To think that Anders and 34 employees would go to microsoft and leave the compiler they loved, and grew up with, just shows you how low humans will go.
Maybe they could go like IBM and sell "consulting" and "database expertise". Or "cluster consulting".
(if anything ever happened to take them out of the search business, as IBM got taken out of the OS business)
Hey but no one says we have to have an open source WWW based education system. It could be offline.
And all your points above seem to remind me of "CVS".. i.e. isn't CVS adn email in theory an inefficient way of coding software projects, versus being in a building together? If everyone was in the same city and building working on software, wouldn't it be easier to communicate?
But in practice, open source software still works with CVS and e-mail, even if the programmers are not setup in the same classroom or building together. Compare to exchanging floppies or CD's in the same room, pair programming, etc.). Those have disadvantages to (i.e. inefficiencies of getting to the building, transportation, lunch breaks not convenient).
I experimented with a "jeopardy" type program that was running on IRC once. Using google I had about a 95 percent average getting hundreds of thousands of questions correct. The time limit for answering each question was also strict (something like 6 seconds?), and this helped motivate me. Of course a DSL/Cable connection was essential.
Stuff that I never even knew about was easy to answer, only if google was available at my side (and with the StayOnTop held on for IRC). I played for about 2 hours getting 95 percent of questions correct.. on stuff I never had the slightest clue about.
It was interesting, but the sad question I asked myself was "why do I need to know any of this? Even if I can get 95 percent right, who's going to care?". It was manly history based questions.. as in jeopardy style. Tough ones too, bizarre things you wouldn't know.
I just used precise quotations around phrases. This was essential: knowing how to quote phrases.
Specific questions in a jeopardy like situation are also easier to find on google since they are so specific. Whereas finding information about the latest popular Sony Widget is not so specific. It may seem specific if you have a part number, but it's not as specific as a question that's 5-6 words or so. So much harder to find valid information on a Sony product than a Jeopardy-like-questoin, from my experience.
I also don't watch Jeopardy or like to play those type games.. so I was surprised to have a 95 percent average. And I never liked history either. It's the future. It's all about the future.
Now imagine how easy it would be to answer exam questions (or questions from your wife) if people had small computers embedded in their ear with "google embedded" installed on them?
p.s. what is the meaning of life, google? I think the answer was 42?
Interesting ideas. That's a lot of work for the standard user but the programmer may attempt it. You could also use something like a spreadsheet or database viewer program that has an HTTP query feature. Some spreadsheets let you view, sort, and display data right inside the software without ever touching your web browser. But how do you actually visit and click the links from there on? Link enabled software?
How 'bout some web browsers with better widgets built in to them? Gridsheets, sorting features, filtering features.. Mozilla may have some stuff for filtering it seems, but I haven't checked into it too much since mozilla I find slow:-( <insert ComponentBasedBrowser plug here>
You could also parse the google results via an online web application on a server, which filtered your results with a CGI regex or PHP regex of your choice (post-processing) after getting the results. IN fact, save the regex, since they can be slow.. Google's data is fairly consistent (rarely change the website) so a parser would probably be better (especially tables, since all google results are just a bunch of tables stacked one after the other).
Or better yet, a thin client built specifically for google, (non browser based) who places results into a software application which is link enabled (fires open your web browser). <insert ComponentBasedBrowser plug here>
Library stats are outdated. For example, go to any library and find me recent Google statistics. She (the libarian) will probably point you to a computer with internet access.
To get KOL working in freepascal you have to go through hell. There are no examples of it in use, and the instructions to get it working on KOL assume you have all the time in the world. That's something that needs to be worked on. I might get it working some time and provide for people...
Examples of using KOL in FPC are needed, and proper installtion downloads for KOL with FPC are needed.
For linux, I'm not sure what is out there to use for light pascal GUI development. I'm sticking to developing non-gui server/client stuff for linux right now, because in my opinion GTK applications currently look bad and hog up way too much screen space. No customization, everything feels fixed. Of course, this could be changed, and I may have just not experimented enough. It's just sad that the win32 version of lazarus looks better, operates better, and takes up less screen space than the GTK based GnuLinux Lazarus.
There was a KOL for kylix, somebody said. If there is, we need some examples of it in use. Then the Kylix stuff could maybe be used in Lazarus some time, or FPC.
You could use pointers when there were no dynamic arrays. But all the latest compilers support dynamic arrays.
The reason there weren't dynamic arrays in the early days was more due to the fact that not as many people were extending the language helping out with creating a compiler that could utilize dynamic arrays.
The language has progressed and you are spreading myths around slashdot that were only true in the 1980's. You sound like Bryan K when he made an ass of himself writing an article called "Why Pascal isn't my favorite Language, 1981".
If you need to create variables on the fly and you don't want to scroll around looking for them:
1. Make a comments. Comment out a variable in your code showing that your variable is declared elsewhere. I.e. Let's say you have some private and public variables declared way at the top of your Pas file. Instead of scrolling up all the time, just make some comments for these variables later on in your code, above the function or procedure where you need to use them. i.e.:
{
var
Str1:string;//public
String2:string;//private
} procedure SearchFile(); begin .. end;
2. Use an IDE which lets you view your variable's in a side panel or IDE extension of some sort, while you program elsewhere in the file.
3.Instead of scrolling around and declaring your variables at the top of the file, just declare them in comments first..above the procedure you are working with. Then actually go and hard code them into the appropriate place later, when you are done some programming and ready to compile.
You should try KOL for delphi. It takes a few months to learn the basics of how the OO programming works in KOL, but once you know it.. you can have a 25KB WordPad program and a 20KB notepad program.
Compare that to Notepad exe which is about 60KB.
Then with UPX you'll have a 15KB program. Couldn't say I've seen this done in C or C++, done visually with Object Oriented programming. Someone show us?
Slashdot Mirror
Please consider the Smart Answers article, which was written a while ago.
works in Opera 7
blame the Post Anonymously checkbox.
MS, since they have money.
To think that Anders and 34 employees would go to microsoft and leave the compiler they loved, and grew up with, just shows you how low humans will go.
Maybe they could go like IBM and sell "consulting" and "database expertise". Or "cluster consulting". (if anything ever happened to take them out of the search business, as IBM got taken out of the OS business)
Yes or type in any topic and include "wiki" in your search phrase. Most likely you will find valuable info. Moreso than web forums many times.
And all your points above seem to remind me of "CVS".. i.e. isn't CVS adn email in theory an inefficient way of coding software projects, versus being in a building together? If everyone was in the same city and building working on software, wouldn't it be easier to communicate?
But in practice, open source software still works with CVS and e-mail, even if the programmers are not setup in the same classroom or building together. Compare to exchanging floppies or CD's in the same room, pair programming, etc.). Those have disadvantages to (i.e. inefficiencies of getting to the building, transportation, lunch breaks not convenient).
http://www.google.com/search?hl=en&q=answer+to+lif e+the+universe+and+everything
Stuff that I never even knew about was easy to answer, only if google was available at my side (and with the StayOnTop held on for IRC). I played for about 2 hours getting 95 percent of questions correct.. on stuff I never had the slightest clue about.
It was interesting, but the sad question I asked myself was "why do I need to know any of this? Even if I can get 95 percent right, who's going to care?". It was manly history based questions.. as in jeopardy style. Tough ones too, bizarre things you wouldn't know.
I just used precise quotations around phrases. This was essential: knowing how to quote phrases.
Specific questions in a jeopardy like situation are also easier to find on google since they are so specific. Whereas finding information about the latest popular Sony Widget is not so specific. It may seem specific if you have a part number, but it's not as specific as a question that's 5-6 words or so. So much harder to find valid information on a Sony product than a Jeopardy-like-questoin, from my experience.
I also don't watch Jeopardy or like to play those type games.. so I was surprised to have a 95 percent average. And I never liked history either. It's the future. It's all about the future.
Now imagine how easy it would be to answer exam questions (or questions from your wife) if people had small computers embedded in their ear with "google embedded" installed on them?
p.s. what is the meaning of life, google? I think the answer was 42?
How 'bout some web browsers with better widgets built in to them? Gridsheets, sorting features, filtering features.. Mozilla may have some stuff for filtering it seems, but I haven't checked into it too much since mozilla I find slow :-( <insert ComponentBasedBrowser plug here>
You could also parse the google results via an online web application on a server, which filtered your results with a CGI regex or PHP regex of your choice (post-processing) after getting the results. IN fact, save the regex, since they can be slow.. Google's data is fairly consistent (rarely change the website) so a parser would probably be better (especially tables, since all google results are just a bunch of tables stacked one after the other).
Or better yet, a thin client built specifically for google, (non browser based) who places results into a software application which is link enabled (fires open your web browser). <insert ComponentBasedBrowser plug here>
Library stats are outdated. For example, go to any library and find me recent Google statistics. She (the libarian) will probably point you to a computer with internet access.
Examples of using KOL in FPC are needed, and proper installtion downloads for KOL with FPC are needed.
For linux, I'm not sure what is out there to use for light pascal GUI development. I'm sticking to developing non-gui server/client stuff for linux right now, because in my opinion GTK applications currently look bad and hog up way too much screen space. No customization, everything feels fixed. Of course, this could be changed, and I may have just not experimented enough. It's just sad that the win32 version of lazarus looks better, operates better, and takes up less screen space than the GTK based GnuLinux Lazarus.
There was a KOL for kylix, somebody said. If there is, we need some examples of it in use. Then the Kylix stuff could maybe be used in Lazarus some time, or FPC.
Just change the program keyword to Library, as a start. And whalla - your DLL or .SO is created.
You can use freepascal or delphi to make DLL's that run and are called in C++ applicatoins.
And vice versa. One example is total commander, who can read your plug-ins in C++ DLL's or in Pascal DLL's.
Example:
program
uses
Units;
var
s:string;
end.
library
uses
Units;
var s:string;
end.
All you have to do to get your library started is take a program and change it's program keyword to library.
Which is the idea of freepascal :)
GnuLinux/Win32 ready. AnsiString ready. Pchar ready. And many more.
Unless you like to strain your pinky, typing begin and end is actually faster than using a curly brace.
@Pascal has ^pointers. Where'd you get that idea from?
You could use pointers when there were no dynamic arrays. But all the latest compilers support dynamic arrays. The reason there weren't dynamic arrays in the early days was more due to the fact that not as many people were extending the language helping out with creating a compiler that could utilize dynamic arrays. The language has progressed and you are spreading myths around slashdot that were only true in the 1980's. You sound like Bryan K when he made an ass of himself writing an article called "Why Pascal isn't my favorite Language, 1981".
Also, look into TeX and LaTeX. These are popular GnuLinux tools...and thought up with yours truly, pascal.
1. Make a comments. Comment out a variable in your code showing that your variable is declared elsewhere. I.e. Let's say you have some private and public variables declared way at the top of your Pas file. Instead of scrolling up all the time, just make some comments for these variables later on in your code, above the function or procedure where you need to use them. i.e.:
{ //public //private
..
var
Str1:string;
String2:string;
}
procedure SearchFile();
begin
end;
2. Use an IDE which lets you view your variable's in a side panel or IDE extension of some sort, while you program elsewhere in the file.
3.Instead of scrolling around and declaring your variables at the top of the file, just declare them in comments first..above the procedure you are working with. Then actually go and hard code them into the appropriate place later, when you are done some programming and ready to compile.
You should try KOL for delphi. It takes a few months to learn the basics of how the OO programming works in KOL, but once you know it.. you can have a 25KB WordPad program and a 20KB notepad program. Compare that to Notepad exe which is about 60KB. Then with UPX you'll have a 15KB program. Couldn't say I've seen this done in C or C++, done visually with Object Oriented programming. Someone show us?
Both. You can mix regular pascal with object pascal. Or just use one alone.
Try Total Commander on Win32. Or Seksi Commander on Linux