> You're talking Object Pascal as opposed to pascal.
Borland Pascal introduced object features in the later versions of the language. Not the same as Delphi, but it laid the groundwork.
> It is possible to implement multiple-inheritance classes in pascal
Example?
> Turbo pascal isn't pascal any more than Turbo C is C
Huh? Be specific.
> This is why the single-inheritance schema was chosen for the VCL.
No, the Delphi compiler and the VCL were developed concurrently, adding features to the compiler (and language) as the need for them arose in the VCL. The design of the VCL was not a consequence of an existing compiler. It is the way it is because the designers WANTED it that way. They came from the same school of thought as the Java designers (in fact, Borland worked closely with Sun on Java), and while cynics like to think that they eschewed multiple inheritance because it's hard to implement (which it is), there's a lot more to that decision than that. After all, they already had a market leading C++ compiler, so they had the expertise. If there is any afterthought in there at all, it's C++ Builder. After much crying and whining from the community (myself included) over the choice of Pascal instead of C++, Borland gave in. Never mind that many of them learned to love OP on its own merits.
> I should have mentioned that not all server-centric apps need be browser-based
Of course not, I should know. I write mainly C/S apps, RPC style (and nowadays, web service-style). In that case you have two apps, one on the desktop, one remote. And both can still benefit handsomely from components--database components on the backend, GUI and communications components on the frontend. It still doesn't signal a move away from components. Face it, if there is one positive upshot of the whole OOP movement, it's component programming. Ok, I read that somewhere in DDJ, but it's true.
There's lots of noise in the reviews, but you can still filter out a useful signal. Many reviews are also obviously dumb, such as the guy that ordered a CD player, loved it, but decided he didn't need it after all, so he gave it one star. Brains entirely optional. After reading through ten or twenty reviews of the CD player, you'd find recurring mentions of good or bad battery life, skipping/no skipping etc, which is usually all you need.
> There's no real reason why it couldn't now be implemented as a multiple-inheritance class > structure. The reason for single-class inheritance is historical.
I don't know where you're getting your facts. Object Pascal was never intended as a multiple-inheritance language, it had nothing to do with Borland being unable to implement MI. Borland's C++ compiler did MI capably long before Delphi came about. Delphi's single inheritance is a school-of-thought thing. Heijlsberg designed Delphi and the VCL to be THE end product, not a transition product on the way to the REAL C++ product. He was always a Pascal guy, ever since he created TP 1. They only reused the VCL in C++ Builder because it already existed and took less work than rewriting it from scratch and using a different inheritance model.
> However, with the advent of server-centric apps, there's not much need for "visual components",
Well, we definitely have different points of view on that. I don't think the browser will become THE platform anytime soon, at least not with current technologies. Relatively few types of applications can be expressed well in a stateless Form paradigm with the bulk of the code executing remotely. A rich user experience will ALWAYS require either lots of client-side code and components, or a very fat pipe to do X-type remoting. "Thin client" computing (as if!) looks good on paper and management loves it for all the obvious reasons, but it will never completely replace client side apps. Some things will always make more sense in a browser and others on the desktop.
Besides, even for server side programming the component approach is gaining hold. Just look at Microsoft's ASP.NET and Borland's web components. Face it, intermingled HTML-and-script programming is a major kludge and in no way the wave of the future. The web didn't all of a sudden make proven concepts such as separation of interface and logic obsolete. It's just that the technologies were still in their infancy and were often driven by neophytes that knew Perl and HTML and thought they ruled the world.
Just? Considering that, after the class library, the IDE is what makes you most productive, that's hardly a minor thing. Sure, you could buy a third party code editor, but that's yet more money, and most people won't.
> hacks caused by the use of single-parent-hierarchy class structure
Well, it sounds like you have issues with the language itself then. Perhaps Delphi isn't for you after all.
> trying to be all things to all people
Borland's biggest problem is that most people aren't using Delphi, not inherent problems in the product itself. Maybe it's something they can't do much about, since it's hard to blame someone for failing to compete well with Microsoft. As a product, I fail to see anything majorly wrong with Delphi, especially when compared to Microsoft's offerings.
> I didn't see anything in it that would make it worth my while
There are three things in particular in 4+ that I can't live without anymore: CTRL-SHIFT-C to create method bodies for all method declarations in a class (quite tedious to do manually if you don't have to), CTRL-SHIFT-Up/Down to switch between method declaration and definition, and CTRL-Click on an identifier to take you to its declaration. Every time I have to go to pre-4 I feel naked without these aids.
> I stopped using Borland products after Delphi 3 (the newer ones didn't offer that much more...)
You stopped at the wrong version. Many useful features were added in 4 and properly debugged in 5. If you use Professional, version 5 is your best bet. There are lots of IDE improvements (in particular navigation and code completion), plus forms are stored as text. I don't think 6 and 7 added much value to Professional, mainly lots of Enterprise and Web stuff.
> While work per cycles is important, work per Watt > is even more important in embedded systems.
Not really, unless you're talking battery operated, or are dealing with extreme heat generation. Judging by many embedded systems nowadays (e.g. set-top boxes from cable providers), heat generation IS an issue, but you didn't specifically address that.
Besides, I was being very generous comparing the 250 MHz PPC to the 1 GHz Duron, because they're certainly not in the same performance class. Don't forget that the Duron itself has a work-per-cycle advantage over Intel, so it's no slouch either. Plus, while I don't know too much about the PPC, it's still an old school RISC design if I'm not wrong, meaning that it has fewer complex instructions and might require more of them than an x86 to accomplish some things (fp math/fancy addressing).
> The cost of this product comes in the engineering work that > went into designing a custom PCB and layout for this product.
That's right, and that's why it might not make much sense for most of the applications people are salivating over here. The cheap (but potentially hotter, though still fanless) Via Epia boards are perfect for these geek-type "embedded" applications, such as MP3 and MP4/DivX players (though the jury is still out on whether the C3 processor has quite enough juice for the latter). The extra heat and power won't matter much, but the very substantial cost savings will.
> if you really think that clock speed is all that counts
Obviously you're one who thinks that the performance of PowerPC expands to beat any competitor at any given time. While the x86 and PPC are not comparable MHz for MHz, you can certainly figure out how much work each one does per cycle. Just for comparison, a dirt cheap throw away CPU nowadays would be a 1GHz Duron (around $35), which runs four times faster than this particular PPC. The PPC would have to execute four times more instructions per cycle than the Duron to be roughly comparable, and you can bet your life that it doesn't. So, all in all, while the PPC might be more efficient per cycle, it is not per dollar. IOW, for any given dollar amount you can buy considerably more raw processing power in an x86.
For example, check out the Epia boards from Via. For around $100 you get roughly the same processing power plus all conceivable peripherals on board. Add a $30 graphical LCD from Crystalfontz, get (or make) a suitable case, and you're set for WAY less than $500. Especially if hacking is what you want anyway, this approach is much cheaper and more flexible (not to mention more fun).
Re:Still can't figure out how it works...
on
Optical Camouflage
·
· Score: 5, Funny
> why doesn't the image show on the guys hands and face.
> Sure, it takes a million years (well, hours) to upload the program
What do you mean, "Sure"? That "Sure" is the key that makes video swapping on the ReplayTV currently a gimmick. Even on a cable modem uploading a 600MB file to someone takes forever (mine only does about 128Kbps upstream), making the entire notion of show swapping purely academic (pun in hindsight not intended). It's not the home users on a cable or 56K modem that pose any file swapping danger, it's the university dorm guys living on multiple T1s. And compared to the numbers of households their numbers are peanuts. They only mentioned Napster because it's the Evil-Consumer trait du jour.
Reaktor by Native Instruments. It's got incredible analog synth support and a very extensible architecture. Since it does full sound modeling, you can create your own custom instruments from scratch. Not the cheapest around, but if you search Usenet, you'll find lots of user praise for it.
> airships can be filled with more things than a lighter than air gas, including heated air.
Or, they could even be--ahem--"filled" with vacuum, resulting in the best possible lift. Of course, the skin with the necessary strength to contain that sort of external air pressure is still mainly a product of Neil Stephenson's imagination.
You didn't originally say series, you said stories. Just about any novel can be published as a series in a paper or magazine (which was indeed very popular at that time), but that doesn't make it a collection of short stories.
> Spielberg, who seems to like sci-fi, but can't help tossing in toys, gimmicks
And kids. Too goddamn many cutsey kids with big gleaming eyes and squeaky voices. I watched the second episode of Taken last night, and while it's not too bad, I almost had to hurt something when I heard the voice of the narrator. My GOD, who in their sensible mind would consider that sort of sound desireable???
> I have to say that the Foundation series is without > a doubt my favorite sci-fi series of all time.
Same here. It's not too often that a scientist is also a great storyteller and writer. That's when you end up with a bit more substance and depth than The Force.
> If someone with the dedication and passion for the story got > their hands on Foundation, like Jackson has with LOTR
That's exactly what I was thinking. Maybe, just maybe Jackson is also a huge Asimov fan. Imagine the entire Foundation series being done in a similarly uncompromising fashion, that would be unreal. To keep costs down, I wouldn't even mind if tons of CGI and unknown actors (now THAT I wouldn't mind) were used, as long as the story were told properly. Heck, Babylon 5 told a very decent story on a TV series budget.
Re:Looks like a "3rd World country" is beating US
on
239 MPG Car
·
· Score: 2
> US GDP: $10 trillion > German GDP: $2.2 trillion > Ratio: 4.5:1
Sorry, my mistake. Still, maybe the OP should define "vast difference." While I whole-heartedly agree that the Germany economy absolutely sucks at the moment (and has sucked for the last few years), there are some mitigating factors, the primary one being the drain on resources of unification. That is only worsed by the deep economic inertia and the general German incapacity for economic change and reform.
The US military leaving Germany might actually be extremely beneficial to the German economy. It would hopefully create the need in the heads of policitians to beef up defence, thus strengthening the domestic defence industry. Traditionally (certainly in the US) that's been one of the driving forces behind lots of high-tech R&D (insert obligatory Nazi references here), and that's something Germany has been neglecting for a LONG time. While they certainly do have some strong players in the defence field, they're hampered by Germany's traditional restraint on defence trade (headline stories of German chemical technology in Lybia and Iraq notwithstanding). If they can't sell submarines to Taiwan and tanks to Turkey, guess who will?
Re:Looks like a "3rd World country" is beating US
on
239 MPG Car
·
· Score: 2
> the vast difference in GDP
Huh? About a factor of three, which incidentally is also roughly the ratio of population sizes. That's why we usually work with percentages rather than raw currency amounts, in which case you'd see that the US spends about 16% on defense while Germany spends about 10%.
Can't say I share that experience. I've used ZA and the Cisco VPN client for several years and they've never been a source of instability on NT/2K, because I haven't had any instability.
It seems their rendering of the letter R is weak, since SPHERES also contains a P instead of an R, and in both of those instances that P seems to have a small imperfection where the diagonal line of the R would join the arc.
Slashdot Mirror
> You're talking Object Pascal as opposed to pascal.
Borland Pascal introduced object features in the later versions of the language. Not the same as Delphi, but it laid the groundwork.
> It is possible to implement multiple-inheritance classes in pascal
Example?
> Turbo pascal isn't pascal any more than Turbo C is C
Huh? Be specific.
> This is why the single-inheritance schema was chosen for the VCL.
No, the Delphi compiler and the VCL were developed concurrently, adding features to the compiler (and language) as the need for them arose in the VCL. The design of the VCL was not a consequence of an existing compiler. It is the way it is because the designers WANTED it that way. They came from the same school of thought as the Java designers (in fact, Borland worked closely with Sun on Java), and while cynics like to think that they eschewed multiple inheritance because it's hard to implement (which it is), there's a lot more to that decision than that. After all, they already had a market leading C++ compiler, so they had the expertise. If there is any afterthought in there at all, it's C++ Builder. After much crying and whining from the community (myself included) over the choice of Pascal instead of C++, Borland gave in. Never mind that many of them learned to love OP on its own merits.
> I should have mentioned that not all server-centric apps need be browser-based
Of course not, I should know. I write mainly C/S apps, RPC style (and nowadays, web service-style). In that case you have two apps, one on the desktop, one remote. And both can still benefit handsomely from components--database components on the backend, GUI and communications components on the frontend. It still doesn't signal a move away from components. Face it, if there is one positive upshot of the whole OOP movement, it's component programming. Ok, I read that somewhere in DDJ, but it's true.
There's lots of noise in the reviews, but you can still filter out a useful signal. Many reviews are also obviously dumb, such as the guy that ordered a CD player, loved it, but decided he didn't need it after all, so he gave it one star. Brains entirely optional. After reading through ten or twenty reviews of the CD player, you'd find recurring mentions of good or bad battery life, skipping/no skipping etc, which is usually all you need.
> There's no real reason why it couldn't now be implemented as a multiple-inheritance class
> structure. The reason for single-class inheritance is historical.
I don't know where you're getting your facts. Object Pascal was never intended as a multiple-inheritance language, it had nothing to do with Borland being unable to implement MI. Borland's C++ compiler did MI capably long before Delphi came about. Delphi's single inheritance is a school-of-thought thing. Heijlsberg designed Delphi and the VCL to be THE end product, not a transition product on the way to the REAL C++ product. He was always a Pascal guy, ever since he created TP 1. They only reused the VCL in C++ Builder because it already existed and took less work than rewriting it from scratch and using a different inheritance model.
> However, with the advent of server-centric apps, there's not much need for "visual components",
Well, we definitely have different points of view on that. I don't think the browser will become THE platform anytime soon, at least not with current technologies. Relatively few types of applications can be expressed well in a stateless Form paradigm with the bulk of the code executing remotely. A rich user experience will ALWAYS require either lots of client-side code and components, or a very fat pipe to do X-type remoting. "Thin client" computing (as if!) looks good on paper and management loves it for all the obvious reasons, but it will never completely replace client side apps. Some things will always make more sense in a browser and others on the desktop.
Besides, even for server side programming the component approach is gaining hold. Just look at Microsoft's ASP.NET and Borland's web components. Face it, intermingled HTML-and-script programming is a major kludge and in no way the wave of the future. The web didn't all of a sudden make proven concepts such as separation of interface and logic obsolete. It's just that the technologies were still in their infancy and were often driven by neophytes that knew Perl and HTML and thought they ruled the world.
> These are just IDE enhancements.
Just? Considering that, after the class library, the IDE is what makes you most productive, that's hardly a minor thing. Sure, you could buy a third party code editor, but that's yet more money, and most people won't.
> hacks caused by the use of single-parent-hierarchy class structure
Well, it sounds like you have issues with the language itself then. Perhaps Delphi isn't for you after all.
> trying to be all things to all people
Borland's biggest problem is that most people aren't using Delphi, not inherent problems in the product itself. Maybe it's something they can't do much about, since it's hard to blame someone for failing to compete well with Microsoft. As a product, I fail to see anything majorly wrong with Delphi, especially when compared to Microsoft's offerings.
> I didn't see anything in it that would make it worth my while
There are three things in particular in 4+ that I can't live without anymore: CTRL-SHIFT-C to create method bodies for all method declarations in a class (quite tedious to do manually if you don't have to), CTRL-SHIFT-Up/Down to switch between method declaration and definition, and CTRL-Click on an identifier to take you to its declaration. Every time I have to go to pre-4 I feel naked without these aids.
> I stopped using Borland products after Delphi 3 (the newer ones didn't offer that much more...)
You stopped at the wrong version. Many useful features were added in 4 and properly debugged in 5. If you use Professional, version 5 is your best bet. There are lots of IDE improvements (in particular navigation and code completion), plus forms are stored as text. I don't think 6 and 7 added much value to Professional, mainly lots of Enterprise and Web stuff.
No something, that's exactly what it means.
> While work per cycles is important, work per Watt
> is even more important in embedded systems.
Not really, unless you're talking battery operated, or are dealing with extreme heat generation. Judging by many embedded systems nowadays (e.g. set-top boxes from cable providers), heat generation IS an issue, but you didn't specifically address that.
Besides, I was being very generous comparing the 250 MHz PPC to the 1 GHz Duron, because they're certainly not in the same performance class. Don't forget that the Duron itself has a work-per-cycle advantage over Intel, so it's no slouch either. Plus, while I don't know too much about the PPC, it's still an old school RISC design if I'm not wrong, meaning that it has fewer complex instructions and might require more of them than an x86 to accomplish some things (fp math/fancy addressing).
> The cost of this product comes in the engineering work that
> went into designing a custom PCB and layout for this product.
That's right, and that's why it might not make much sense for most of the applications people are salivating over here. The cheap (but potentially hotter, though still fanless) Via Epia boards are perfect for these geek-type "embedded" applications, such as MP3 and MP4/DivX players (though the jury is still out on whether the C3 processor has quite enough juice for the latter). The extra heat and power won't matter much, but the very substantial cost savings will.
> if you really think that clock speed is all that counts
Obviously you're one who thinks that the performance of PowerPC expands to beat any competitor at any given time. While the x86 and PPC are not comparable MHz for MHz, you can certainly figure out how much work each one does per cycle. Just for comparison, a dirt cheap throw away CPU nowadays would be a 1GHz Duron (around $35), which runs four times faster than this particular PPC. The PPC would have to execute four times more instructions per cycle than the Duron to be roughly comparable, and you can bet your life that it doesn't. So, all in all, while the PPC might be more efficient per cycle, it is not per dollar. IOW, for any given dollar amount you can buy considerably more raw processing power in an x86.
For example, check out the Epia boards from Via. For around $100 you get roughly the same processing power plus all conceivable peripherals on board. Add a $30 graphical LCD from Crystalfontz, get (or make) a suitable case, and you're set for WAY less than $500. Especially if hacking is what you want anyway, this approach is much cheaper and more flexible (not to mention more fun).
> why doesn't the image show on the guys hands and face.
What guy?
> Sure, it takes a million years (well, hours) to upload the program
What do you mean, "Sure"? That "Sure" is the key that makes video swapping on the ReplayTV currently a gimmick. Even on a cable modem uploading a 600MB file to someone takes forever (mine only does about 128Kbps upstream), making the entire notion of show swapping purely academic (pun in hindsight not intended). It's not the home users on a cable or 56K modem that pose any file swapping danger, it's the university dorm guys living on multiple T1s. And compared to the numbers of households their numbers are peanuts. They only mentioned Napster because it's the Evil-Consumer trait du jour.
Reaktor by Native Instruments. It's got incredible analog synth support and a very extensible architecture. Since it does full sound modeling, you can create your own custom instruments from scratch. Not the cheapest around, but if you search Usenet, you'll find lots of user praise for it.
> airships can be filled with more things than a lighter than air gas, including heated air.
Or, they could even be--ahem--"filled" with vacuum, resulting in the best possible lift. Of course, the skin with the necessary strength to contain that sort of external air pressure is still mainly a product of Neil Stephenson's imagination.
Although for a while Borland was trying to pull precisely this trick with their development tools.
> Depends on how you define 'short series'.
You didn't originally say series, you said stories. Just about any novel can be published as a series in a paper or magazine (which was indeed very popular at that time), but that doesn't make it a collection of short stories.
> Spielberg, who seems to like sci-fi, but can't help tossing in toys, gimmicks
And kids. Too goddamn many cutsey kids with big gleaming eyes and squeaky voices. I watched the second episode of Taken last night, and while it's not too bad, I almost had to hurt something when I heard the voice of the narrator. My GOD, who in their sensible mind would consider that sort of sound desireable???
> Perhaps "Foundation" is waiting for Peter Jackson to finish with the LOTR series
Hear, hear, just in case he's reading Slashdot.
> The first books are short stories collections,
No, they're not. You must be thinking of some of his other books, like the robot stories.
> I have to say that the Foundation series is without
> a doubt my favorite sci-fi series of all time.
Same here. It's not too often that a scientist is also a great storyteller and writer. That's when you end up with a bit more substance and depth than The Force.
> If someone with the dedication and passion for the story got
> their hands on Foundation, like Jackson has with LOTR
That's exactly what I was thinking. Maybe, just maybe Jackson is also a huge Asimov fan. Imagine the entire Foundation series being done in a similarly uncompromising fashion, that would be unreal. To keep costs down, I wouldn't even mind if tons of CGI and unknown actors (now THAT I wouldn't mind) were used, as long as the story were told properly. Heck, Babylon 5 told a very decent story on a TV series budget.
> US GDP: $10 trillion
> German GDP: $2.2 trillion
> Ratio: 4.5:1
Sorry, my mistake. Still, maybe the OP should define "vast difference." While I whole-heartedly agree that the Germany economy absolutely sucks at the moment (and has sucked for the last few years), there are some mitigating factors, the primary one being the drain on resources of unification. That is only worsed by the deep economic inertia and the general German incapacity for economic change and reform.
The US military leaving Germany might actually be extremely beneficial to the German economy. It would hopefully create the need in the heads of policitians to beef up defence, thus strengthening the domestic defence industry. Traditionally (certainly in the US) that's been one of the driving forces behind lots of high-tech R&D (insert obligatory Nazi references here), and that's something Germany has been neglecting for a LONG time. While they certainly do have some strong players in the defence field, they're hampered by Germany's traditional restraint on defence trade (headline stories of German chemical technology in Lybia and Iraq notwithstanding). If they can't sell submarines to Taiwan and tanks to Turkey, guess who will?
> the vast difference in GDP
Huh? About a factor of three, which incidentally is also roughly the ratio of population sizes. That's why we usually work with percentages rather than raw currency amounts, in which case you'd see that the US spends about 16% on defense while Germany spends about 10%.
Can't say I share that experience. I've used ZA and the Cisco VPN client for several years and they've never been a source of instability on NT/2K, because I haven't had any instability.
It seems their rendering of the letter R is weak, since SPHERES also contains a P instead of an R, and in both of those instances that P seems to have a small imperfection where the diagonal line of the R would join the arc.
> will be less splattered body matter in Solaris, and less
> people pulling their eyes out of their sockets.
Yeah, they will be doing that AFTER the movie, when realizing what a waste of time it was.