What if he's pissed off at being forced to resign, and this is as close as he can come to saying, look, 90% of the machines in the world are running monitoring software that we're able to introduce specifically because security is so damned lax. Fix security, and the CIA would have a much harder time doing its job.
One of the funniest replies I've read in this thread claimed that researchers cannot get funding unless they shout "Doomsday is coming". Not even the most imaginary hypothetical example is cited - the report is simply made up by some liberal asshole.
Ah... why would this person be liberal. Aren't liberals the ones screaming doomsday? I'm a moderate liberal who thinks that the left is far too entrenched with the idea that humans are "destroying the planet". I see daily evidence of this in many forms. One: when interviewed by Nova, a researcher who had discoverd massive greenhouse gas emmisions from recent forest fires that burned permafrost material (greenhouse gasses that would not have been released before, due to the change in fire prevention strategies and capabilities) said that this was probably responsible for more greenhouse gasses (CO2 and water vapor) than fossil fuels, but "we don't know how much it takes to tip the balance." Even in the face of cold, hard evidence that the single largest problem was not fossil fuels, this guy's very first reaction was to defend the efforts to curb fossil fuels... not the reaction of someone who is interested in the objective facts.
Two: I have a good friend whose Solar observatory was labeled in congressional funding circles as "an enemy of the planet" (I kid you not), for having published data (not conclusions, data) that supported a non-anthropomorphic warming model and could not secure further funding.
Three: I watch as scientists around the world attempt to discredit any research done with private dollars, without any real concern for the quality of the research itself. In every other field, private research is judged on the basis of its scientific merit or lack thereof, but if you get money from an oil company, car company or heavy industry, and do atmospheric research you can't get published in a reputable journal regardless of the quality of your work.
Put those three together, and come up with a model for me of how research is not skewed toward finding a link between human activity and global warming.
"Reading Slashdot on such topics makes you think the world is really only divided into two kinds of persons - the coporate man/politicians and the crazy gaians"
No, being polar about politics makes you think that.
Ok, clearly I'm addicted to this thread... I need help;-)
"Understandable, but this assumes that their PAYING customers know what they are talking about. I use MySQL every day and I've used RDBMSs quite a bit. The feeling that I get from MySQL is that they think a lot of "features" of RDBMSs are just "fluff" (like transactions, stored procedures, foreign key constraints at one time or another)"
They don't, and if you get the chance to listen to them speak about the development of the database at a conference or the like, I highly recommend that you do so.
What they've said is that those features are great, but their existing user base was made up of people who were able to extract value from MySQL specifically because things that required those features were not their core concerns.
MySQL is often used as a front-end cache for other databases, for highly volatile data such as Web content and other information which simply does not require those sorts of features, and which oftne suffers from their implementation.
That's not to say that MySQL should not implement those features, just that they must be taken slowly and carefully, and implemented in such a way that the existing user-base don't have to pay the price for features they didn't ask for.
For example, let me take one that you did not touch on: sub-selects. MySQL doesn't do sub-selects.
Of course, you can emulate them in your code like so:
select x from a where a.y in (select y from b where z = 1); # Becomes: create temporary table tmp1 select y from b where z = 1; select x from a, tmp1 where a.y = tmp1.y;
This eliminates a key optimization that sub-selects can perform, but is otherwise identical. That key optimization is to implement the temporary table in such a way that values are returned before either statement is finished executing. This is a major win in terms of the time between request and first answer, but should result in roughly the same execution time overall.
In this case, the small win from the extra optimization and shorter code would have come at the cost of other MySQL development that existing, paying customers wanted more and the changes to select logic that would have allowed for lazy evaluation might have resulted in more checks along the way.
So the feature waited on the back-burner while they worked on things like InnoDB integration and full-text searching. Personally I see the results as a win, and in the end I'll get sub-selects, but not at the cost of features I need.
"Understandable, but this assumes that their PAYING customers know what they are talking about."
Not really. It assumes that implementing what their paying customers want will give them the time and resources to implement what others do. This is a good thing.
" it is *not* an RDBMS and shouldn't be used in place of one."
This is completely untrue. Seriously, unlearn this "goodfact" that's been circulating on Slashdot and other places, as it's simply false.
An RDBMS is a simple category of datastore which implements a style of set logic which relies on relations between like set elements. To put it in crude terms, anything that can implement the most basic SQL select with an inner join is an RDBMS. The logic that has been used to corrupt the term is this: We need more features for modern applications than simple relations, so RDBMS is an insufficient term. Thus, to be an RDBMS, you must perform some arbitrary (and often ill-defined) set of other functions. People have thrown in ACID (a term you don't hear as much any more since modern MySQL engines are ACID), replication, object oriented relationships, and any number of other features which are quite simply not part of the meaning of the term RDBMS.
Say what you mean. MySQL is most certainly a fully-functional RDBMS. That it may not have some of the features of some other RDBMS or even some features that MOST other modern RDBMSs give you is no excuse for arbitrarily changing the defintion of RDBMS.
There's an open source network called Gnutella, developed using open standards, and with many open source implementations.
Originally it was only a point-to-point download service negotiated by a peer-to-peer query-flooding system. Times have changed.
Today's Gnutella (go check out gtk-gnutella or any of the other fine clients for details) is more like a fully decentrallized BT. There are still some flaws in the protocol with respect to poisoning, but it's certainly at the point now where I can use it to fetch operating systems, television (I've used it to catch up on shows that my TiVo has missed several times now, shame on me for show-loyalty;-) and lots of great digitial photography (yes, you can find some amazing art that actually involves no skin-tones).
"I use MySQL in maybe 60% of my projects, and I have been bitten by a majority of those issues"
Ok, look someone's not reading here... I'll make this one last post, but I really have to stop this thread.
I never said that the items listed were incorrect. They would not work as rhetoric if they were lies. However, the page is designed, not to help (all of the items listed are in the documentation on the MySQL site), but to discourage the use of the database. Any database can be ranted about, but for some unknown reason people choose to rant about MySQL quite a lot.
I think part of this is that PostgreSQL as a project has felt that their time and effort on full ANSI compliance was not rewarded, and so a lot of specifically PostgreSQL folks want to rail on about how MySQL did not make such an attempt, but instead focused on non-standard features like their full-text searching.
That's fine, but let's not confuse those disagreements for any actual problems with either product.
MySQL has some serious problems as an SQL-compliant database (let alone as a *relational* database)
You got those backwards. SQL compliance implies that you are a relational (e.g. set theory applied to a table structure) database. MySQL might be non ANSI SQL 9X compliant, but that does not make it a non-relational database. There are ANSI SQL features which MySQL simply does not support.
and that site simply points them out in an objective manner. In fact he says they aren't bugs or mistakes, but *documented, yet unexpected, behavior*.
It is most certainly not objective, and if you really think that it is, then I don't think we can have this conversation rationally.
Ok, that's it. Thanks guys, but I really do need to go get on with my life. If my comments here have been helpful, great, if not... well, good luck with your work anyway.
I was refering to the fact that this link shows up in EVERY MySQL related article, within minutes. That, I call carpet bombing. Use your own terminology as you see fit.
In fact, the only single person I see prolonging this thread (by responing to these so-called "anti-MySQL" people as you like to label them) is you.
One, I've never refered to anyone in this thread as an anti-MySQL person. Not a one. Please, don't re-write my comments.
Two, you are right: I've put too much effort into this, and no one really cares anyway besides folks like you that feel so strongly that over the course of this thread, you actually marked me as a foe.
Well, that's hard to say. Peer reviewed journals from that period do not publish their tables of contents that far back on the Web (Nature, for example, goes back to 1980), so someone would have to go hit a physical library to get you that citation. Perhaps you know of an online index of such documents that I don't... I rarely need documents older than 20 years in my line of work.
HOWEVER, we can be fairly certain that such articles do exist. Phrenology was a very popular theory, and the scientific community would have welcomed papers on the subject.
The real topic was that peer review does not guarantee correctness, only mainstream scientific respect. That's a decent generic baseline, but saying, "this guy on slashdot made a typo, so he can't have seen a flaw in this research," doesn't make a whole lot of sense. The current environment (no pun intended) in the scientific community surrounding global warming is hostile to any information (not even theories, but raw data) which suggests that humans may not play a significant role in said warming. In that environment, you are going to see some junk science published in respected journals. This happens every time there is a major political topic dominating a field.
Global warming may well be human induced, I have no idea, but I can tell you with certainty that we're going to have to dig ourselves out of at least 20 years of heavily biased research before we understand how. One example: scientists discovered that forest fires that burn hotter now due to fire prevention efforts over the last 100 years are able to burn permafrost and release HUGE amounts of carbon dioxide and water vapor (greenhouse gasses). Upon discovering this, the immediate reaction was, "well, this is probably reponsible for more greenhouse emmisions than humans, but we can't know what would tip the balance." There is no objectivity in this field and all data comes with a set of preconceptions that are going to be very hard to break through.
Good points. However, if you want the benefit of such a system without the development time, just use existing Gnutella.
When Gnutella started, it was a simple query-flooding system. It's now much, much more and truly resembles a decentralized BT.
When you search, you get a result. When you request a file, all of the servents that the servent from which you madethe request knows about are told to send you a UDP packet, informing you that they can help with your reuqest. This sort of decentralized swarming allows you to skip right past the centralized server or even a network of them.
The only down-side that I see after discussiong the protocol differences with a friend is this: BT can validate each chunk. Gnutella can validate each chunk (and does), but it's kind of moot since the source of the chunk gave you the checksum in the first place. If someone poisons a chunk you can find that out heuristically over time, but you're going to get to the end of a lot of files and find that they're corrupt before you de-prioritize the poisoner sufficiently.
A smart poisoner could even make it near impossible to be found. This is fixable, but would require protocol extensions for requesting and sending the checksums of all chunks from the original servent that told you about the file. Discovering that such a servent was lying would be much easier over time.
Of course, these problems are hard to avoid in any truly distributed model.
"The point of the many of the 'gotchas' is that this software behaves erractically: rather than produce an exception during an overflow or conversion error, for example, it just silently modifies the data and returns no warning to the user."
A well documented convinience for applications which always behaves the same way.
Please explain your definition of erratic for me, because mine doesn't seem to match the example you give.
"Two years ago the folks from MySQL stated that transactions, views, subselects, etc weren't needed by 99% of the applications out there"
Wrong. Simply wrong. What they said is that their PAYING CUSTOMERS prioritize their work, and while having everything would be nice, they have to do things in the order that puts the bread on the table.
Transactions were a nice little freebie that fell out of BerkelyDB and InnoDB table types, so fairly little work was required. For something like sub-selects, vies, and other goodies, I'd much rather have all of the features of 4.0 and 4.1 before those. Your milage may vary, and you're welcome to use something else. It's just that you don't need to carpet-bomb every article about products you don't care for.
"Anyhow - please don't blame us victims"
Ok, you really need to take a step back and breath... you're not a "victim" of a database. Please.
"MySQL AB admits their product is a "lite" database, and doesn't outright disparage basic features of databases as unnecessary"
Well, actually they don't. What they've said is that their PAYING customers direct their efforts and those customers prioritize differently than the "MySQL sucks" crowd. It's not terribly surprising which one gets their features implemented first.
What they have said is that they will not implement features that come with (what they percieve to be) major performance hits without finding a way to do so whithout slowing down everything. In many cases, they have found those ways (e.g. by providing InnoDB table types, but still providing MYISAM tables).
Other databases also contribute to this by discovering ways to make the tradeoffs less painful (or even not at all painful). I applaud, for example, much of the work done by Oracle on their unique caching strategies, and while it's not open source, they still contribute valuable data points. PostgreSQL and other open source databases also contribute hugely, just as MySQL contributes to them (would PostgreSQL have moved beyond its difficult-to-admin stage if MySQL were not so damned simple to install and set up?)
We're the OPEN SOURCE COMMUNITY, and we really should start acting like it.
Are we so mainstream in the open source world now that we have bought into the utterly bankrupt idea that every story has two (and only two) sides? Someone who rails against PostgreSQL and MySQL must have a balanced view, and deserves to be linked to from anything that discusses either? Eh, how's that?
Please, it's just a rant. He can rant about all the products he wants, I just don't see why every mention of MySQL has to link to this diatribe.
What is this... I'm really freaking confused. MySQL is used by tens if not hundreds of thousands of people around the world. It has always done what I want it to do faithfully. I'm a Sybase weenie with significant Oracle and Ingres experience. I've been working with and writing software for 15 years and I honestly do not understand what is so damned scary about MySQL. Send query, get answer. Life good. Carry on.
If it doesn't do what you want, fine don't use it, but please stop acting as if it's going to poison the water supply.
This is a bit like disclaiming your political views by saying you've voted constitution party AND green party.
A lot of folks think that MySQL is a good idea
It's software that does what you tell it to do. Thus, it is a good idea. You think software product X is beter, meets some criteria that you find compelling, can do things that you want... great, that's fine, but like I said, that opinion doesn't need to be spattered all over every occurance of software product Y on slashdot.
"How is this a rant? Are you saying the problems they list don't exist?"
Not at all. What I'm saying is that it's a page that was meant to discredit MySQL by putting together a sort of reverse-FAQ. Instead of saying, "people run into X, well here's what they do when they run into that," it's just a diatribe about how lacking someone thinks MySQL is. That, as they say, ain't news, and it most CERTAINLY does not deserve a fresh link from every MySQL article to hit Slashdot.
I guess it's nice, but a decentralized BT is called Gnutella. Check out the current state of the Gnutella protocol. Swarming, UDP chunk notification, etc. All of the BT features you expect are there, and if you're downloading sufficiently popular data, it's screamingly fast (just like BT).
What I'm waiting for is a non-proprietary client that recognizes torrents over Gnutella and can fetch them via the BT protocol. That would eliminate that rare need I have to fire up a BT client.
I use gtk-gnutella, which is quite nice, but I hear good things about other clients these days too.
That is NOT a list of "common gotchas" (the misleading name of the page aside), it's an anti-MySQL rant that has lived on far too long.
Please, can we for once have a post about a piece of software on Slashdot without the pro-X or anti-Y folks flocking to it to bash or praise it? Can we just for once talk about the damn book?
Ah ok, this is a misunderstanding then. Most of what I would refer to as "support" is either being defered (after all your time is expensive, so your clients don't contact you unless they NEED you) or isn't being done at all (e.g. your comment, "who wants to pay to have every possible hole patched up to complete").
Being a consultant makes a world of difference. Sure, you can sell support for 1000 machines, but that's very different from being the IT staff in a company with 1000 machines. VERY different.
Still, offering support for 1000 end-users is no small task, and my hat's off to you.
I can say for certain, that I can maintain 1000 desktop systems by myself, and about 50 servers
I'm sorry, I just don't buy that. If you mean that you can support one aspect of running those machines, sure, but you're doing it all? How do you have enough time to dedicate to security updates (and response to any security problems), external complaints, administrative requests, new users and hardware, MTBF replacements, user questions, new software install requests, expansion of physical space and networks, physical upgrades, and the inevitable development of tools and small software? Over 1000 systems? On top of that, you don't have 1000 machines in a vacuum. There must be some corporate, government or educational body that purchased all of those boxes, and there's going to be beauracratic and technical interaction with the organization, other networks and systems within it, etc.
If the average machine on someone's desk is usable for about 3 years, 1000 machines means (unless you foolishly just wait for the 3 years to elapse and then panic), that you'll be replacing a machine a bit more than once every work-day to keep desktops up to date in terms of performance and supported hardware (3 years is also a nice period of time to depreciate the cost of the system for the company). Of course, a bit faster than that, since you have to wait at least SOME time (probably a year or year and a half) before you replace the first system.
If you do all of that yourself, then I suggest you consider quitting your job and hiring yourself out to 10 smaller companies and/or getting some sleep, but I suspect that you're thinking only in terms of the "support" work that you need to do.
If you are just saying that you can perform updates and software tuning for 1000 machine, sure so can any monkey with sufficient training, but that's not the whole job that Gartner is considering.
Re:What's a dead virus?
on
HIV Vaccine
·
· Score: 1
More information on Viruses and Vaccination can be found at your local neighborhood Wikipedia.
10-20 years from now, it will take 50% or less of the operations staff that it takes today to manage machines... I can buy that. I look at the history.
In the 50s-60s we had entire departments of large corporations supporting one machine (mainframe).
In the 70s-80s we had entire departments of large corporations supporting several machines (minis).
In the 90s-00s we have entire departments of large corporations supporting hundreds of machines (micros).
So, if we project forward, I certainly see what they're saying, but what happens when I can support 1000 machines at a time on my own the way I do about 1/10th of the support work for those thousand today, but my company needs 10s or even 100s of thousands of machines? Answer: the more things change, the more they stay the same.
Al Gore was just starting to bury the first fiber connections that would become the internet
Re:Insert Internet Inventor Joke Here (Score:5, Informative) by Anonymous Coward on Wednesday March 19, @06:03PM (#5548002)
Gore's words in a CNN interview, as quoted by Wired News, were as follows:
"During my service in the United States Congress, I took the initiative in creating the Internet."
Gore meaning, obvious to anyone who knew the record, was that he did the political work and articulated the public vision that made the Internet possible. No reasonable person could conclude that Gore was claiming to have invented the Internet in any technical sense. The first half of his sentence makes this clear: he is talking about work he did in the context of his service in the Congress. The creation of the Internet was a process that had several phases and took several years, and Gore is claiming the principal credit for the political side of that effort. It is a substantial claim, but an accurate one.
And even more specifically, I think '84 would be a bit later. I think his major push to fund DARPA and NSF were earlier, but I could be wrong.
If you were paying attention to politics concerning the Internet in the 80s (I was), you would have heard Gore's name quite a lot.
Why would you argue with someone's first-hand observations like this. If he said it was hard for him, then it was. Period. You don't get to argue.
When you say "a properly configured..." what you mean to say is, "if you get past the installation, which you found hard..." and yes, you're right. That also changes nothing in his observations.
You're assuming that what you've read about the study is a complete and accurate representation. Ignoring the irony of calling it a "lie", given the subject matter, the simple fact is that unless you're reading the paper written by the researchers themselves, you're likely reading a poor summary of the study.
Re:E uses one config system. It is called vi Use i
on
E17 Available From CVS
·
· Score: 1
What you probably mean is you want a gui
Not at all. vi is not, in any way, a config system. It lacks any kind of consistent data representation across applications. How, for example, do you interpret a config file with high-bits? Is it ISO-Latin-1? Is it UTF-8? Is it Big5? vi is just an editor, and has no way of being sure of the answer to that. Only cross-application standards (which is about 30% of what a desktop environment is for) can solve this problem.
Now E might be a fine window manager for your average desktop environment (assuming we front-end its configuration with something more standard in that DE), but it's just a window manager with delusions of grandeur.
Slashdot Mirror
Are you sure about that?
<tinfoil hat>
What if he's pissed off at being forced to resign, and this is as close as he can come to saying, look, 90% of the machines in the world are running monitoring software that we're able to introduce specifically because security is so damned lax. Fix security, and the CIA would have a much harder time doing its job.
</tinfoil>
One of the funniest replies I've read in this thread claimed that researchers cannot get funding unless they shout "Doomsday is coming". Not even the most imaginary hypothetical example is cited - the report is simply made up by some liberal asshole.
Ah... why would this person be liberal. Aren't liberals the ones screaming doomsday? I'm a moderate liberal who thinks that the left is far too entrenched with the idea that humans are "destroying the planet". I see daily evidence of this in many forms. One: when interviewed by Nova, a researcher who had discoverd massive greenhouse gas emmisions from recent forest fires that burned permafrost material (greenhouse gasses that would not have been released before, due to the change in fire prevention strategies and capabilities) said that this was probably responsible for more greenhouse gasses (CO2 and water vapor) than fossil fuels, but "we don't know how much it takes to tip the balance." Even in the face of cold, hard evidence that the single largest problem was not fossil fuels, this guy's very first reaction was to defend the efforts to curb fossil fuels... not the reaction of someone who is interested in the objective facts.
Two: I have a good friend whose Solar observatory was labeled in congressional funding circles as "an enemy of the planet" (I kid you not), for having published data (not conclusions, data) that supported a non-anthropomorphic warming model and could not secure further funding.
Three: I watch as scientists around the world attempt to discredit any research done with private dollars, without any real concern for the quality of the research itself. In every other field, private research is judged on the basis of its scientific merit or lack thereof, but if you get money from an oil company, car company or heavy industry, and do atmospheric research you can't get published in a reputable journal regardless of the quality of your work.
Put those three together, and come up with a model for me of how research is not skewed toward finding a link between human activity and global warming.
"Reading Slashdot on such topics makes you think the world is really only divided into two kinds of persons - the coporate man/politicians and the crazy gaians"
No, being polar about politics makes you think that.
"Understandable, but this assumes that their PAYING customers know what they are talking about. I use MySQL every day and I've used RDBMSs quite a bit. The feeling that I get from MySQL is that they think a lot of "features" of RDBMSs are just "fluff" (like transactions, stored procedures, foreign key constraints at one time or another)"
They don't, and if you get the chance to listen to them speak about the development of the database at a conference or the like, I highly recommend that you do so.
What they've said is that those features are great, but their existing user base was made up of people who were able to extract value from MySQL specifically because things that required those features were not their core concerns.
MySQL is often used as a front-end cache for other databases, for highly volatile data such as Web content and other information which simply does not require those sorts of features, and which oftne suffers from their implementation.
That's not to say that MySQL should not implement those features, just that they must be taken slowly and carefully, and implemented in such a way that the existing user-base don't have to pay the price for features they didn't ask for.
For example, let me take one that you did not touch on: sub-selects. MySQL doesn't do sub-selects.
Of course, you can emulate them in your code like so:This eliminates a key optimization that sub-selects can perform, but is otherwise identical. That key optimization is to implement the temporary table in such a way that values are returned before either statement is finished executing. This is a major win in terms of the time between request and first answer, but should result in roughly the same execution time overall.
In this case, the small win from the extra optimization and shorter code would have come at the cost of other MySQL development that existing, paying customers wanted more and the changes to select logic that would have allowed for lazy evaluation might have resulted in more checks along the way.
So the feature waited on the back-burner while they worked on things like InnoDB integration and full-text searching. Personally I see the results as a win, and in the end I'll get sub-selects, but not at the cost of features I need.
"Understandable, but this assumes that their PAYING customers know what they are talking about."
Not really. It assumes that implementing what their paying customers want will give them the time and resources to implement what others do. This is a good thing.
" it is *not* an RDBMS and shouldn't be used in place of one."
This is completely untrue. Seriously, unlearn this "goodfact" that's been circulating on Slashdot and other places, as it's simply false.
An RDBMS is a simple category of datastore which implements a style of set logic which relies on relations between like set elements. To put it in crude terms, anything that can implement the most basic SQL select with an inner join is an RDBMS. The logic that has been used to corrupt the term is this: We need more features for modern applications than simple relations, so RDBMS is an insufficient term. Thus, to be an RDBMS, you must perform some arbitrary (and often ill-defined) set of other functions. People have thrown in ACID (a term you don't hear as much any more since modern MySQL engines are ACID), replication, object oriented relationships, and any number of other features which are quite simply not part of the meaning of the term RDBMS.
Say what you mean. MySQL is most certainly a fully-functional RDBMS. That it may not have some of the features of some other RDBMS or even some features that MOST other modern RDBMSs give you is no excuse for arbitrarily changing the defintion of RDBMS.
There's an open source network called Gnutella, developed using open standards, and with many open source implementations.
;-) and lots of great digitial photography (yes, you can find some amazing art that actually involves no skin-tones).
Originally it was only a point-to-point download service negotiated by a peer-to-peer query-flooding system. Times have changed.
Today's Gnutella (go check out gtk-gnutella or any of the other fine clients for details) is more like a fully decentrallized BT. There are still some flaws in the protocol with respect to poisoning, but it's certainly at the point now where I can use it to fetch operating systems, television (I've used it to catch up on shows that my TiVo has missed several times now, shame on me for show-loyalty
"I use MySQL in maybe 60% of my projects, and I have been bitten by a majority of those issues"
Ok, look someone's not reading here... I'll make this one last post, but I really have to stop this thread.
I never said that the items listed were incorrect. They would not work as rhetoric if they were lies. However, the page is designed, not to help (all of the items listed are in the documentation on the MySQL site), but to discourage the use of the database. Any database can be ranted about, but for some unknown reason people choose to rant about MySQL quite a lot.
I think part of this is that PostgreSQL as a project has felt that their time and effort on full ANSI compliance was not rewarded, and so a lot of specifically PostgreSQL folks want to rail on about how MySQL did not make such an attempt, but instead focused on non-standard features like their full-text searching.
That's fine, but let's not confuse those disagreements for any actual problems with either product.
MySQL has some serious problems as an SQL-compliant database (let alone as a *relational* database)
You got those backwards. SQL compliance implies that you are a relational (e.g. set theory applied to a table structure) database. MySQL might be non ANSI SQL 9X compliant, but that does not make it a non-relational database. There are ANSI SQL features which MySQL simply does not support.
and that site simply points them out in an objective manner. In fact he says they aren't bugs or mistakes, but *documented, yet unexpected, behavior*.
It is most certainly not objective, and if you really think that it is, then I don't think we can have this conversation rationally.
Ok, that's it. Thanks guys, but I really do need to go get on with my life. If my comments here have been helpful, great, if not... well, good luck with your work anyway.
One post hardly consitutes carpet bombing.
I was refering to the fact that this link shows up in EVERY MySQL related article, within minutes. That, I call carpet bombing. Use your own terminology as you see fit.
In fact, the only single person I see prolonging this thread (by responing to these so-called "anti-MySQL" people as you like to label them) is you.
One, I've never refered to anyone in this thread as an anti-MySQL person. Not a one. Please, don't re-write my comments.
Two, you are right: I've put too much effort into this, and no one really cares anyway besides folks like you that feel so strongly that over the course of this thread, you actually marked me as a foe.
Well, that's hard to say. Peer reviewed journals from that period do not publish their tables of contents that far back on the Web (Nature, for example, goes back to 1980), so someone would have to go hit a physical library to get you that citation. Perhaps you know of an online index of such documents that I don't... I rarely need documents older than 20 years in my line of work.
HOWEVER, we can be fairly certain that such articles do exist. Phrenology was a very popular theory, and the scientific community would have welcomed papers on the subject.
The real topic was that peer review does not guarantee correctness, only mainstream scientific respect. That's a decent generic baseline, but saying, "this guy on slashdot made a typo, so he can't have seen a flaw in this research," doesn't make a whole lot of sense. The current environment (no pun intended) in the scientific community surrounding global warming is hostile to any information (not even theories, but raw data) which suggests that humans may not play a significant role in said warming. In that environment, you are going to see some junk science published in respected journals. This happens every time there is a major political topic dominating a field.
Global warming may well be human induced, I have no idea, but I can tell you with certainty that we're going to have to dig ourselves out of at least 20 years of heavily biased research before we understand how. One example: scientists discovered that forest fires that burn hotter now due to fire prevention efforts over the last 100 years are able to burn permafrost and release HUGE amounts of carbon dioxide and water vapor (greenhouse gasses). Upon discovering this, the immediate reaction was, "well, this is probably reponsible for more greenhouse emmisions than humans, but we can't know what would tip the balance." There is no objectivity in this field and all data comes with a set of preconceptions that are going to be very hard to break through.
Good points. However, if you want the benefit of such a system without the development time, just use existing Gnutella.
When Gnutella started, it was a simple query-flooding system. It's now much, much more and truly resembles a decentralized BT.
When you search, you get a result. When you request a file, all of the servents that the servent from which you madethe request knows about are told to send you a UDP packet, informing you that they can help with your reuqest. This sort of decentralized swarming allows you to skip right past the centralized server or even a network of them.
The only down-side that I see after discussiong the protocol differences with a friend is this: BT can validate each chunk. Gnutella can validate each chunk (and does), but it's kind of moot since the source of the chunk gave you the checksum in the first place. If someone poisons a chunk you can find that out heuristically over time, but you're going to get to the end of a lot of files and find that they're corrupt before you de-prioritize the poisoner sufficiently.
A smart poisoner could even make it near impossible to be found. This is fixable, but would require protocol extensions for requesting and sending the checksums of all chunks from the original servent that told you about the file. Discovering that such a servent was lying would be much easier over time.
Of course, these problems are hard to avoid in any truly distributed model.
Hmmm... Let's see what would describe the Windows of SQL DBMSes... what would be the criteria... Well, it would have to be
- Proprietary
- Attempt to use closed technologies to lock in users
- Add useless, heavy features that impact routine tasks and stability
- Be very expensive
- Have executive management that seeks to dominate the market through manipulation of sales channels
You're right about one thing... there is an RDBMS that fits your description, but I'm afraid it's not MySQL."The point of the many of the 'gotchas' is that this software behaves erractically: rather than produce an exception during an overflow or conversion error, for example, it just silently modifies the data and returns no warning to the user."
A well documented convinience for applications which always behaves the same way.
Please explain your definition of erratic for me, because mine doesn't seem to match the example you give.
"Two years ago the folks from MySQL stated that transactions, views, subselects, etc weren't needed by 99% of the applications out there"
Wrong. Simply wrong. What they said is that their PAYING CUSTOMERS prioritize their work, and while having everything would be nice, they have to do things in the order that puts the bread on the table.
Transactions were a nice little freebie that fell out of BerkelyDB and InnoDB table types, so fairly little work was required. For something like sub-selects, vies, and other goodies, I'd much rather have all of the features of 4.0 and 4.1 before those. Your milage may vary, and you're welcome to use something else. It's just that you don't need to carpet-bomb every article about products you don't care for.
"Anyhow - please don't blame us victims"
Ok, you really need to take a step back and breath... you're not a "victim" of a database. Please.
"MySQL AB admits their product is a "lite" database, and doesn't outright disparage basic features of databases as unnecessary"
Well, actually they don't. What they've said is that their PAYING customers direct their efforts and those customers prioritize differently than the "MySQL sucks" crowd. It's not terribly surprising which one gets their features implemented first.
What they have said is that they will not implement features that come with (what they percieve to be) major performance hits without finding a way to do so whithout slowing down everything. In many cases, they have found those ways (e.g. by providing InnoDB table types, but still providing MYISAM tables).
Other databases also contribute to this by discovering ways to make the tradeoffs less painful (or even not at all painful). I applaud, for example, much of the work done by Oracle on their unique caching strategies, and while it's not open source, they still contribute valuable data points. PostgreSQL and other open source databases also contribute hugely, just as MySQL contributes to them (would PostgreSQL have moved beyond its difficult-to-admin stage if MySQL were not so damned simple to install and set up?)
We're the OPEN SOURCE COMMUNITY, and we really should start acting like it.
Are we so mainstream in the open source world now that we have bought into the utterly bankrupt idea that every story has two (and only two) sides? Someone who rails against PostgreSQL and MySQL must have a balanced view, and deserves to be linked to from anything that discusses either? Eh, how's that?
Please, it's just a rant. He can rant about all the products he wants, I just don't see why every mention of MySQL has to link to this diatribe.
people think twice about using MySQL
What is this... I'm really freaking confused. MySQL is used by tens if not hundreds of thousands of people around the world. It has always done what I want it to do faithfully. I'm a Sybase weenie with significant Oracle and Ingres experience. I've been working with and writing software for 15 years and I honestly do not understand what is so damned scary about MySQL. Send query, get answer. Life good. Carry on.
If it doesn't do what you want, fine don't use it, but please stop acting as if it's going to poison the water supply.
FWIW, I've used both PostgreSQL and MySQL.
This is a bit like disclaiming your political views by saying you've voted constitution party AND green party.
A lot of folks think that MySQL is a good idea
It's software that does what you tell it to do. Thus, it is a good idea. You think software product X is beter, meets some criteria that you find compelling, can do things that you want... great, that's fine, but like I said, that opinion doesn't need to be spattered all over every occurance of software product Y on slashdot.
"How is this a rant? Are you saying the problems they list don't exist?"
Not at all. What I'm saying is that it's a page that was meant to discredit MySQL by putting together a sort of reverse-FAQ. Instead of saying, "people run into X, well here's what they do when they run into that," it's just a diatribe about how lacking someone thinks MySQL is. That, as they say, ain't news, and it most CERTAINLY does not deserve a fresh link from every MySQL article to hit Slashdot.
I guess it's nice, but a decentralized BT is called Gnutella. Check out the current state of the Gnutella protocol. Swarming, UDP chunk notification, etc. All of the BT features you expect are there, and if you're downloading sufficiently popular data, it's screamingly fast (just like BT).
What I'm waiting for is a non-proprietary client that recognizes torrents over Gnutella and can fetch them via the BT protocol. That would eliminate that rare need I have to fire up a BT client.
I use gtk-gnutella, which is quite nice, but I hear good things about other clients these days too.
That is NOT a list of "common gotchas" (the misleading name of the page aside), it's an anti-MySQL rant that has lived on far too long.
Please, can we for once have a post about a piece of software on Slashdot without the pro-X or anti-Y folks flocking to it to bash or praise it? Can we just for once talk about the damn book?
Ah ok, this is a misunderstanding then. Most of what I would refer to as "support" is either being defered (after all your time is expensive, so your clients don't contact you unless they NEED you) or isn't being done at all (e.g. your comment, "who wants to pay to have every possible hole patched up to complete").
Being a consultant makes a world of difference. Sure, you can sell support for 1000 machines, but that's very different from being the IT staff in a company with 1000 machines. VERY different.
Still, offering support for 1000 end-users is no small task, and my hat's off to you.
I can say for certain, that I can maintain 1000 desktop systems by myself, and about 50 servers
I'm sorry, I just don't buy that. If you mean that you can support one aspect of running those machines, sure, but you're doing it all? How do you have enough time to dedicate to security updates (and response to any security problems), external complaints, administrative requests, new users and hardware, MTBF replacements, user questions, new software install requests, expansion of physical space and networks, physical upgrades, and the inevitable development of tools and small software? Over 1000 systems? On top of that, you don't have 1000 machines in a vacuum. There must be some corporate, government or educational body that purchased all of those boxes, and there's going to be beauracratic and technical interaction with the organization, other networks and systems within it, etc.
If the average machine on someone's desk is usable for about 3 years, 1000 machines means (unless you foolishly just wait for the 3 years to elapse and then panic), that you'll be replacing a machine a bit more than once every work-day to keep desktops up to date in terms of performance and supported hardware (3 years is also a nice period of time to depreciate the cost of the system for the company). Of course, a bit faster than that, since you have to wait at least SOME time (probably a year or year and a half) before you replace the first system.
If you do all of that yourself, then I suggest you consider quitting your job and hiring yourself out to 10 smaller companies and/or getting some sleep, but I suspect that you're thinking only in terms of the "support" work that you need to do.
If you are just saying that you can perform updates and software tuning for 1000 machine, sure so can any monkey with sufficient training, but that's not the whole job that Gartner is considering.
More information on Viruses and Vaccination can be found at your local neighborhood Wikipedia.
10-20 years from now, it will take 50% or less of the operations staff that it takes today to manage machines... I can buy that. I look at the history.
In the 50s-60s we had entire departments of large corporations supporting one machine (mainframe).
In the 70s-80s we had entire departments of large corporations supporting several machines (minis).
In the 90s-00s we have entire departments of large corporations supporting hundreds of machines (micros).
So, if we project forward, I certainly see what they're saying, but what happens when I can support 1000 machines at a time on my own the way I do about 1/10th of the support work for those thousand today, but my company needs 10s or even 100s of thousands of machines? Answer: the more things change, the more they stay the same.
And even more specifically, I think '84 would be a bit later. I think his major push to fund DARPA and NSF were earlier, but I could be wrong.
If you were paying attention to politics concerning the Internet in the 80s (I was), you would have heard Gore's name quite a lot.
Why would you argue with someone's first-hand observations like this. If he said it was hard for him, then it was. Period. You don't get to argue.
When you say "a properly configured..." what you mean to say is, "if you get past the installation, which you found hard..." and yes, you're right. That also changes nothing in his observations.
You're assuming that what you've read about the study is a complete and accurate representation. Ignoring the irony of calling it a "lie", given the subject matter, the simple fact is that unless you're reading the paper written by the researchers themselves, you're likely reading a poor summary of the study.
What you probably mean is you want a gui
Not at all. vi is not, in any way, a config system. It lacks any kind of consistent data representation across applications. How, for example, do you interpret a config file with high-bits? Is it ISO-Latin-1? Is it UTF-8? Is it Big5? vi is just an editor, and has no way of being sure of the answer to that. Only cross-application standards (which is about 30% of what a desktop environment is for) can solve this problem.
Now E might be a fine window manager for your average desktop environment (assuming we front-end its configuration with something more standard in that DE), but it's just a window manager with delusions of grandeur.