> Its a big trilogy. It has very nice CG. Theme > music is great. > > But is this worthy of a "Best Director" award?
Well James Cameron won for the super-cheesy "Titanic" and Robert Zemeckis for the even worse "Forrest Gump".
Peter Jackson pulled off something that most people (including myself) thought was impossible - he did justice to Tolkien's novel and also changed the way films may be made in the future by filming three at once. All three movies are likely to be in the top 10 most popular movies of all time on IMDB. What more would you want?
It's difficult to think of another director who could have pulled it off.
>Peter Jackson is not a good director, as can be
>easily proven by looking at pretty much any
>dramatic scene in the movie (especially the ones
>where elves are present). They are all horribly
>framed, overlighted in an horrible, cheesy manner,
>and the music is always way overboard
Most of the things you are talking about have got very little to do with the director and writer. It's the cinematographer and lighting people who frame and light the shots, and the music director who chooses the music.
> "Considering his typical tendency to use > high-quality, large-format cameras and his > desire that it be handy and convenient, I > suspect he would be attracted to our D100, for > its size and versatility and overall digital > image quality."
He may have shot on film and manipulated the photos on computer after scanning it with a drum scanner - but there's no way he'd be using such low resolution digital cameras as a D100.
Remember Ansel Adams used a view camera that took 8x10 inch film.
Consider that the Nikon D100 can nowhere near match the resolution of high grain 35mm film. It only has a 6.1 megapixel image sensor, whereas any decent film scanner can extract atleast 12 megapixels from 35mm film - and nowhere near get down the grain level (eg. try finding the grain on Fuji velvia on a Nikon coolscan scanner).
Since 10x8 inches is roughly 14 times the area of 35mm * 23mm to get that level of detail you'd need atleast a 168 megapixel image.
Consider that Adam's was not in a hurry to get his images he was after the maximal control and quality possible (he spent a lot of the late years of his life re-printing his earlier photos).
> The movie studios should NOT make the same > mistakes that the music industry did. They > should start offering legitimate good quality > > legal downloads NOW, before too many people > start thinking about movies the way they do > mp3s.
Right, Job's comment about no instant gratifcation on movie downloads doesn't really make sense. It might take 30 mins to download the entire movie, but you can start watching the movie right away if you can download the movie in a shorter time than it's duration.
The real reason he isn't interested in movie downloads is that the iTunes store is more a vehicle to sell iPods, whereas an online movie store wouldn't help them to sell any hardware.
It's a pity since they have all the infracture there, along with Quicktime - they could build a great online movie store.
> Don't misunderstand me: nanotech can be useful.
> Dumb computers are useful right now. Things like micro-mechanical machines may be useful. Limited, one-task-only, expert system can be useful. But real intelligence? Real nanotech? I don't think so.
I agree. A lot of potential nano-applications (fabrication, repair etc) require autonomous, intelligent nanobots which are able to navigate and propel themselves around at a nano-scale.
Considering robotics and AI are nowhere near the state of being able to do this at a human scale, you'd have to think that these applications must be far, far off.
The most intriguing explanation for the Big Bang I've seen recently come from String theory.
The idea is that the Big Bang may have been another universe colliding with our own at a single point in 11-dimension space. The energy of the collision resulting in a huge amount of mass being created.
If this is true, this means that there may be more than one Big Bang (or more in the future). For more on this read the Elegant Universe by Brian Greene, or watch tv series at here.
Unfortunately, like the 'review' - it doesn't mention which frameworks the book covers though.
Re:5 years in the business... WHERE???
on
Effective XML
·
· Score: 1
> XSLT providing tremendous advantages in
> transforming data for a variety of other purposes (you
> probably hated lisp/scheme based language, too).
Gah. I'm tired of people comparing XSLT to Lisp or Scheme. Okay XSLT can transform and generate itself just like Lisp, but that's where the similarities end. In almost every other design aspect it is the opposite of Lisp.
XSLT is an incredibly baroque, verbose language, only useful for a very limited set of trivial XML transformations, (ie. surprise style sheets!) that involve no I/O or complex computations. If you do a lot of this - then maybe it is worth learning, but my experience is that you can hit it's limits very quickly.
Lisp on the other hand is a incredibly elegant, compact and powerful general purpose language that has been used in almost every application domain imaginable. Maybe the most elegant, clear and powerful single programming language ever invented, where very complex functionality can often be written in an amazingly small amount of understandable code.
Wedding photographers may be overpaid, but equipment and printing costs are also rather high. Consider that the Zeiss lenses for Hassleblads can cost up to $3000, and 8x10 prints at a professional photo lab will be around $30 each, more for custom prints. It must all add up pretty quickly.
Also consider the stress involved on the day and the required mix of technical and people skills involved. Certainly not an easy job to do well.
> I'm a professional software developer. I am *far* better served by Linux than Mac OS X
Really? I use Linux at work to do Java development, all the standard tools I use are also available on OS X (plus more, eg. xcode). Emacs, Eclipse, Code warrior, ant, maven, Jboss, JEdit etc.
That is true for all Linux software, since it's based on open source software and OS X is unix based - if it's popular there's generally going to be a port (even if you might have to use X11).
That's why the quality software available for OS X is always going to be a superset of what is available for Linux.
> And why is Photoshop better again? Is it because
> you shelled out $$$ for it? Unless you need CMYK
> support (and you don't need that for electronic
> publishing), GIMP is just fine. And people use
> ImageMagick for converting files in UNIX.
GIMP may be fine for preparing graphics for the web - because accurate colorspace management is useless when PC monitors vary so much, but if you are doing work that needs to go into a print publication, it seems like it lacks some pretty fundamental features. CMYK support yes, but also other color management features, support for ICC device profiles and colorspace conversions.
Without these features there is no way to get accurate colors in from a scanner or out to a printer, or to mix images from multiple sources in the same document. It's not just a matter of adjusting gamma values, you need to have profiles of the various RGB colorspaces that describe how various points in RGB spaces map to each other (how they are suppose to 'look').
There have been attempts to provide this type of functionality in Gimp through plugins but these are in fairly early stages. Also i believe a lot of the fundament color management algorithms are protected by patents.
Thoughts about j2ee
on
Bitter EJB
·
· Score: 3, Interesting
I've been doing J2ee development for the last couple of years and on the whole I don't think EJBs give you many advantages but they do add a lot of confusion.
The main problem is that it isn't clear the best way to use them, there are so many techniques and patterns out there but there is nothing you can point at that is the 'best practice' - it all depends on your application.
The advantages of easily scalability of business objects over multiple servers is useful, but since the database is the main bottleneck in most enterprise applications the scalability of the business logic layer doesn't give you a lot. Also many applications require communication between the various clients (through the server) but since EJBs are designed to be location independent, there is no nice ways of communicating between session EJBs on various machines. It is possible to do it with message EJBs but it's not easy. Another thing to note is that you don't need EJBs to separate the business layer from the presentation layer, it's easy to do with normal classes. You also don't need EJBs to have the business logic running on another server - you can just use RMI.
The automatic EJB transaction management is useful but since it relies on exceptions being thrown and it can get very confusing, and often isn't flexible enough - we seem to resort a lot of the time to UserTransactions.
On top of that a complex database schema seems also impossible to map to Entity EJBs in a useful way. We have 300+ tables and the presentation layer needs all sorts of combinations of relations between the data that mapping the data into 'Entities' just seem to make data access inefficient.
There are many alternative theories for petrolium
formation, many are 'abiogenic' theories that say that 'fossil fuels' are actually primordial, that have existed since the Earth was created.
For more info read see this and "The Deep Not Biosphere" by Thomas Gold of Cornell university.
> In many ways the enviroments are inverses of each other, much as the East-West cultural issues.
UNIX was designed to manage multiple users connecting via text terminals, whereas Windows was designed to manage single users running GUI applications in a client/server environment.
ESR says that Windows NT lacks any underlying metaphor, but I would argue it is much more 'object'-based than UNIX, from devices managed by the Object Manager in the kernel, right up to OLE/COM object interface for applications/components.
It's not hard to see how objects with Properties and Methods could provide advantages over Files and Pipes/Signals (if done right), even if it adds a certain amount of complexity.
In my office we doing Java development using Sun JDK 1.4.1 & JBoss on PCs running Linux. I'm wondering if anyone has experience of using Macs/OS x in a similarly environment? On the surface of things, everything should be fine since OS X supports all the standard tools, apache, tomcat, Ant, JBoss, perl, python, Emacs etc and the compile time and JRE spped on a G5 should be comparible to a high-end PC (or is it?).
Does anyone have any experience of using a Mac such a PC/Linux-based environment? I'm especially interested in performance comparisons and compatibility btw Apple's JDK and Sun JDK and JBoss performance, of a G5 compared to a PC
The motivation for switching to OS x would be for the non-development related software that it runs natively (eg. Office, Lotus Notes, Photoshop etc) etc that i have to Wine or VMWare to use under Linux (as well as Aqua generally seeming a nicer environment to use).
> The media tends to run stories for the purpose of > ratings/circulation. Left/Right does not come into > it very much at all.
Although ratings and circulation are important, the media's real clients and main source of income are advertizers (ie. large companies). Large companies definitely have political agendas, and by investing or withdrawing advertising revenue from a newspaper they have a real effect on what you read.
All of these stories have been mentioned in the mainstream media, but they have not become reoccuring stories, but they certainly are issues talked about in the foreign press.
For example... #1 It is *fact* that the Bush administration is made up of hardline conservatives with the stated ideology of increasing the US's power in the world including increased control of the middle east and optionally sidelining the UN.
This should be an issue to you if you are conservative or liberal. To any thinking person this should be a huge issue - whether you agree with them or not, because they have a pretty radical vision for the US's role in the world that most people would find shocking.
How could this group's ideology NOT be a important determining factor in attacking Iraq?
How could the US press not investigate this link further without being derelect in their duties?
It should be the press's role to keep the Government honest, it is sad that many Americans don't understand that questioning the motives of your rulers isn't always politically motivated.. it is what the press are there for.
Matt Ridley interview
on
The Red Queen
·
· Score: 4, Informative
There's an interesting video interview with Matt Ridley where he talks about his latest book 'Nature via Nurture' on edge.org
http://www.edge.org/3rd_culture/ridley03/ridley_in dex.html
If someone was *really* from 250 years in the future, would it be obvious just by talking to them? Would it be easy to distinguish between a real time traveller and some delusional person who thinks they are from the future or a deliberate fake?
On one hand they probably speak very strangely, using words, grammar and ideas that have not been developed yet but are consistent. They would also have a very detailed and consistent knowledge of a future world way past the ability of the average person to make up. They probably would also have knowledge of future technologies and so forth and be able to describe them in the detail. But none of this would be verifiable, and most of it would probably sound outlandish.
If you were zapped back to 1750 you'd probably be able to tell people about aeroplanes, tungsten light bulbs, telephones, computers etc. but you would be hard to build any of them from scratch. Would anyone believe you?
I think the author makes a good point that different programming paradigms are suited to particular problem domains. I think it is not well understood when different models are the most suitable, and even when we do it is too hard to mix paradigms within the same program, or the language of choice doesn't support the best option.
Modelling a GUI interface is done very naturally with OOP because is easy to isolate separate independent components, but OOPs seem less well suited to problems involving relationships across many types of data where structured, relational or functional paradigms are more suited. (I think the average database backed website is a good example - or a mathemetical engineering application)
Perhaps in the future programmers will better understand better which mixtures of paradigms to use and languages will allow this better. Progress will probably also be made in bridging the gulf between programming models so they may be more easily used together. An example of this is EJB-QL (part of the EJB 2.1 spec), a relational-like query language that works on top of EJB objects.
I tried out Eclipse 2.0, although it seemed to have a number of good features, on the whole I found the user interface very clunky and confusing to use (as well as a bit slow). It certainly didn't make me want to swap from Emacs and ant.
Open a new 'Perspective', change your 'View', select an 'Implementor', navigate 'Resources'. Sure these concepts may be useful ideas for people wanting to add new components to the Eclipse platform but why should you need to understand these terms to use the software itself?
It feels Eclipse was designed by a group of architecture astronauts) with the target audience being developers who will extend the platform more than the actual end users themselves.
Re:Obligatory Anti-Pattern Viewpoint
on
Design Patterns
·
· Score: 1
Wow you are very mixed up with your terminology (and in your thinking in general).
Sure SQL as a declarative relational language is quite powerful for certain tasks, but by itself it isn't a general purpose programming language (for a start it's associated with a permanent data store, it doesn't have any I/O etc). There are also computation tasks which are very difficult to do in it (try maintaining recurse data structures such as trees).
So you clump it together with 'procedural' programming and then mention something vague about about "formulas instead of patterns".
> IOW, "formulas" versus "patterns". > Formulas are not only more compact, but don't > clutter up the larger-scale structure of the > code. (LISP fans say similar things about GOF > patterns, I would note).
The concept of a pattern is as applicable to procedural or relational programming as it is to object oriented design (or architecture or whatever other design problems you have). It just happens that the GOF book is about OO design patterns.
What do you mean by 'formulas'? Do you mean declarative statements (like SQL) versus imperative (for loops etc)? Your statement is meaingless to me.
> Thus, you are not only reinventing a (halfass) > database with GOF, but reinventing an > out-of-style one on top of that. A double-whammy > no-no in my book.
The OO design patterns in GOF are pretty fundamental and the problems they solve have been around forever. eg. database 'triggers' are an example of the observer pattern. Most aren't applicable to relational model by itself at all (eg. facade, factory etc)...
On your website you map the OO patterns into the database equivalents, but it's ridiculous, all you are doing it mapping classes to tables, objects to rows, attributes to fields. So what? This isn't relational programming! You still need procedural code on top of it to implement the pattern, and then you are just simulating OO programming with procedures.
> The GOF Patterns movement is mostly a rehash > of technology and philosophy that died in the > early 70's, as variations of Codd's relational > model proved superior
And Codd's relational model was influenced by earlier research on knowledge representation in AI frames and semantic networks etc, inheritance in object orientation was also used in frames.
Patterns are new because the give a name to design problems that previously had no names. Most of the OO design patterns are not applicable for relational programming but there is no reason to stop you from creating relational SQL patterns.
Slashdot Mirror
> I'll say it.
> Why should Peter Jackson get the Oscar?
> Its a big trilogy. It has very nice CG. Theme
> music is great.
>
> But is this worthy of a "Best Director" award?
Well James Cameron won for the super-cheesy "Titanic" and Robert Zemeckis for the even worse "Forrest Gump".
Peter Jackson pulled off something that most people (including myself) thought was impossible - he did justice to Tolkien's novel and also changed the way films may be made in the future by filming three at once. All three movies are likely to be in the top 10 most popular movies of all time on IMDB. What more would you want?
It's difficult to think of another director who could have pulled it off.
>Peter Jackson is not a good director, as can be
>easily proven by looking at pretty much any
>dramatic scene in the movie (especially the ones
>where elves are present). They are all horribly
>framed, overlighted in an horrible, cheesy manner,
>and the music is always way overboard
Most of the things you are talking about have got very little to do with the director and writer. It's the cinematographer and lighting people who frame and light the shots, and the music director who chooses the music.
> "Considering his typical tendency to use > high-quality, large-format cameras and his
> desire that it be handy and convenient, I
> suspect he would be attracted to our D100, for
> its size and versatility and overall digital
> image quality."
He may have shot on film and manipulated the photos on computer after scanning it with a drum scanner - but there's no way he'd be using such low resolution digital cameras as a D100.
Remember Ansel Adams used a view camera that took 8x10 inch film.
Consider that the Nikon D100 can nowhere near match the resolution of high grain 35mm film. It only has a 6.1 megapixel image sensor, whereas any decent film scanner can extract atleast 12 megapixels from 35mm film - and nowhere near get down the grain level (eg. try finding the grain on Fuji velvia on a Nikon coolscan scanner).
Since 10x8 inches is roughly 14 times the area of 35mm * 23mm to get that level of detail you'd need atleast a 168 megapixel image.
Consider that Adam's was not in a hurry to get his images he was after the maximal control and quality possible (he spent a lot of the late years of his life re-printing his earlier photos).
> The movie studios should NOT make the same
> mistakes that the music industry did. They
> should start offering legitimate good quality >
> legal downloads NOW, before too many people
> start thinking about movies the way they do
> mp3s.
Right, Job's comment about no instant gratifcation on movie downloads doesn't really make sense. It might take 30 mins to download the entire movie, but you can start watching the movie right away if you can download the movie in a shorter time than it's duration.
The real reason he isn't interested in movie downloads is that the iTunes store is more a vehicle to sell iPods, whereas an online movie store wouldn't help them to sell any hardware.
It's a pity since they have all the infracture there, along with Quicktime - they could build a great online movie store.
This algorithm has a slightly different result from the recursive algorithm.
For an even number of disks it your code moves the disks from stack 0 to 1, and for an odd number of disks from stack 0 to 2.
(The recursive algorithm in the 102 examples moves the disks from stack 0 to 2 always).
> Don't misunderstand me: nanotech can be useful.
> Dumb computers are useful right now. Things like micro-mechanical machines may be useful. Limited, one-task-only, expert system can be useful. But real intelligence? Real nanotech? I don't think so.
I agree. A lot of potential nano-applications (fabrication, repair etc) require autonomous, intelligent nanobots which are able to navigate and propel themselves around at a nano-scale.
Considering robotics and AI are nowhere near the state of being able to do this at a human scale, you'd have to think that these applications must be far, far off.
The most intriguing explanation for the Big Bang I've seen recently come from String theory.
The idea is that the Big Bang may have been another universe colliding with our own at a single point in 11-dimension space. The energy of the collision resulting in a huge amount of mass being created.
If this is true, this means that there may be more than one Big Bang (or more in the future). For more on this read the Elegant Universe by Brian Greene, or watch tv series at here.
You can read the first chapter here.
Unfortunately, like the 'review' - it doesn't mention which frameworks the book covers though.
> XSLT providing tremendous advantages in
> transforming data for a variety of other purposes (you
> probably hated lisp/scheme based language, too).
Gah. I'm tired of people comparing XSLT to Lisp or Scheme. Okay XSLT can transform and generate itself just like Lisp, but that's where the similarities end. In almost every other design aspect it is the opposite of Lisp.
XSLT is an incredibly baroque, verbose language, only useful for a very limited set of trivial XML transformations, (ie. surprise style sheets!) that involve no I/O or complex computations. If you do a lot of this - then maybe it is worth learning, but my experience is that you can hit it's limits very quickly.
Lisp on the other hand is a incredibly elegant, compact and powerful general purpose language that has been used in almost every application domain imaginable. Maybe the most elegant, clear and powerful single programming language ever invented, where very complex functionality can often be written in an amazingly small amount of understandable code.
Wedding photographers may be overpaid, but equipment and printing costs are also rather high. Consider that the Zeiss lenses for Hassleblads can cost up to $3000, and 8x10 prints at a professional photo lab will be around $30 each, more for custom prints. It must all add up pretty quickly.
Also consider the stress involved on the day and the required mix of technical and people skills involved. Certainly not an easy job to do well.
Really? I use Linux at work to do Java development, all the standard tools I use are also available on OS X (plus more, eg. xcode). Emacs, Eclipse, Code warrior, ant, maven, Jboss, JEdit etc.
That is true for all Linux software, since it's based on open source software and OS X is unix based - if it's popular there's generally going to be a port (even if you might have to use X11).That's why the quality software available for OS X is always going to be a superset of what is available for Linux.
> you shelled out $$$ for it? Unless you need CMYK
> support (and you don't need that for electronic
> publishing), GIMP is just fine. And people use
> ImageMagick for converting files in UNIX.
GIMP may be fine for preparing graphics for the web - because accurate colorspace management is useless when PC monitors vary so much, but if you are doing work that needs to go into a print publication, it seems like it lacks some pretty fundamental features. CMYK support yes, but also other color management features, support for ICC device profiles and colorspace conversions.
Without these features there is no way to get accurate colors in from a scanner or out to a printer, or to mix images from multiple sources in the same document. It's not just a matter of adjusting gamma values, you need to have profiles of the various RGB colorspaces that describe how various points in RGB spaces map to each other (how they are suppose to 'look').
There have been attempts to provide this type of functionality in Gimp through plugins but these are in fairly early stages. Also i believe a lot of the fundament color management algorithms are protected by patents.
I've been doing J2ee development for the last couple of years and on the whole I don't think EJBs give you many advantages but they do add a lot of confusion.
The main problem is that it isn't clear the best way to use them, there are so many techniques and patterns out there but there is nothing you can point at that is the 'best practice' - it all depends on your application.
The advantages of easily scalability of business objects over multiple servers is useful, but since the database is the main bottleneck in most enterprise applications the scalability of the business logic layer doesn't give you a lot. Also many applications require communication between the various clients (through the server) but since EJBs are designed to be location independent, there is no nice ways of communicating between session EJBs on various machines. It is possible to do it with message EJBs but it's not easy. Another thing to note is that you don't need EJBs to separate the business layer from the presentation layer, it's easy to do with normal classes. You also don't need EJBs to have the business logic running on another server - you can just use RMI.
The automatic EJB transaction management is useful but since it relies on exceptions being thrown and it can get very confusing, and often isn't flexible enough - we seem to resort a lot of the time to UserTransactions.
On top of that a complex database schema seems also impossible to map to Entity EJBs in a useful way. We have 300+ tables and the presentation layer needs all sorts of combinations of relations between the data that mapping the data into 'Entities' just seem to make data access inefficient.
There are many alternative theories for petrolium formation, many are 'abiogenic' theories that say that 'fossil fuels' are actually primordial, that have existed since the Earth was created.
For more info read see this and "The Deep Not Biosphere" by Thomas Gold of Cornell university.
That's true. According to this this the encryption key is your login password. (Or a master password).
It's hard to see how this couldn't be vulnerable to brute force/dictionary attacks.
> In many ways the enviroments are inverses of each other, much as the East-West cultural issues.
UNIX was designed to manage multiple users connecting via text terminals, whereas Windows was designed to manage single users running GUI applications in a client/server environment.
ESR says that Windows NT lacks any underlying metaphor, but I would argue it is much more 'object'-based than UNIX, from devices managed by the Object Manager in the kernel, right up to OLE/COM object interface for applications/components.
It's not hard to see how objects with Properties and Methods could provide advantages over Files and Pipes/Signals (if done right), even if it adds a certain amount of complexity.
A slightly offtopic but related question....
In my office we doing Java development using Sun JDK 1.4.1 & JBoss on PCs running Linux. I'm wondering if anyone has experience of using Macs/OS x in a similarly environment? On the surface of things, everything should be fine since OS X supports all the standard tools, apache, tomcat, Ant, JBoss, perl, python, Emacs etc and the compile time and JRE spped on a G5 should be comparible to a high-end PC (or is it?).
Does anyone have any experience of using a Mac such a PC/Linux-based environment? I'm especially interested in performance comparisons and compatibility btw Apple's JDK and Sun JDK and JBoss performance, of a G5 compared to a PC
The motivation for switching to OS x would be for the non-development related software that it runs natively (eg. Office, Lotus Notes, Photoshop etc) etc that i have to Wine or VMWare to use under Linux (as well as Aqua generally seeming a nicer environment to use).
> The media tends to run stories for the purpose of > ratings/circulation. Left/Right does not come into > it very much at all.
Although ratings and circulation are important, the media's real clients and main source of income are advertizers (ie. large companies). Large companies definitely have political agendas, and by investing or withdrawing advertising revenue from a newspaper they have a real effect on what you read.
All of these stories have been mentioned in the mainstream media, but they have not become reoccuring stories, but they certainly are issues talked about in the foreign press.
For example...
#1 It is *fact* that the Bush administration is made up of hardline conservatives with the stated ideology of increasing the US's power in the world including increased control of the middle east and optionally sidelining the UN.
This should be an issue to you if you are conservative or liberal. To any thinking person this should be a huge issue - whether you agree with them or not, because they have a pretty radical vision for the US's role in the world that most people would find shocking.
How could this group's ideology NOT be a important determining factor in attacking Iraq?
How could the US press not investigate this link further without being derelect in their duties?
It should be the press's role to keep the Government honest, it is sad that many Americans don't understand that questioning the motives of your rulers isn't always politically motivated.. it is what the press are there for.
There's an interesting video interview with Matt Ridley where he talks about his latest book 'Nature via Nurture' on edge.org http://www.edge.org/3rd_culture/ridley03/ridley_in dex.html
If someone was *really* from 250 years in the future, would it be obvious just by talking to them? Would it be easy to distinguish between a real time traveller and some delusional person who thinks they are from the future or a deliberate fake?
On one hand they probably speak very strangely, using words, grammar and ideas that have not been developed yet but are consistent. They would also have a very detailed and consistent knowledge of a future world way past the ability of the average person to make up. They probably would also have knowledge of future technologies and so forth and be able to describe them in the detail. But none of this would be verifiable, and most of it would probably sound outlandish. If you were zapped back to 1750 you'd probably be able to tell people about aeroplanes, tungsten light bulbs, telephones, computers etc. but you would be hard to build any of them from scratch. Would anyone believe you?
I think the author makes a good point that different programming paradigms are suited to particular problem domains. I think it is not well understood when different models are the most suitable, and even when we do it is too hard to mix paradigms within the same program, or the language of choice doesn't support the best option.
Modelling a GUI interface is done very naturally with OOP because is easy to isolate separate independent components, but OOPs seem less well suited to problems involving relationships across many types of data where structured, relational or functional paradigms are more suited. (I think the average database backed website is a good example - or a mathemetical engineering application)
Perhaps in the future programmers will better understand better which mixtures of paradigms to use and languages will allow this better. Progress will probably also be made in bridging the gulf between programming models so they may be more easily used together. An example of this is EJB-QL (part of the EJB 2.1 spec), a relational-like query language that works on top of EJB objects.
I tried out Eclipse 2.0, although it seemed to have a number of good features, on the whole I found the user interface very clunky and confusing to use (as well as a bit slow). It certainly didn't make me want to swap from Emacs and ant.
Open a new 'Perspective', change your 'View', select an 'Implementor', navigate 'Resources'. Sure these concepts may be useful ideas for people wanting to add new components to the Eclipse platform but why should you need to understand these terms to use the software itself?
It feels Eclipse was designed by a group of architecture astronauts) with the target audience being developers who will extend the platform more than the actual end users themselves.
Wow you are very mixed up with your terminology (and in your thinking in general).
Sure SQL as a declarative relational language is quite powerful for certain tasks, but by itself it isn't a general purpose programming language (for a start it's associated with a permanent data store, it doesn't have any I/O etc). There are also computation tasks which are very difficult to do in it (try maintaining recurse data structures such as trees).
So you clump it together with 'procedural' programming and then mention something vague about about "formulas instead of patterns".
> IOW, "formulas" versus "patterns".
> Formulas are not only more compact, but don't
> clutter up the larger-scale structure of the
> code. (LISP fans say similar things about GOF
> patterns, I would note).
The concept of a pattern is as applicable to procedural or relational programming as it is to object oriented design (or architecture or whatever other design problems you have). It just happens that the GOF book is about OO design patterns.
What do you mean by 'formulas'? Do you mean declarative statements (like SQL) versus imperative (for loops etc)? Your statement is meaingless to me.
> Thus, you are not only reinventing a (halfass)
> database with GOF, but reinventing an
> out-of-style one on top of that. A double-whammy
> no-no in my book.
The OO design patterns in GOF are pretty fundamental and the problems they solve have been around forever. eg. database 'triggers' are an example of the observer pattern. Most aren't applicable to relational model by itself at all (eg. facade, factory etc)...
On your website you map the OO patterns into the database equivalents, but it's ridiculous, all you are doing it mapping classes to tables, objects to rows, attributes to fields. So what? This isn't relational programming! You still need procedural code on top of it to implement the pattern, and then you are just simulating OO programming with procedures.
> The GOF Patterns movement is mostly a rehash
> of technology and philosophy that died in the
> early 70's, as variations of Codd's relational
> model proved superior
And Codd's relational model was influenced by earlier research on knowledge representation in AI frames and semantic networks etc, inheritance in object orientation was also used in frames.
Patterns are new because the give a name to design problems that previously had no names. Most of the OO design patterns are not applicable for relational programming but there is no reason to stop you from creating relational SQL patterns.