I guess they made it more Java-like. Like with the Java 2DAPI, where can I find IDirectDraw::WaitForVerticalRetract()? Where can I find ScrollWindowEx()? The Windows API has neat stuff because MS wanted game developers to switch from DOS to Windows, and key elements are unavailable anywhere else.
Having taken Brinch Hansen's data-structures course at Caltech more years ago than I care to admit and having been indoctrinated by Brinch Hansen in the Pascal religion, I always thought that a properly restrictive language could help a lot with reliability.
Ada was supposed to be that silver bullet (although Pascal diehards have their issues with Ada), but the darned trouble is that Ada is not showing a large-enough (maybe 5 percent) quantifiable improvement over C++.
Java is another of these silver bullets, and the claim is that people churn out a lot more stuff, but I have not heard about reliability.
Maybe it is how you structure a design and the code implementation is more important than the hand holding (or hand slapping) of a particular language.
Celine Dion is a singer from Canada. Slashdot just had a story that Sony was releasing her latest album with a particularly malicious copy protection. The postings were divided between "Sony is exposing itself to liability for using copy protection that could harm your firmware in your drive" to "Who cares? Any scheme to prevent us from listening to Celine Dion is a Good Thing."
Much of the "content" that the recording industry is worried about is schlock, and I took the poster who wanted to play Celine in the car as expressing that viewpoint through clever sarcasm.
How about this for a kludge.
Borland Kylix compiles to native Linux using the Qt libraries. Qt is C++, and it has its own layer on top of C++ (Signals and Slots and the MOC), so Kylix cannot link directly to Qt. Kylix links to a C-language wrapper (looks an awful like Gnome/GTK way of faking objects) to the Qt C++ classes, and this wrapper library was generated automatically using some kind of text editing program in an effort to make sure stuff wasn't left out.
But thats not all! The Kylix IDE is not (yet) written in Kylix -- it is a Windows app, and it runs on Linux using WINE, and now you know the rest of the story.
Can you give a reference -- Web site, study, anything? The current run-out-of-oil scenario is the bell-curve resource depletion model (do a Google search on K. Humbert, a USGS geologist who predicted peak of US production for 1970's). When you are on the back side of world oil production, the stuff won't be had for love, war or money.
A more optimistic scenario is that if we are willing to pay more (accept slightly lower grade deposits), there is a lot more to be had. Abundant $33/bbl oil means energy is not a problem (I can easily drive my Taurus on $1.50/gal gas) -- the only sticking point would be CO2 emissions.
Please fill me in on sedimentary oil geology.
The Russians could have pulled it off, but they lacked test stands where they could have seen if the darned thing ran before trying to fly it.
Let this be a lesson to software developers.
1. I surfed Web sites on Ada evangalism, and I think they have a hard time quantifying an improvement over C++ -- either C++ is not that bad or programmers manage to make mistakes in Ada.
2. I get the occasional GP fault in Borland Object Pascal. Most common is accessing nil object references -- they should do some kind of Objective C/Smalltalk nil object type to at least throw a meaningful exception. Forgeting to invoke default constructors and destructors is another big one -- maybe C++ has the right idea of automatically invoking constructors and destructors on base classes. More infrequent but tremendous time waster is Windows itself. I was resizing a window I hadn't done Show() on -- gave a crash in Windows 95 but not Windows 98 and took forever to track down. There is still stuff with palatte management that is the source of annoying performance bugs and screen-doesn't-look-right.
If it weren't for Da Shuttle, we could have had Moon bases by now. The Saturn V could take crews and payloads to the Moon -- Shuttle can barely make low-Earth orbit. Saturn launches probably run a billion dollars each, but each Shuttle launch runs a cool half billion, depending on who is doing your accounting. Besides, the Saturns were already designed while with the Shuttle they had to sink in several billion to get it going.
Budgeting, say 3 billion a year, doing 3 launches a year to the Moon, by now you could have had over 30 years tons and tons of stuff delivered to the lunar surface. Instead, this same money was pissed away on the Shuttle and the stupid space station.
Has anyone out there, ever, on a board-level system, or something you get from Dell or Compaq, even seen a desktop system that hibernates (the S3 level?)? I believe the hibernate mode along with all the other capabilities of ACPI are all vaporware.
Theoretically, a hibernating machine should use less watts than even a telephone answering machine, and I would like to have answering and FAX capability on the promised 2 watts -- I am one of these power conservation fanatics.
The best I can do is S2-standby, and even that has bugs -- can't use a serial mouse with it, and can't print after coming out of standby without rebooting.
Don't know if this will change, but the.NET SDK with a command-line C# compiler is part of the Beta (free).NET runtime, and I have written "Hello world" programs in both console and GUI mode.
My SOURCE codes are copy protected because they are written in Object Pascal. The unwashed masses can't crack 'em because they don't program while persons savy enough to make sense of the code will sniff that anything not in C/C++/Java/Perl/CLisp is not worth bothering to read.
GNOME will always be GNOME, and while you are creating a CLR/Mono binding to it, and if CLR/Mono runs contrary to a person's aesthetic sensibilities, they can still program GNOME in good old ANSI C, now and in the forseeable future.
By the way, I have looked at the Mono web site, and as a Windows developer, I am not quite sure what I have to download in what order to get the Mono C# and class libraries going, but I will keep checking the site as things develop.
Having tried C#/.NET at the command line, the performance hit over C++ is maybe 2-3 (18 months of Moore) instead of 5-10 (about 5 years).
Given that performance is not a show-stopper anymore and given that Managed Software (class library at OS level, GC, runtime checks) is the Next Thing (hey, there was a time when we though C was too much a layer over assembly language), your choices are Java or CLI/CLR.
Java has some nice stuff to it -- friendly documentation at the Sun site compared to that gibberish that passes for documentation at MS, a nice software-engineered feel instead of that steaming pile of stuff that makes up an MS API (I develop for MS API's). But Java is Java and Sun is Sun, and you have to take the whole thing or leave it.
Since MS has flopped this "CLR/CLI/.NET" standard out there, it really there for the implementing. Oh, the Borg we hear, we are about to get assimilated into the Collective.
My understanding is that the effort is not simply to try to clone.NET but to implement an Open Source managed software thingy, and if it forks from MS, who cares. MS can have all the proprietary extensions it wants and we can have our own extensions.
Why not clone Java? Sun won't let you. Why not invent our own managed software thingy? We could, but there is one already out there.
My personal pantheon (in this order)
1. Guy who invented Novacaine for dental procedures
2. Thomas Crapper -- inventor of indoor plumbing
3. Norman Borlaug -- developer of high-yield
cereal grains (he is supported on foundation
money, the robber barons (i.e. Gates's) of
last century, so his stuff is "Open Source")
4. Chester Carlson -- inventor of Xerox photocopy
process (I love my laser printer -- THAT is
the best thing since Guttenberg)
Slashdot Mirror
I guess they made it more Java-like. Like with the Java 2DAPI, where can I find IDirectDraw::WaitForVerticalRetract()? Where can I find ScrollWindowEx()? The Windows API has neat stuff because MS wanted game developers to switch from DOS to Windows, and key elements are unavailable anywhere else.
Ada was supposed to be that silver bullet (although Pascal diehards have their issues with Ada), but the darned trouble is that Ada is not showing a large-enough (maybe 5 percent) quantifiable improvement over C++.
Java is another of these silver bullets, and the claim is that people churn out a lot more stuff, but I have not heard about reliability.
Maybe it is how you structure a design and the code implementation is more important than the hand holding (or hand slapping) of a particular language.
Much of the "content" that the recording industry is worried about is schlock, and I took the poster who wanted to play Celine in the car as expressing that viewpoint through clever sarcasm.
How about this for a kludge. Borland Kylix compiles to native Linux using the Qt libraries. Qt is C++, and it has its own layer on top of C++ (Signals and Slots and the MOC), so Kylix cannot link directly to Qt. Kylix links to a C-language wrapper (looks an awful like Gnome/GTK way of faking objects) to the Qt C++ classes, and this wrapper library was generated automatically using some kind of text editing program in an effort to make sure stuff wasn't left out. But thats not all! The Kylix IDE is not (yet) written in Kylix -- it is a Windows app, and it runs on Linux using WINE, and now you know the rest of the story.
Can you give a reference -- Web site, study, anything? The current run-out-of-oil scenario is the bell-curve resource depletion model (do a Google search on K. Humbert, a USGS geologist who predicted peak of US production for 1970's). When you are on the back side of world oil production, the stuff won't be had for love, war or money. A more optimistic scenario is that if we are willing to pay more (accept slightly lower grade deposits), there is a lot more to be had. Abundant $33/bbl oil means energy is not a problem (I can easily drive my Taurus on $1.50/gal gas) -- the only sticking point would be CO2 emissions. Please fill me in on sedimentary oil geology.
The Russians could have pulled it off, but they lacked test stands where they could have seen if the darned thing ran before trying to fly it. Let this be a lesson to software developers.
1. I surfed Web sites on Ada evangalism, and I think they have a hard time quantifying an improvement over C++ -- either C++ is not that bad or programmers manage to make mistakes in Ada. 2. I get the occasional GP fault in Borland Object Pascal. Most common is accessing nil object references -- they should do some kind of Objective C/Smalltalk nil object type to at least throw a meaningful exception. Forgeting to invoke default constructors and destructors is another big one -- maybe C++ has the right idea of automatically invoking constructors and destructors on base classes. More infrequent but tremendous time waster is Windows itself. I was resizing a window I hadn't done Show() on -- gave a crash in Windows 95 but not Windows 98 and took forever to track down. There is still stuff with palatte management that is the source of annoying performance bugs and screen-doesn't-look-right.
If it weren't for Da Shuttle, we could have had Moon bases by now. The Saturn V could take crews and payloads to the Moon -- Shuttle can barely make low-Earth orbit. Saturn launches probably run a billion dollars each, but each Shuttle launch runs a cool half billion, depending on who is doing your accounting. Besides, the Saturns were already designed while with the Shuttle they had to sink in several billion to get it going. Budgeting, say 3 billion a year, doing 3 launches a year to the Moon, by now you could have had over 30 years tons and tons of stuff delivered to the lunar surface. Instead, this same money was pissed away on the Shuttle and the stupid space station.
Yes, and if the Kernel were written in Borland Object Pascal, you could get a 23 second compile on a Celeron 500.
The way Office blows chunks, it really looks like thats the case.
Has anyone out there, ever, on a board-level system, or something you get from Dell or Compaq, even seen a desktop system that hibernates (the S3 level?)? I believe the hibernate mode along with all the other capabilities of ACPI are all vaporware. Theoretically, a hibernating machine should use less watts than even a telephone answering machine, and I would like to have answering and FAX capability on the promised 2 watts -- I am one of these power conservation fanatics. The best I can do is S2-standby, and even that has bugs -- can't use a serial mouse with it, and can't print after coming out of standby without rebooting.
He may have misspoke, but Microsoft is going to pull Windows when pigs to to the skies.
Don't know if this will change, but the .NET SDK with a command-line C# compiler is part of the Beta (free) .NET runtime, and I have written "Hello world" programs in both console and GUI mode.
My SOURCE codes are copy protected because they are written in Object Pascal. The unwashed masses can't crack 'em because they don't program while persons savy enough to make sense of the code will sniff that anything not in C/C++/Java/Perl/CLisp is not worth bothering to read.
GNOME will always be GNOME, and while you are creating a CLR/Mono binding to it, and if CLR/Mono runs contrary to a person's aesthetic sensibilities, they can still program GNOME in good old ANSI C, now and in the forseeable future.
By the way, I have looked at the Mono web site, and as a Windows developer, I am not quite sure what I have to download in what order to get the Mono C# and class libraries going, but I will keep checking the site as things develop.
Given that performance is not a show-stopper anymore and given that Managed Software (class library at OS level, GC, runtime checks) is the Next Thing (hey, there was a time when we though C was too much a layer over assembly language), your choices are Java or CLI/CLR.
Java has some nice stuff to it -- friendly documentation at the Sun site compared to that gibberish that passes for documentation at MS, a nice software-engineered feel instead of that steaming pile of stuff that makes up an MS API (I develop for MS API's). But Java is Java and Sun is Sun, and you have to take the whole thing or leave it.
Since MS has flopped this "CLR/CLI/.NET" standard out there, it really there for the implementing. Oh, the Borg we hear, we are about to get assimilated into the Collective.
My understanding is that the effort is not simply to try to clone .NET but to implement an Open Source managed software thingy, and if it forks from MS, who cares. MS can have all the proprietary extensions it wants and we can have our own extensions.
Why not clone Java? Sun won't let you. Why not invent our own managed software thingy? We could, but there is one already out there.
My personal pantheon (in this order) 1. Guy who invented Novacaine for dental procedures 2. Thomas Crapper -- inventor of indoor plumbing 3. Norman Borlaug -- developer of high-yield cereal grains (he is supported on foundation money, the robber barons (i.e. Gates's) of last century, so his stuff is "Open Source") 4. Chester Carlson -- inventor of Xerox photocopy process (I love my laser printer -- THAT is the best thing since Guttenberg)