Minsky isn't angry that his AI "theories" aren't panning out. He's angry that AI researchers aren't making an attempt to think upnew AI theories that aren't glorified vaccuum cleaners.
He's been living on the taxpayer's dime for too long, then. Something like an automatic vacuum cleaner might actually be useful to the people who pay for all this research to be done. Academics have got to learn that the ivory tower doesn't isolate them from having to justify their existance by doing something useful.
Heh. If he gets too many hits, I think there's a good chance that I will be downgraded from Friend V2.0 to Acquaintance V1.5...
Serving only static files, on a threaded webserver and a decent OS, even a lowly Pentium can saturate a T1 without breaking a sweat. Hell, I remember saturating 155 Mbit/s ATMs (100x faster than a T1) on 1994-vintage DEC Alphas. I would say that most Slashdottings flatten routers and pipes rather than servers.
I wonder, why does RMS not call it "Gnunix" or something like that? It's snappier than "GNU/Linux", that's for sure. The domain name gnunix.org is still available too!
POSIX only defines the API's and the most minimal implementation details required[1] but does not detail how those API's should be implemented. It doesn't matter how you implement the scheduler as long as you expose the POSIX API's.
My point was, the definition of what is Unix and what isn't isn't straightforward. You cannot say "it has this API so it is Unix" nor "it has this shell so it is Unix" and so on. If the term "Unix" is to be meaningful at all (as distinct from "Unix-like") then a decision has to be made, is this Unix or isn't it? That's what the standards body is for.
There's nothing wrong with being "Unix-like" of course, but that's what a Unix-like product should call itself, not Unix.
Open Source code is its own standard. Standards are for secretive companies, for companies that don't trust each other, and for monopolists.
Uhh, right. Let's take something simple, like SMTP or POP3. There are multiple implementations of these standards, from sendmail/popper to Exchange. If they don't conform to the same standard, no-one gets their email. But since they do, not only can email get from A to B, but you can feasibly replace one with the other. How does that benefit a monopolist in any way? You want to talk open source, what if sendmail and qmail don't use the same SMTP standard? What if Apache and Mozilla don't use the same HTTP standard? See, saying "the code is the standard" only works if there is only one implementation. For everything else, you need a neutral third party to make sure everyone plays by the rules.
There hasn't been much movement on formal standards, at least among Unices.
POSIX, NFS, DCE, CDE/Motif, X11, Kerberos, etc etc etc. How can you not have heard of these?
moving KDE from being a collection of applications to a collection of pluggable components, with things like Konqueror becoming complicated wrappers for these components. The whole desktop is totally integrated - that's big, isn't it?
No, Microsoft did that first - that is after all what they got sued for, integrating their browser into their desktop!
Never mind that fact that they offered a seemless transition over the years from 68000 to PowerPC, from MacOS Classic to MacOS X.
Hardly seamless... if you used the 68040 FPU, you were outta luck when Apple didn't bother to emulate or thunk it cleanly. Sure you could port and recompile, but that doesn't help when you're using pre-written apps, and it's still hardly "seamless". I would have preferred to see Apple go to the 68060 (which was superior to PPC601 at the time) and who knows, maybe the 68080 after that...
C:\Documents and Settings\Dave>su 'su' is not recognized as an internal or external command, operable program or batch file.
That's funny, because on my Windows machine it says: Microsoft Windows 2000 [Version 5.00.2195] (C) Copyright 1985-2000 Microsoft Corp.
H:\>su su: user root does not exist
H:\>
An argument about what is or isn't installed by default is not the same as arguing about the merits of who copied who! After all, Cygwin predates MacOS X... could you say that Apple copied NT+Cygwin by creating a GUI with a Unix-style command line?
They're enough like Unix that they might as well be
Well, I don't know how much certification costs, but I'm sure Apple or IBM could afford it. IBM probably will do so in the post-AIX5 era, just to reassure the corporates that their in-house AIX apps can be recompiled without too much trauma. Apple probably don't care right now, but if they want to sell to the US govt they'll need FIPS, and they might do UNIX at the same time.
However, _under the hood_, Linux looks and smells like Unix, surely?
It's not as simple as that. Consider: Solaris(tm) is UNIX(tm) and its scheduler (for example) works a certain way. Whether this is the best way to do it is a matter of debate. Linux' scheduler works a different way, so is it the same under the hood, even tho' all the programmer sees is fork(), nice and all the rest?
That's why Open Group exists - to make these kinds of decisions of what's UNIX(tm) and what's not. Interestingly, Linux per se will never be UNIX(tm), it will have to be certified as individual distributions, for example Red Hat might get certified while SuSE doesn't. That is because POSIX includes criteria about the shell, libc and all the stuff outside the kernel too.
For everyone else, it looks like Unix, it acts like Unix, it smells like Unix. It's Unix.
It amazes me that Slashbots can criticize players like Microsoft for ignoring standards when it suits them, then turn around and do exactly the same thing themselves. Standards exist and are worth protecting because they make everyone's lives easier. If an OS is UNIX98 or POSIX compliant, then if means if you want to port your software to that platform, you can make certain assumptions before you start work that will vastly increase your chances of success within time and budget. And what "looks and smells" like Unix covers a wide range of ground, even Minix "looks and smells" a lot like Unix, but it simply doesn't have the capability of Linux let alone Solaris. An OS like OpenVMS isn't Unix, but you can compile and run plenty of Unix software on it, because of its POSIX API. NT with Cygwin can "look and smell" like Unix, but under the hood it's totally different.
If anyone can come along and write an OS that has $ as its prompt and you can type ls to get a list of files, does that make it a Unix? No, there's more to it than that. And that's why the Unix(r) brand exists.
Wouldn't some sort of software solution be the Hurd (if/when it becomes ready) in that as each system is a micro-kernel you just restart that bit of the operating system. As said in another post this is like/etc/rc.d but at a lower level.
QNX, I believe, already does this, and has been in production use throughout the world for years.
Nonsense. MS (for example) is NOT about to produce 3000 different versions of Office for it's customers nor will it provide them with the source so they can do that themselves. Those customizations simply won't happen at all.
It's simply a matter of scale. Oracle or SAP for example, will fork their codebase for a major customer, then merge those patches back into the trunk in a later version (I have personal experience of this). In the vertical application space, pretty much all the new features in a new version come from customer requests/suggestions - when you get enough of those outstanding, you simply roll them up and release a new version (again, personal experience).
Secondly, altho' Microsoft (obviously) do not release source code, their applications are insanely scriptable from VBA or anything that can script COM (perl, if you prefer). Almost every object exposes an interface. You know why there are all these macro viruses? Because Outlook/Exchange isn't just a mail system, it's a complete platform for workflow and groupware application development, and like any platform (Linux included) it's not bulletproof. So there is a lot of scope for customization and adding plugins, and a lot of people do do it.
The efficiency comes from slashing several layers of administrative overhead from the transaction. That is, MS devotes non-productive (to the economy as a whole) resources to vendor lock-in 'features' and anti copying measures, then throws more resources at a massive legal department and lobbying in order to back up those measures.
That is quite possibly true (not enough data to call it either way). Is the overhead of this outweighed by the benefits of centralizing software development? (Probably) no-one knows.
is quite likely that members of different companie's IT departments would work together on larger projects
A company might help its customers and suppliers, that happens a lot today, but so long as software has the potential to be a competitive advantage, sharing will be very limited. At the moment, choice of generic office software is not a competitive advantage, but say you create a set of tools that are very specific to your industry (say, publishing or law or something). Now you do have a competitive advantage, but if you want to share (distribute) that to your suppliers and customers, GPL forces you to distribute it to your competitors too. But under the present system, everyone starts with out-of-the-box MS Office, customizes it with VBA macros, COM objects, etc, and they wholly own their own enhancements.
Short summary, Free software has a good bit to offer users but will likely INCREASE demnd for programmers. Competition will be increased throughout the industry.
Not a dig at you personally, but it amazes me that most Slashbots can simultaneously complain about wages being driven down by Indian/Russian outsourcing and H1Bs, yet promote the use of free software.
Let's say that a feature is "nice to have" but not essential. A single vendor, like MS or Oracle or SAP can afford to include it, because the cost is spread among many customers, and everyone gets their nice-to-have feature, and a programmer has a job. But if there was no central vendor, then no company can individually justify coding the feature themselves, so no-one gets it and the programmer is out of a job. It's lose-lose.
We decided to focus our efforts on improving Internet site software.... Because of the constant need to upgrade the hardware and software of Internet sites, many of the engineering techniques used previously to help maintain system dependability are too expensive to be deployed.
(etc)
Translation: "when we started this project, we thought we'd be able to spin it off into a hot IPO and get rich!!"
There are a great many proprietary apps out there that various businesses wish they could customise feature X or add feature Y. Proprietary software means that those customizations are simply out of the question, so that's one less position for a programmer.
How do you figure that? All you're doing is moving the job, at best, from the vendor to the customer.
The net result of Free software taking over is that a huge inefficiency in the economy will be removed.
No, you're creating an inefficiency by removing economies of scale and specialization. Instead of one organization that knows the software inside out, you create hundreds who just know their little bit of it.
Imagine if all companies had their own couriers instead of using Fedex, or all companies generated their own electricity instead of buying it from the grid. Imagine Ford owned its own iron ore mines. Vertical integration of that kind has been thoroughly discredited. But that's what you're proposing recreating.
I like getting paid to write code. I'm pretty sure that a lot of other people do. If the companies don't sell the products, and make a lot of money, then the whole idea of a paid programmer will go away. That would be a bummer.
Indeed, I can imagine the same "geek" a decade or so apart:
20 years old, college sophomore: Dude, everything should be, like, totally free dude. Like the sun, dude, nobody has to like, pay for that and stuff. Pass the bong, dude...
30 years old, wife and 2 kids, mortgage, etc: Shit, man, why can't I find a job? My kids gotta eat!
RMS has written a lot about the economics of open source, but he makes some basic assumptions: that everyone involved in writing software has a foundation grant or academic tenure and the programmer not the user is the best person to decide how software should work and what software should be written. Essentially he totally decouples the producer and consumer. Maybe the world really looks like that from an ivory tower at MIT, but the market in software exists so that programmers can earn a living writing the software that the users want written. The present system, selling software as a product, works very well. RMS and all open source advocates would like to go back to the days where computer operators were the "high priests" of technology. The real crime of the software industry in their eyes is that it brought computing to the "unwashed masses".
If you think about it, any Array has all the ACID properties (if you ignore writing out to disk). You'll at least get a runtime error upon ACID failure.
Only if you are only updating one array element - analogous to one row in one table - within a single transaction. If you have a transaction that affects many rows in many tables, then you need full ACID. In practice, a database will use full ACID even if it only has to do a single row in a transaction. Speaking from personal experience, it's not uncommon to write a single transaction that updates multiple rows in each of 10 or 12 tables at once, and sometimes many more.
This is not just a problem with free, Free or open source software, but also with planned, structured development of commerical model software such as Windows.
That's not quite true. On Windows there's a standard mechanism for software to integrate with other software: COM. Expose a COM interface and the objects in your program can be reused, or you can script the application in VBA or whatever you want. Try it sometime; you can reuse bits of Word and bits of Excel to create an application that draws graphs, pastes them into documents, then prints them automatically, for example. Or you can write a custom browser by reusing bits of IE.
In the Open Source world, there's no-one with enough clout to say "right, we're using COM" or DCE or CORBA or OAF or SOAP or XML-RPC or anything like that. Everyone has their own ideas - which is both good and bad. Good because choice is good, bad because too much choice is if anything worse than too little. Open source has years of work before even cut-and-paste works as well as it does on Windows, let alone the Mac, and let alone anything more complex.
I like to call this number "software snobbery". Many people compare software applications feature for feature, paying no attention to what their requirements are. The fact is, a very large percent of the database market not need more than a single GB database for their current task.
Well, Monty started it by talking about SouthWest's booking system. Airlines are among the heaviest IT users in the world. If you want to drop names, be prepared to be called on your assertions by someone who does know. Someone asked me what a typical DB was in finance; I posted links to case studies on a couple of multi-terabyte databases. Also, SouthWest are an Oracle site, so they aren't even using MySQL for their seat reservations, so I don't know what Monty was talking about.
The fact is our MySQL deploys outnumber the Oracle deployments, and over time as MySQL and Postressql get better I'm expecting that MySQL will creep into Oracle space as well.
You assume that Oracle is a static target. It is not. PostgreSQL is a perfectly adequate database for small tasks; its SQL parser and core transaction processing is infinitely superior to MySQL. But both of them are beaten hands down by open source databases like SAP/DB and Borland Interbase (or Firebird, whatever it's called these days). I'd happily use either of those with databases of a few hundred M or even a few G.
Now, if open source advocates talked about those two when they talked database, I'd show some respect. But the constant worship of MySQL just shows that they don't know.
I dunno, i find it hard to believe that all "older" programmers would step gleefully into management after being paid code monkeys for x years. I love programming, but i'd hate to be a programmer's boss...
THe question is what sucks more, being a manager, or being managed by someone who hasn't a clue?
Programmers are much more manageable if one of their own kind is doing the managing.
Just from curiosity, how much data are we talking about for a large corporation, say SW Air or BofA?
Impossible to put a figure on the total amount of data that exists within an organization, but a typical SAN in a major financial institution has terabytes online. UBS Warburg has 2 Tb in just its general ledger database. Acxion has 25 Tb in its data warehouse, which will mainly be used for queries, whereas the GL database will be more transaction heavy. SouthWest is an Oracle customer, but it doesn't say here how much data they have.
Slashdot Mirror
Minsky isn't angry that his AI "theories" aren't panning out. He's angry that AI researchers aren't making an attempt to think upnew AI theories that aren't glorified vaccuum cleaners.
He's been living on the taxpayer's dime for too long, then. Something like an automatic vacuum cleaner might actually be useful to the people who pay for all this research to be done. Academics have got to learn that the ivory tower doesn't isolate them from having to justify their existance by doing something useful.
By scheduler you mean cron, right?
No, the thing that decides which process gets CPU time next.
Heh. If he gets too many hits, I think there's a good chance that I will be downgraded from Friend V2.0 to Acquaintance V1.5...
Serving only static files, on a threaded webserver and a decent OS, even a lowly Pentium can saturate a T1 without breaking a sweat. Hell, I remember saturating 155 Mbit/s ATMs (100x faster than a T1) on 1994-vintage DEC Alphas. I would say that most Slashdottings flatten routers and pipes rather than servers.
GNU/Polyx.
I wonder, why does RMS not call it "Gnunix" or something like that? It's snappier than "GNU/Linux", that's for sure. The domain name gnunix.org is still available too!
POSIX only defines the API's and the most minimal implementation details required[1] but does not detail how those API's should be implemented. It doesn't matter how you implement the scheduler as long as you expose the POSIX API's.
My point was, the definition of what is Unix and what isn't isn't straightforward. You cannot say "it has this API so it is Unix" nor "it has this shell so it is Unix" and so on. If the term "Unix" is to be meaningful at all (as distinct from "Unix-like") then a decision has to be made, is this Unix or isn't it? That's what the standards body is for.
There's nothing wrong with being "Unix-like" of course, but that's what a Unix-like product should call itself, not Unix.
Open Source code is its own standard. Standards are for secretive companies, for companies that don't trust each other, and for monopolists.
Uhh, right. Let's take something simple, like SMTP or POP3. There are multiple implementations of these standards, from sendmail/popper to Exchange. If they don't conform to the same standard, no-one gets their email. But since they do, not only can email get from A to B, but you can feasibly replace one with the other. How does that benefit a monopolist in any way? You want to talk open source, what if sendmail and qmail don't use the same SMTP standard? What if Apache and Mozilla don't use the same HTTP standard? See, saying "the code is the standard" only works if there is only one implementation. For everything else, you need a neutral third party to make sure everyone plays by the rules.
There hasn't been much movement on formal standards, at least among Unices.
POSIX, NFS, DCE, CDE/Motif, X11, Kerberos, etc etc etc. How can you not have heard of these?
moving KDE from being a collection of applications to a collection of pluggable components, with things like Konqueror becoming complicated wrappers for these components. The whole desktop is totally integrated - that's big, isn't it?
No, Microsoft did that first - that is after all what they got sued for, integrating their browser into their desktop!
Never mind that fact that they offered a seemless transition over the years from 68000 to PowerPC, from MacOS Classic to MacOS X.
Hardly seamless... if you used the 68040 FPU, you were outta luck when Apple didn't bother to emulate or thunk it cleanly. Sure you could port and recompile, but that doesn't help when you're using pre-written apps, and it's still hardly "seamless". I would have preferred to see Apple go to the 68060 (which was superior to PPC601 at the time) and who knows, maybe the 68080 after that...
C:\Documents and Settings\Dave>su
'su' is not recognized as an internal or external command,
operable program or batch file.
That's funny, because on my Windows machine it says:
Microsoft Windows 2000 [Version 5.00.2195]
(C) Copyright 1985-2000 Microsoft Corp.
H:\>su
su: user root does not exist
H:\>
An argument about what is or isn't installed by default is not the same as arguing about the merits of who copied who! After all, Cygwin predates MacOS X... could you say that Apple copied NT+Cygwin by creating a GUI with a Unix-style command line?
Thus, you can have whatever damn scheduler you like and still be UNIX(tm)-compliant.
By that argument, NT is more UNIX than Linux is, because NT has POSIX!
That is why I say that the definition of what is Unix and what is not is complex, and that's why a standards body exists to make the decision.
They're enough like Unix that they might as well be
Well, I don't know how much certification costs, but I'm sure Apple or IBM could afford it. IBM probably will do so in the post-AIX5 era, just to reassure the corporates that their in-house AIX apps can be recompiled without too much trauma. Apple probably don't care right now, but if they want to sell to the US govt they'll need FIPS, and they might do UNIX at the same time.
However, _under the hood_, Linux looks and smells like Unix, surely?
It's not as simple as that. Consider: Solaris(tm) is UNIX(tm) and its scheduler (for example) works a certain way. Whether this is the best way to do it is a matter of debate. Linux' scheduler works a different way, so is it the same under the hood, even tho' all the programmer sees is fork(), nice and all the rest?
That's why Open Group exists - to make these kinds of decisions of what's UNIX(tm) and what's not. Interestingly, Linux per se will never be UNIX(tm), it will have to be certified as individual distributions, for example Red Hat might get certified while SuSE doesn't. That is because POSIX includes criteria about the shell, libc and all the stuff outside the kernel too.
For everyone else, it looks like Unix, it acts like Unix, it smells like Unix. It's Unix.
It amazes me that Slashbots can criticize players like Microsoft for ignoring standards when it suits them, then turn around and do exactly the same thing themselves. Standards exist and are worth protecting because they make everyone's lives easier. If an OS is UNIX98 or POSIX compliant, then if means if you want to port your software to that platform, you can make certain assumptions before you start work that will vastly increase your chances of success within time and budget. And what "looks and smells" like Unix covers a wide range of ground, even Minix "looks and smells" a lot like Unix, but it simply doesn't have the capability of Linux let alone Solaris. An OS like OpenVMS isn't Unix, but you can compile and run plenty of Unix software on it, because of its POSIX API. NT with Cygwin can "look and smell" like Unix, but under the hood it's totally different.
If anyone can come along and write an OS that has $ as its prompt and you can type ls to get a list of files, does that make it a Unix? No, there's more to it than that. And that's why the Unix(r) brand exists.
Wouldn't some sort of software solution be the Hurd (if/when it becomes ready) in that as each system is a micro-kernel you just restart that bit of the operating system. As said in another post this is like /etc/rc.d but at a lower level.
QNX, I believe, already does this, and has been in production use throughout the world for years.
Nonsense. MS (for example) is NOT about to produce 3000 different versions of Office for it's customers nor will it provide them with the source so they can do that themselves. Those customizations simply won't happen at all.
It's simply a matter of scale. Oracle or SAP for example, will fork their codebase for a major customer, then merge those patches back into the trunk in a later version (I have personal experience of this). In the vertical application space, pretty much all the new features in a new version come from customer requests/suggestions - when you get enough of those outstanding, you simply roll them up and release a new version (again, personal experience).
Secondly, altho' Microsoft (obviously) do not release source code, their applications are insanely scriptable from VBA or anything that can script COM (perl, if you prefer). Almost every object exposes an interface. You know why there are all these macro viruses? Because Outlook/Exchange isn't just a mail system, it's a complete platform for workflow and groupware application development, and like any platform (Linux included) it's not bulletproof. So there is a lot of scope for customization and adding plugins, and a lot of people do do it.
The efficiency comes from slashing several layers of administrative overhead from the transaction. That is, MS devotes non-productive (to the economy as a whole) resources to vendor lock-in 'features' and anti copying measures, then throws more resources at a massive legal department and lobbying in order to back up those measures.
That is quite possibly true (not enough data to call it either way). Is the overhead of this outweighed by the benefits of centralizing software development? (Probably) no-one knows.
is quite likely that members of different companie's IT departments would work together on larger projects
A company might help its customers and suppliers, that happens a lot today, but so long as software has the potential to be a competitive advantage, sharing will be very limited. At the moment, choice of generic office software is not a competitive advantage, but say you create a set of tools that are very specific to your industry (say, publishing or law or something). Now you do have a competitive advantage, but if you want to share (distribute) that to your suppliers and customers, GPL forces you to distribute it to your competitors too. But under the present system, everyone starts with out-of-the-box MS Office, customizes it with VBA macros, COM objects, etc, and they wholly own their own enhancements.
Short summary, Free software has a good bit to offer users but will likely INCREASE demnd for programmers. Competition will be increased throughout the industry.
Not a dig at you personally, but it amazes me that most Slashbots can simultaneously complain about wages being driven down by Indian/Russian outsourcing and H1Bs, yet promote the use of free software.
Let's say that a feature is "nice to have" but not essential. A single vendor, like MS or Oracle or SAP can afford to include it, because the cost is spread among many customers, and everyone gets their nice-to-have feature, and a programmer has a job. But if there was no central vendor, then no company can individually justify coding the feature themselves, so no-one gets it and the programmer is out of a job. It's lose-lose.
Translation: "when we started this project, we thought we'd be able to spin it off into a hot IPO and get rich!!"
From the article:
Something is seriously fucked up here...
There are a great many proprietary apps out there that various businesses wish they could customise feature X or add feature Y. Proprietary software means that those customizations are simply out of the question, so that's one less position for a programmer.
How do you figure that? All you're doing is moving the job, at best, from the vendor to the customer.
The net result of Free software taking over is that a huge inefficiency in the economy will be removed.
No, you're creating an inefficiency by removing economies of scale and specialization. Instead of one organization that knows the software inside out, you create hundreds who just know their little bit of it.
Imagine if all companies had their own couriers instead of using Fedex, or all companies generated their own electricity instead of buying it from the grid. Imagine Ford owned its own iron ore mines. Vertical integration of that kind has been thoroughly discredited. But that's what you're proposing recreating.
I like getting paid to write code. I'm pretty sure that a lot of other people do. If the companies don't sell the products, and make a lot of money, then the whole idea of a paid programmer will go away. That would be a bummer.
Indeed, I can imagine the same "geek" a decade or so apart:
20 years old, college sophomore: Dude, everything should be, like, totally free dude. Like the sun, dude, nobody has to like, pay for that and stuff. Pass the bong, dude...
30 years old, wife and 2 kids, mortgage, etc: Shit, man, why can't I find a job? My kids gotta eat!
RMS has written a lot about the economics of open source, but he makes some basic assumptions: that everyone involved in writing software has a foundation grant or academic tenure and the programmer not the user is the best person to decide how software should work and what software should be written. Essentially he totally decouples the producer and consumer. Maybe the world really looks like that from an ivory tower at MIT, but the market in software exists so that programmers can earn a living writing the software that the users want written. The present system, selling software as a product, works very well. RMS and all open source advocates would like to go back to the days where computer operators were the "high priests" of technology. The real crime of the software industry in their eyes is that it brought computing to the "unwashed masses".
If you think about it, any Array has all the ACID properties (if you ignore writing out to disk). You'll at least get a runtime error upon ACID failure.
Only if you are only updating one array element - analogous to one row in one table - within a single transaction. If you have a transaction that affects many rows in many tables, then you need full ACID. In practice, a database will use full ACID even if it only has to do a single row in a transaction. Speaking from personal experience, it's not uncommon to write a single transaction that updates multiple rows in each of 10 or 12 tables at once, and sometimes many more.
Yeah but lots of companies went out of business for doing it (one of my former included)
Companies go out of business for many reasons. Their choice of word processor isn't one of them.
This is not just a problem with free, Free or open source software, but also with planned, structured development of commerical model software such as Windows.
That's not quite true. On Windows there's a standard mechanism for software to integrate with other software: COM. Expose a COM interface and the objects in your program can be reused, or you can script the application in VBA or whatever you want. Try it sometime; you can reuse bits of Word and bits of Excel to create an application that draws graphs, pastes them into documents, then prints them automatically, for example. Or you can write a custom browser by reusing bits of IE.
In the Open Source world, there's no-one with enough clout to say "right, we're using COM" or DCE or CORBA or OAF or SOAP or XML-RPC or anything like that. Everyone has their own ideas - which is both good and bad. Good because choice is good, bad because too much choice is if anything worse than too little. Open source has years of work before even cut-and-paste works as well as it does on Windows, let alone the Mac, and let alone anything more complex.
I like to call this number "software snobbery". Many people compare software applications feature for feature, paying no attention to what their requirements are. The fact is, a very large percent of the database market not need more than a single GB database for their current task.
Well, Monty started it by talking about SouthWest's booking system. Airlines are among the heaviest IT users in the world. If you want to drop names, be prepared to be called on your assertions by someone who does know. Someone asked me what a typical DB was in finance; I posted links to case studies on a couple of multi-terabyte databases. Also, SouthWest are an Oracle site, so they aren't even using MySQL for their seat reservations, so I don't know what Monty was talking about.
The fact is our MySQL deploys outnumber the Oracle deployments, and over time as MySQL and Postressql get better I'm expecting that MySQL will creep into Oracle space as well.
You assume that Oracle is a static target. It is not. PostgreSQL is a perfectly adequate database for small tasks; its SQL parser and core transaction processing is infinitely superior to MySQL. But both of them are beaten hands down by open source databases like SAP/DB and Borland Interbase (or Firebird, whatever it's called these days). I'd happily use either of those with databases of a few hundred M or even a few G.
Now, if open source advocates talked about those two when they talked database, I'd show some respect. But the constant worship of MySQL just shows that they don't know.
I dunno, i find it hard to believe that all "older" programmers would step gleefully into management after being paid code monkeys for x years. I love programming, but i'd hate to be a programmer's boss...
THe question is what sucks more, being a manager, or being managed by someone who hasn't a clue?
Programmers are much more manageable if one of their own kind is doing the managing.
Just from curiosity, how much data are we talking about for a large corporation, say SW Air or BofA?
Impossible to put a figure on the total amount of data that exists within an organization, but a typical SAN in a major financial institution has terabytes online. UBS Warburg has 2 Tb in just its general ledger database. Acxion has 25 Tb in its data warehouse, which will mainly be used for queries, whereas the GL database will be more transaction heavy. SouthWest is an Oracle customer, but it doesn't say here how much data they have.