If the government is the client then by definition the public is the client, since the government is only acting on behalf of the public.
No, the government is the client, by definition, in a project that was supposed to be for the benefit for the public. That does not mean that the public is the client, not even indirectly. You don't become a client by proxy just because you are a taxpayer, that is not how projects work.
And yes, I shudder at the thought of uncounted number of backseat drivers trying give feedback. Letting more people look into something isn't trivial and it costs. What government projects should we allow full and instantaneous insight (aka absolute transparency) into and to what cost? What is an acceptable overhead cost for adding absolute transparency to every project? How much do you want to pay in extra tax just to humor this 'need' of yours? I rather prefer to pay less taxes and let the politicians heads roll if they get out of line, and that the money I pay in taxes go to healhcare, education etc. If we add overhead costs, we should make sure that we save money in the long run. Anything else is much more irresponsible than a failed project.
I do think we need transparency, to a certain degree. Absolute transparency only exists in utopia or where there is no regard for cost. Even if we have transparency things can slip through. Every project is a gamble.
If the UK taxpayers see that there has been something wrong going on, make 'em pay. Use your democratic vote.
Just because taxpayers somehow pay for something doesn't mean that every taxpayer should be able to get full real-time insight into or control of it. I shudder at the thought of the extra cost when gazillions of wannabe software developers consider themselves to be the clients of a government project and mess it up.
I didn't see the following tweets until Major Nelson put out the apology, and I was rather horrified by the way Adam Orth expresses himself to a potential customer. Still, I am not sure kicking someone is the right way to go, but I do think they need to give at least the management some media training and make sure that everyone is aware of a company media policy. So many people are ignorant of how the internet ecosystem works and how things spread.
Personally, I refrain myself from publicly commenting on matters regarding the organization where I work. We have people whose job is to take care of these matters. When I see something I can tell them, say what I think and let them decide the correct course of action. I am entitled to my opinion, but that doesn't mean that I need to express it at all times. I know that my word might be taken for the official position and that might not be true, anyway I am not paid to comment on my employers decisions.
One of my first lecturers at the university said that the traditional lectures are the worlds worst Xerox-machine. One guy standing by the blackboard and 200+ students copying everything verbatim.
In his classes we got all the slides he used at the beginning of the course, so we had no need to copy everything down. Most of the time the topics were covered in the textbook, but if he chose to cover material that wasn't in the textbook, it was in the hand-outs. The notes I took in those classes were mainly writing down some extra words on a slide printout if I thought something could use extra clarification.
I don't see *why* the lectures should give you something that is *only* available to those attending the lectures. The lectures might give a different angle on the same subject, but essentially everything you need should be in the material you are already given. (Why on earth should the professor withhold information from the students? Isn't s/he interested in them learning?) This way, lectures can focus on trying to explain that and allow the students to ask questions for clarification of that material. Since we were given hand-outs and slides for all the classes for the whole semester, I could easily browse through next class in advance to be more prepared and ask more relevant questions in class, improving my learning even further. Each lecture was followed up by a smaller class with a TA, digging deeper into the subject of the day and possibly preparing for a lab session.
Attendance in a class should not be a goal in itself, learning is the goal, and the classes should facilitate that.
This method of course requires the lecturers to prepare the classes well in advance. Not all lecturers seem to have the ability and focus to do that. I wonder why they require any more of the students then.
This class was one of the best I attended to at my university, and the following four years I returned as a TA in that class, teaching a total of six classes of EE and CS students. For me it set a standard of teaching which I still today find exemplary. (The course in question was modeled on MIT 6.001 but had gone through a number of revisions throughout the years.)
This to me is such a fail, as most web devs need to be sure of the versions they are compatible with...
No, the "fail" is in that very chain of thought. Those web devs should not call themselves web devs since they do not understand the fundamental differences between the old media they used to work with and the new media, having to resort to web browser version to achieve what they foolishly are striving for.
At first I was not too keen on version number inflation, but thinking about it I couldn't care less. Actually, I find it good if it rids the world of people targeting web browser versions when they develop for the web. Target standards, not web browsers.
The only problem as I see is the plugins. That could be handled if Mozilla decided to create a stable API for plugin development and have version numbers on that API instead. This could even create a more stable browser with less unpredictability when multiple plugins are used. Another way, although more anarchistic, is to create a crowd sourced database of version compatibility between browsers and plugins, not having installers contain that information, but rather let us (the users) try it out and report.
Yes, and the OP wrote "[...]technology being crippled for no good reason" and "[...]IP nonsense!".
No one is questioning whether or not anything is against the spec, what was questioned is why you cripple technology (through a spec or otherwise) without a good reason, and as of yet no good reason has been produced. Thus: IP nonsense. I don't think anyone is questioning Phillips move as anything other than "good reason" (with the possible exception of the DRM advocates).
If HDMI LLC can give some good reason, they might sway my opinion. If I dare guess, the only reason I think they can give is that the HDMI spec is supposed to ensure that unauthorized copies cannot be made, and if you are able to produce HDMI-to-anything cables you could connect your HDMI capable output device to something that can record the information. But that I do not consider to be a good enough reason. I am not interested in making copies but I do want to be able to (or at least have the possibility to) connect my legacy HDMI products to newer products that might not have HDMI, in the future.
I don't understand what you mean with "both ways". Phillips is not crippling technology with their stance on the CD, in fact, they are doing the opposite by telling manufacturers that DRM is not a part of the CD-specification and might prevent consumers from playing those discs. Thus they are not allowed to be called CDs. The DRM is the crippling part, not the fact that the manufacturers that insist on having DRM on their discs can't call them CDs.
It might not have cost you a dime, but hasn't it cost you time? For me time is more valuable than money. Money I can get more of, time I can never get back. You might think it is worth your time to be on Facebook, good for you. Still, it is not free, you gave up something to use Facebook.
If your personal data is of no value, then you might say that it is free in that sense, otherwise Facebook is very costly.
So, "free" is a very loose term, and money is a very minor factor in determining if something is free.
"Serious restrictions on freedom of expression"? It would be nice to know WHAT part of Europe you are referring to. IMHO, the country I live in stille are more free when it comes to expression than the US. The copyright restrictions pushed on us now are backed by interests over in the great media publishing giant in the west, and the limits on competition as you call it is to *protect* competition from being destroyed by a monopoly-like situation.
I seriously doubt that software patents are pushed onto the european member countries by the EU unless there were a precedent in the US backed by very strong and determined economic interests.
Keep calling? What are you mumbling about? In the GP there is ONE mention of "Abserger's" and TWO mentions of "Asberger's".
I seriously doubt that this counts as "keep calling". Being a wise-ass about a typo is just stupid. Then again, what can one expect from an AC at slashdot?
No. They might both be of the same magnitude of order, as in around 1GHz, but there definitely is a real noticable difference between the old 1GHz processor i bought around a decade ago, the 1GHz processor in my iPad and the 1.42GHz on my relatively old desktop.
Clock speed is not comparable when you have different architectures and the surrounding hardware differs greatly.
How many developers contact Apple before starting to design an app, while just in the idea phase, just to check if it might get approval? I doubt that there are that many.
Since I am not an InfoWorld subscriber I could not read the report by Andrew Borg of the Aberdeen Group that Galen Gruman wrote about (nice plug about your own article BTW). Thus I have a hard time to see what Borg really meant and what got lost in the filtering of TFA.
But of course we will have to think of wired and wireless networks as two separate entities. Not that we cannot think about them at the same time and how they should work together, but because of their different characteristics.
For an end-user the experience should be roughly the same, but from an engineering point-of-view, you have to take all factors into account when designing your network. The limitations, security concerns, cost, etc of each medium is important to acknowledge.
So even if I might agree on that we shouldn't view wi-fi as the "neglected stepchild", we cannot dismiss the differences. Doing so would be plain stupid.
Since I am a bit tired I should probably clarify my previous post. Most posters seem to have got stuck on the claim that 10 lines could deobfuscate most addresses. It doesn't matter if it is 10, 100, 1000 or more. Deobfuscation either works (since most users use simple obfuscation techniques) or don't work. In the cases where it don't work people would probably have to think twice to deobfuscate manually and the probability of sending the mail to a non-existant adress is higher. You will be harder to reach. So yes we have decreased the utility and the spam continues.
Use obfuscation if you think it works for you, I don't think it is necessarily a bad thing unless used over-excessively. Personally, I rather rely on filtering techniques and make it easy for people wanting to mail me to do so. This is the reason I have an email address, otherwise I would force people to use regular mail to get in touch with me.
Let this be a lesson to you kids: sleep before you post.
Of course it will be harder to write such a filter (Please note that the spammer wouldn't care about the filter max beeing 10 lines with 80 chars each and no modules, if 2000 lines do the job the script will be 2000 lines.)
But for the non-german speaking person visiting your website it would perhaps be impossible to get to mail you, and besides you missed the running point in the original posting:
So if we're not helping hold back the flood of spam, why are we decreasing the utility of the web by eliminating mailto tags and forcing users to hand-correct the addresses in their mail clients?
Or perhaps you have already answered the question; Yes we have definitely decreased the utility of the web by eliminating mailto tags and forced the users to hand-correct (deobscure) email addresses.
I can agree with you that many XXX fanatics slams YYY fanatics (in this case XXX=PostgreSQL and YYY=MySQL). The opposite is very much true also. I mean, you brought up how other people in this thread avoided the "8k row limit" and "vacuum locking", two historical issues you just brought up to illustrate how badly people behaved in the past! That would be as if I tried to talk about pre-InnoDB/Berkley DB MySQL. Hardly a constructive discussion?
Many people could really do with transactions (and triggers and foreign keys), just to make sure that their data is intact, especially since they don't need that extra speed non-transactional tables buys them. But my standpoint is clear in this case, if you choose away transactions you should know what you are doing, many people do but even more don't. There are many different kinds of solutions to storing data, but when it comes to databases my first priority is data integrity. Obviously others have other priorities. Consider that there are many ways of storing data, filesystems, LDAP, RDBMS, in-memory, etc. Use the solution that suits you.
And come on. You too would reply if you were being called things.
A last thing: do try PostgreSQL out, it doesn't take much time or effort. A friend of mine with only very limited MySQL experience got started with PostgreSQL in less than an hour, someone with your experience should be able to do it in notime. And it is able to run on very crappy hardware, I don't think it requires much more than MySQL.
I have re-read your post and I find it very disturbing that you run around calling other people, including me, arrogand and elitist, just because our definition of a RDBMS does not happen match your. I am inclined to agree with those you call arrogant elitists that a database should always use transactions to protect your data. After all, transactions are there to protect your data, hence database. Note that transactions alone don't do magic, other features as triggers and foreign keys are needed too.
So sorry, your post don't convice me at all either, just confirms that our standpoint are completely different. What disturbs me is that you have to resort to labeling people as "arrogant" and "elitists" just because their viewpoint differ from yours. That kind of behaviour makes me sick. I feel sorry for you.
And to all other out there, check what your application need. If it is raw speed and no data safety, MySQL/MyISAM is great. If you have to have transactions MySQL/InnoDB might work, especially if you want to combine both transactions and non-transaction tables. Otherwise, check out PostgreSQL for example (there are a few others like Firebird but that I have no personal experience with so get recommendations from someone else). It is not as bad as people said. Easy to set up, get running and the docs aren't bad. I have learned from them, I think you can too.
I am sorry if I wasn't clear in my previous post. Of course you can, but most people don't normalize their tables. In fact many use one, the table, to store everything. Then they have little if any use for more "advanced" features. Most people don't need referential integrity simply because everything is stored in a huge table without references. Then they never have to use subqueries to find the complex key from other tables when inserting new data. They do not worry with handling several tables, so often no transaction need. For them MySQL works fine and probably faster than any other database since they don't bother with transactions. In fact, they use MySQL more as a data storage not a relational database. It's ok, it works for their application. My objection was when people say "most people don't need..." without looking at their application. "Most people" don't need a RDBMS. This doesn't somehow make MySQL automatically and univerally superior. In some applications, yes. As a full fledged RDBMS many seem to object, I am inclined to agree with them. Misunderstand me right; it is even better to use MySQL even if in a bad way than for every programmer to implement his own datastorage from scratch. The development cycle becomes much more rapid when you just have to open a database connection, throw in a few INSERTs and SELECTs and have abstracted your project to a higher level than to dig around in your filesystem and parse homebrewed fileformats.
No RDBMS has a monopoly on bad design. None has a monopoly on good design, either.
Of course not. I have seen one of the ugliest database designs ever done on an Oracle 8i-server, then to be ported over to tha PostgreSQL-server I once set up at my old work. I almost wept. But IMHO a system can "encourage" or "discourage" bad design. I think that allowing users to choose away transactions is bad, perhaps you are of a different opinion, ok.
MySQL has supported transactions and foreign key constraints for ages, using InnoDB tables.
Yes, but it is not used for every table, only where the designer has agreed to use InnoDB-tables. This is not something I really like with a system that is to be a RDBMS. Today you can claim that this is good because it allows users to choose between raw speed and data integrity. But this hasn't always been like this and to me it seems like the lack of transactions in orignal MySQL was a poor design decision that has been "fixed" through the design decision to allow multiple types of tables and the later InnoDB. Two ways of looking at the same issue.
MySQL works very well for a very large number of people, thanks very much.
This is because a very large number of people don't really need a database to store their data. They could just use a flat file, but MySQL is much more convenient to use, it allows for higher level programming. Good, use it. That alone does not make it more of a database system. Still, it is very useful and to the myriad of developers out there that don't really need a full fledged database system and knows MySQL like the back of their hand, this is a recommendation from me: use it! Keep an open mind though and try to learn other systems also.
Calling someone who challenges your viewpoint (that transactions MUST be used everywhere and if you don't do subqueries then you're somehow a charletan and not a "real" developer of "serious" databases) is just more of the same drivel I've been reading all this time...
Personally I don't run a critical system but I find it very comforting to know that transactions allow my tables to be updated without any raceconditions or unwanted side effects. Because I model my databases and normalizes them to 3NF or BCNF I always have several tables and thus transactions and subqueries are essential to me!
Many people out there don't know what normalization means and have one big table for everything, effectively duplicatind data, but for their use it works. I don't look down on them, but if they try convincing me that transactions and subqueries are bogus and can be solved within the application layer I know that they really have little clue of what they are speaking about.
You conveniently ignored the 8k row limits and the VACUUM blocking yet again, because it doesn't fit with your cozy world view.
Yes, 8k row limits might be a pain in the butt. But they are history now, and has been so for quite a while. Personally I have never had a problem with it for my projects, although I agree that that was the one of the big issues I had with PostgreSQL when I initially had to choose a system. But since MySQL lacked a few features i really needed, MySQL was never the choise for me.
Again, VACUUM blocking I have never had a problem with, but YMMV. Also, this is history, so why go on about it? Lets talk about todays limitations in the systems. If someone say something incorrect about todays MySQL, feel free to correct them.
What problems do you have with todays PostgreSQL? What could MySQL do for you that PostgreSQL couldn't?
If Oracle or MS SQL were free I would probably use them, but they aren't. Now i use PostgreSQL. I have yet to be shown where MySQL could be a better replacement for me, but if you can convice me, i'll switch!
I remember reading this article and thought of the similarity between the thread arcs of IBM and conversation clues of Microsoft. I'd sure like to see my Thunderbird support them!
Interesting though, that I can't recall seeing such a discussion about how IBM "reinvented" threading when the previous Slashdot article was posted. Or is it just that the average Slashdot ranter is unable to read or understand anything after having seen the character sequence M I C R O S O F T ?
Please. Everybody. Read the article. A bit too brief but nevertheless interesting. Thanks glawrie for the reference.
Slashdot Mirror
If the government is the client then by definition the public is the client, since the government is only acting on behalf of the public.
No, the government is the client, by definition, in a project that was supposed to be for the benefit for the public. That does not mean that the public is the client, not even indirectly. You don't become a client by proxy just because you are a taxpayer, that is not how projects work.
And yes, I shudder at the thought of uncounted number of backseat drivers trying give feedback. Letting more people look into something isn't trivial and it costs. What government projects should we allow full and instantaneous insight (aka absolute transparency) into and to what cost? What is an acceptable overhead cost for adding absolute transparency to every project? How much do you want to pay in extra tax just to humor this 'need' of yours? I rather prefer to pay less taxes and let the politicians heads roll if they get out of line, and that the money I pay in taxes go to healhcare, education etc. If we add overhead costs, we should make sure that we save money in the long run. Anything else is much more irresponsible than a failed project.
I do think we need transparency, to a certain degree. Absolute transparency only exists in utopia or where there is no regard for cost. Even if we have transparency things can slip through. Every project is a gamble.
If the UK taxpayers see that there has been something wrong going on, make 'em pay. Use your democratic vote.
Someone please mod parent up.
Just because taxpayers somehow pay for something doesn't mean that every taxpayer should be able to get full real-time insight into or control of it. I shudder at the thought of the extra cost when gazillions of wannabe software developers consider themselves to be the clients of a government project and mess it up.
I didn't see the following tweets until Major Nelson put out the apology, and I was rather horrified by the way Adam Orth expresses himself to a potential customer. Still, I am not sure kicking someone is the right way to go, but I do think they need to give at least the management some media training and make sure that everyone is aware of a company media policy. So many people are ignorant of how the internet ecosystem works and how things spread.
Personally, I refrain myself from publicly commenting on matters regarding the organization where I work. We have people whose job is to take care of these matters. When I see something I can tell them, say what I think and let them decide the correct course of action. I am entitled to my opinion, but that doesn't mean that I need to express it at all times. I know that my word might be taken for the official position and that might not be true, anyway I am not paid to comment on my employers decisions.
Yesterday I summed up some of my thoughts in the matter: http://mzomborszki.wordpress.com/2013/04/05/how-to-be-an-insensitive-clod/.
One of my first lecturers at the university said that the traditional lectures are the worlds worst Xerox-machine. One guy standing by the blackboard and 200+ students copying everything verbatim.
In his classes we got all the slides he used at the beginning of the course, so we had no need to copy everything down. Most of the time the topics were covered in the textbook, but if he chose to cover material that wasn't in the textbook, it was in the hand-outs. The notes I took in those classes were mainly writing down some extra words on a slide printout if I thought something could use extra clarification.
I don't see *why* the lectures should give you something that is *only* available to those attending the lectures. The lectures might give a different angle on the same subject, but essentially everything you need should be in the material you are already given. (Why on earth should the professor withhold information from the students? Isn't s/he interested in them learning?) This way, lectures can focus on trying to explain that and allow the students to ask questions for clarification of that material. Since we were given hand-outs and slides for all the classes for the whole semester, I could easily browse through next class in advance to be more prepared and ask more relevant questions in class, improving my learning even further. Each lecture was followed up by a smaller class with a TA, digging deeper into the subject of the day and possibly preparing for a lab session.
Attendance in a class should not be a goal in itself, learning is the goal, and the classes should facilitate that.
This method of course requires the lecturers to prepare the classes well in advance. Not all lecturers seem to have the ability and focus to do that. I wonder why they require any more of the students then.
This class was one of the best I attended to at my university, and the following four years I returned as a TA in that class, teaching a total of six classes of EE and CS students. For me it set a standard of teaching which I still today find exemplary. (The course in question was modeled on MIT 6.001 but had gone through a number of revisions throughout the years.)
This to me is such a fail, as most web devs need to be sure of the versions they are compatible with...
No, the "fail" is in that very chain of thought. Those web devs should not call themselves web devs since they do not understand the fundamental differences between the old media they used to work with and the new media, having to resort to web browser version to achieve what they foolishly are striving for.
At first I was not too keen on version number inflation, but thinking about it I couldn't care less. Actually, I find it good if it rids the world of people targeting web browser versions when they develop for the web. Target standards, not web browsers.
The only problem as I see is the plugins. That could be handled if Mozilla decided to create a stable API for plugin development and have version numbers on that API instead. This could even create a more stable browser with less unpredictability when multiple plugins are used. Another way, although more anarchistic, is to create a crowd sourced database of version compatibility between browsers and plugins, not having installers contain that information, but rather let us (the users) try it out and report.
Yes, and the OP wrote "[...]technology being crippled for no good reason" and "[...]IP nonsense!".
No one is questioning whether or not anything is against the spec, what was questioned is why you cripple technology (through a spec or otherwise) without a good reason, and as of yet no good reason has been produced. Thus: IP nonsense. I don't think anyone is questioning Phillips move as anything other than "good reason" (with the possible exception of the DRM advocates).
If HDMI LLC can give some good reason, they might sway my opinion. If I dare guess, the only reason I think they can give is that the HDMI spec is supposed to ensure that unauthorized copies cannot be made, and if you are able to produce HDMI-to-anything cables you could connect your HDMI capable output device to something that can record the information. But that I do not consider to be a good enough reason. I am not interested in making copies but I do want to be able to (or at least have the possibility to) connect my legacy HDMI products to newer products that might not have HDMI, in the future.
I don't understand what you mean with "both ways". Phillips is not crippling technology with their stance on the CD, in fact, they are doing the opposite by telling manufacturers that DRM is not a part of the CD-specification and might prevent consumers from playing those discs. Thus they are not allowed to be called CDs. The DRM is the crippling part, not the fact that the manufacturers that insist on having DRM on their discs can't call them CDs.
Who wrote such a narrow-minded license and for what purpose? I would like how they thought this would benefit end-users.
It smells like greed, incompetence and arrogance.
Why do you only measure "free" in terms of money?
It might not have cost you a dime, but hasn't it cost you time? For me time is more valuable than money. Money I can get more of, time I can never get back. You might think it is worth your time to be on Facebook, good for you. Still, it is not free, you gave up something to use Facebook.
If your personal data is of no value, then you might say that it is free in that sense, otherwise Facebook is very costly.
So, "free" is a very loose term, and money is a very minor factor in determining if something is free.
"Serious restrictions on freedom of expression"? It would be nice to know WHAT part of Europe you are referring to. IMHO, the country I live in stille are more free when it comes to expression than the US. The copyright restrictions pushed on us now are backed by interests over in the great media publishing giant in the west, and the limits on competition as you call it is to *protect* competition from being destroyed by a monopoly-like situation.
I seriously doubt that software patents are pushed onto the european member countries by the EU unless there were a precedent in the US backed by very strong and determined economic interests.
D'oh! :-)
Keep calling? What are you mumbling about? In the GP there is ONE mention of "Abserger's" and TWO mentions of "Asberger's".
I seriously doubt that this counts as "keep calling". Being a wise-ass about a typo is just stupid. Then again, what can one expect from an AC at slashdot?
No. They might both be of the same magnitude of order, as in around 1GHz, but there definitely is a real noticable difference between the old 1GHz processor i bought around a decade ago, the 1GHz processor in my iPad and the 1.42GHz on my relatively old desktop.
Clock speed is not comparable when you have different architectures and the surrounding hardware differs greatly.
How many developers contact Apple before starting to design an app, while just in the idea phase, just to check if it might get approval? I doubt that there are that many.
Since I am not an InfoWorld subscriber I could not read the report by Andrew Borg of the Aberdeen Group that Galen Gruman wrote about (nice plug about your own article BTW). Thus I have a hard time to see what Borg really meant and what got lost in the filtering of TFA.
But of course we will have to think of wired and wireless networks as two separate entities. Not that we cannot think about them at the same time and how they should work together, but because of their different characteristics.
For an end-user the experience should be roughly the same, but from an engineering point-of-view, you have to take all factors into account when designing your network. The limitations, security concerns, cost, etc of each medium is important to acknowledge.
So even if I might agree on that we shouldn't view wi-fi as the "neglected stepchild", we cannot dismiss the differences. Doing so would be plain stupid.
Since I am a bit tired I should probably clarify my previous post. Most posters seem to have got stuck on the claim that 10 lines could deobfuscate most addresses. It doesn't matter if it is 10, 100, 1000 or more. Deobfuscation either works (since most users use simple obfuscation techniques) or don't work. In the cases where it don't work people would probably have to think twice to deobfuscate manually and the probability of sending the mail to a non-existant adress is higher. You will be harder to reach. So yes we have decreased the utility and the spam continues.
Use obfuscation if you think it works for you, I don't think it is necessarily a bad thing unless used over-excessively. Personally, I rather rely on filtering techniques and make it easy for people wanting to mail me to do so. This is the reason I have an email address, otherwise I would force people to use regular mail to get in touch with me.
Let this be a lesson to you kids: sleep before you post.
Of course it will be harder to write such a filter (Please note that the spammer wouldn't care about the filter max beeing 10 lines with 80 chars each and no modules, if 2000 lines do the job the script will be 2000 lines.)
But for the non-german speaking person visiting your website it would perhaps be impossible to get to mail you, and besides you missed the running point in the original posting:
Or perhaps you have already answered the question; Yes we have definitely decreased the utility of the web by eliminating mailto tags and forced the users to hand-correct (deobscure) email addresses.
I can agree with you that many XXX fanatics slams YYY fanatics (in this case XXX=PostgreSQL and YYY=MySQL). The opposite is very much true also. I mean, you brought up how other people in this thread avoided the "8k row limit" and "vacuum locking", two historical issues you just brought up to illustrate how badly people behaved in the past! That would be as if I tried to talk about pre-InnoDB/Berkley DB MySQL. Hardly a constructive discussion?
Many people could really do with transactions (and triggers and foreign keys), just to make sure that their data is intact, especially since they don't need that extra speed non-transactional tables buys them. But my standpoint is clear in this case, if you choose away transactions you should know what you are doing, many people do but even more don't. There are many different kinds of solutions to storing data, but when it comes to databases my first priority is data integrity. Obviously others have other priorities. Consider that there are many ways of storing data, filesystems, LDAP, RDBMS, in-memory, etc. Use the solution that suits you.
And come on. You too would reply if you were being called things.
A last thing: do try PostgreSQL out, it doesn't take much time or effort. A friend of mine with only very limited MySQL experience got started with PostgreSQL in less than an hour, someone with your experience should be able to do it in notime. And it is able to run on very crappy hardware, I don't think it requires much more than MySQL.
I have re-read your post and I find it very disturbing that you run around calling other people, including me, arrogand and elitist, just because our definition of a RDBMS does not happen match your. I am inclined to agree with those you call arrogant elitists that a database should always use transactions to protect your data. After all, transactions are there to protect your data, hence database. Note that transactions alone don't do magic, other features as triggers and foreign keys are needed too.
So sorry, your post don't convice me at all either, just confirms that our standpoint are completely different. What disturbs me is that you have to resort to labeling people as "arrogant" and "elitists" just because their viewpoint differ from yours. That kind of behaviour makes me sick. I feel sorry for you.
And to all other out there, check what your application need. If it is raw speed and no data safety, MySQL/MyISAM is great. If you have to have transactions MySQL/InnoDB might work, especially if you want to combine both transactions and non-transaction tables. Otherwise, check out PostgreSQL for example (there are a few others like Firebird but that I have no personal experience with so get recommendations from someone else). It is not as bad as people said. Easy to set up, get running and the docs aren't bad. I have learned from them, I think you can too.
Yes, many SELECTs can be solved using joins, INSERTS, DELETES and UPDATES are more troublesome.
And you can't normalise MySQL tables?
I am sorry if I wasn't clear in my previous post. Of course you can, but most people don't normalize their tables. In fact many use one, the table, to store everything. Then they have little if any use for more "advanced" features. Most people don't need referential integrity simply because everything is stored in a huge table without references. Then they never have to use subqueries to find the complex key from other tables when inserting new data. They do not worry with handling several tables, so often no transaction need. For them MySQL works fine and probably faster than any other database since they don't bother with transactions. In fact, they use MySQL more as a data storage not a relational database. It's ok, it works for their application. My objection was when people say "most people don't need..." without looking at their application. "Most people" don't need a RDBMS. This doesn't somehow make MySQL automatically and univerally superior. In some applications, yes. As a full fledged RDBMS many seem to object, I am inclined to agree with them. Misunderstand me right; it is even better to use MySQL even if in a bad way than for every programmer to implement his own datastorage from scratch. The development cycle becomes much more rapid when you just have to open a database connection, throw in a few INSERTs and SELECTs and have abstracted your project to a higher level than to dig around in your filesystem and parse homebrewed fileformats.
No RDBMS has a monopoly on bad design. None has a monopoly on good design, either.
Of course not. I have seen one of the ugliest database designs ever done on an Oracle 8i-server, then to be ported over to tha PostgreSQL-server I once set up at my old work. I almost wept. But IMHO a system can "encourage" or "discourage" bad design. I think that allowing users to choose away transactions is bad, perhaps you are of a different opinion, ok.
MySQL has supported transactions and foreign key constraints for ages, using InnoDB tables.
Yes, but it is not used for every table, only where the designer has agreed to use InnoDB-tables. This is not something I really like with a system that is to be a RDBMS. Today you can claim that this is good because it allows users to choose between raw speed and data integrity. But this hasn't always been like this and to me it seems like the lack of transactions in orignal MySQL was a poor design decision that has been "fixed" through the design decision to allow multiple types of tables and the later InnoDB. Two ways of looking at the same issue.
MySQL works very well for a very large number of people, thanks very much.
This is because a very large number of people don't really need a database to store their data. They could just use a flat file, but MySQL is much more convenient to use, it allows for higher level programming. Good, use it. That alone does not make it more of a database system. Still, it is very useful and to the myriad of developers out there that don't really need a full fledged database system and knows MySQL like the back of their hand, this is a recommendation from me: use it! Keep an open mind though and try to learn other systems also.
Calling someone who challenges your viewpoint (that transactions MUST be used everywhere and if you don't do subqueries then you're somehow a charletan and not a "real" developer of "serious" databases) is just more of the same drivel I've been reading all this time...
Personally I don't run a critical system but I find it very comforting to know that transactions allow my tables to be updated without any raceconditions or unwanted side effects. Because I model my databases and normalizes them to 3NF or BCNF I always have several tables and thus transactions and subqueries are essential to me!
Many people out there don't know what normalization means and have one big table for everything, effectively duplicatind data, but for their use it works. I don't look down on them, but if they try convincing me that transactions and subqueries are bogus and can be solved within the application layer I know that they really have little clue of what they are speaking about.
You conveniently ignored the 8k row limits and the VACUUM blocking yet again, because it doesn't fit with your cozy world view.
Yes, 8k row limits might be a pain in the butt. But they are history now, and has been so for quite a while. Personally I have never had a problem with it for my projects, although I agree that that was the one of the big issues I had with PostgreSQL when I initially had to choose a system. But since MySQL lacked a few features i really needed, MySQL was never the choise for me.
Again, VACUUM blocking I have never had a problem with, but YMMV. Also, this is history, so why go on about it? Lets talk about todays limitations in the systems. If someone say something incorrect about todays MySQL, feel free to correct them.
What problems do you have with todays PostgreSQL? What could MySQL do for you that PostgreSQL couldn't?
If Oracle or MS SQL were free I would probably use them, but they aren't. Now i use PostgreSQL. I have yet to be shown where MySQL could be a better replacement for me, but if you can convice me, i'll switch!
I remember reading this article and thought of the similarity between the thread arcs of IBM and conversation clues of Microsoft. I'd sure like to see my Thunderbird support them!
Interesting though, that I can't recall seeing such a discussion about how IBM "reinvented" threading when the previous Slashdot article was posted. Or is it just that the average Slashdot ranter is unable to read or understand anything after having seen the character sequence M I C R O S O F T ?
Please. Everybody. Read the article. A bit too brief but nevertheless interesting. Thanks glawrie for the reference.