint that Linux has to cope with on a daily basis.
In fact, that's one of my overriding complaints about Linux software. There's this sort of loose assumption that backward compatibility isn't required because you can just download the source for something new. But, when you work in an environment where you have 10 applications, each with its own realease cycle, you have to adopt a platform from hardware all the way up to OS and tools for those applications to target. You can't just upgrade at the drop of a hat without chaning the deadlines for half a dozen of your projects.
So when you discover that project A's new widget will require a security update to the software it depends on, and that will require a new version of libc, you're totally screwed. It's nice to live in the "it's my machine, and I'll upgrade when I like," world, but if you're going to compare OSes for the enterprise, you're talking about a very different ball of wax.
First off, I've provided server arrays to such organizations before, and honestly I've never seen a "big iron" server that was purchased with any more justification than the phrase "big iron" turned on some executive. I've seen Oracle installations on single (or few) large systems work as well and as poorly as their open source competition from MySQL and PostgreSQL in VERY large environments. Some also hybridize (I think Sabre announced on a Webcast at one point that they used MySQL front-ends with DB2 back-ends), and that works well too.
The question is: what are your costs, and what do you get for those costs? In my experience the costs are far higher in terms of systems and administration overhead with Oracle. If your experience differs, then we'll have to agree to disagree until one of us is exposed to a sample of the others' experience.
They did not just rebuild source RPMS because that would have violated business constraints, which were the basis for comparison.
He did comment that thre admins provided feedback saying that they would have considered a distribution upgrade over the glibc upgrade if they were allowed to. That would seem to me to be a more likely path for a business to have taken. Still, for the constraints posed, this was a fairly valid test (and remember that the constraints were posed on both sides).
Stu, you've clearly had different experiences from me. I've fought with Oracle and watched others do the same many, many times. If you've come away unscathed, then more power to you (drop me a line if you're interested in interviewing), but I've just seen so many horror stories now that I'm tired of trying. Every time I deploy any other database it's a 5 minute install and little or no tweeking. Every time I or a DBA installs Oracle it's a 1-2 day install that requires care and feeding for days or weeks before it's in a state that can be left unattended.
Mind you, claims that more extensive reading of documentation would help seem a tad off the mark, since other databases get the same cursory treatment that 99% of system tools get. Your database simply shouldn't be a major learning experience, any more than your filesystem is. You should be able to set it up with a few simple commands that are packaged appropriately for the operating system and applications that need it should then be able to use it.
"A little known fact (not described in the High School history books) is that the Pilgrims landed at Plymouth because they had run out of beer and needed to make more."
This is a common misrepresentation of the situation. They were running low on provisions in general, but the most urgent was drinkable liquid. Because beer keeps even better than water, but has enough nutrients to keep you alive for quite a while, it's a last-resort foodstuff. The fact that they were running low on ale probably indicates that they were dangerously close to dying.
I would not say that they stopped to make more beer, so much as they stopped to avoid dying because all they had left was beer, and not much of it. Those are very different situations.
Generally any view of history that makes the Plymouth pilgims sound cooler should be regarded with heavy suspicion.
I've watched dozens of projects in a handful of companies move from other databases to Oracle. These other databases have included Fame, Sybase, BDB, MSSQL, MS Access, MySQL, Postgress, PostgreSQL, Ingres, and a variety of lesser known databases.
NEVER, have I seen such a project which I would refer to as "painless". Oracle is a monolithic beast which requires constant care and feeding by experts who have been so steeped in its ways that they are prohibitively expensive. Oracle perpetuates this situation and, as best I can tell, deliberately obfuscates their product in order to continue to rake in huge fees for training and services. Mind you, I've done Oracle DBA work in the past, so I'm not talking through my hat, here.
I would never, ever rely on Oracle. I use it as a bargaining tool in project management ("if we hit a wall with PostgreSQL/MySQL/etc. we can always switch to Oracle"), but unless there is a practical reason such as customer requirements which cannot be altered by me, I dodge it, and I would suggest that others do the same.
This has nothing to do with petty in-fighting about what feature-set is better. This is about a company that makes its money by abusing its customers. Why would you stand for that, even if it means giving up some feature, or working harder to implement some feature with a different tool? I would rather hire 5 MySQL or PostgreSQL engineers that work on expanding their feature-sets full-time, than pay Oracle a licensing fee.