What's the HARDEST program you've written? - Page 2

[Natso]

A MMOG engine.
It had a global view of all players & you could only see the people on your own 'map'... there was a limit of 14,000 maps or something like that... 256X256.
The neat part about it was that you could have people online playing & somebody could edit pat #24 & EVERYBODY would immidiatly see the new mat 24, the old one would be history! It allowed moderators to build the world without the players having to redownload the program itself.
However it kinda went outta date after I never added any goal to it whatsoever.
<It was made with game maker... pretty good too seeing as it ran at its full speed without much lag at all!>

[Toba]

Well, the hardest thing I've ever made would have to be the programming language I invented in April this year. It's a very low-level two-dimensional progamming language for DOS that compiles to around 3K for most programs. It can do self-modifying code, PUT/GET in it's little world of two-dimensional memory, etc. I made an interpreter first in VB, but it became difficult to maintain and buggy (spaghetti code doesn't even begin to describe that mess), so I decided to make an inline interpreter in assembler. The compiler is relatively simple (it opens the assembled inline interpreter, inserts the code, and saves as a .COM file), but the inline interpreter is a real challenge to work on. If you want to try it out, I decided to make it freeware: http://www.parabolagames.com/?page=catalog&p=twirq_1_0_0

If you try it out, tell me what you think :) .

[K-1]

Hmm. my hardest program is probably a recursive binary tree program in C++, that could be sorted with a few different algorithms. Nothing too hard really, but I have had some large pain in the ass "Projects" that I wouldn't call programs.

[Chainsaw]

I worked for Intuit for four years working on QuickBooks. Man, that was a huge program! And the hard part was that they had so many users and the new versions had to be better/cleaner/faster than the older versions. When the app loaded in Windows, 135 DLLs would load! about 50 of those were MS dlls, of course, but still...

[Puckdropper]

Hardest project (Program?) I have had so far is to write a package in Ada using and manipulating strings. The problem was, rather than using get_line for the input, we had to use get. The computer had no idea how long the real string was, it only knew it was dimensioned to be 80 characters.

Well, I put a nice interface on it, and hoped for the best...

[bentkey]

The hardest thing I ever wrote was just a fun project. It was intended to be a screen saver and was written to run under dos of all things! At the time 640x480 resolution was HiRes and nobody did any real 3d graphics mapping in that res. My screen saver was a 3 dimensional cube that bounces inside of a 3 dimensional box. The cube has pictures mapped on it's sides which stay in proportion as it bounces, spins and goes toward and away. The box changes in size also. It is written in Borland C and the innermost operations are in Turbo Assembler. The code had to be optimized down to the very last processor instruction in order to get it to run on the machines of the day. The last version was written using the original Win G library that MS put out with windows 3.1. heheh Anyone remember that? After I finished it, I put it in the drawer and its remained there ever since. Reading this thread made me remember it. If anyone is curious, I'm going to post a copy as an attachement. I realize, no one outside the house has ever seen it until now. You have to set your screen color to 256 or it won't run.

[Alcides]

This has been pretty inactive for a couple of weeks, but here I go...

The hardest thing I've written was not really hard, but tedious. In some of my spare time I work with numbers for different reasons and thought it would be convenient to write something to the effect of a spreadsheet only not a spreadsheet. There are math rules and extra rules. I think I have it around somewhere. I hardly use it like I used to.

Example... If I collect the data of select stars and planets and then want to have a list of furthest from and closest to certain positions or have them listed by apparent magnitude and apparent luminosity side by side. Sometimes I just want to calculate how many hours my boss is allotted to schedule me and my coworkers. The thing is that I wanted one program where I create a function I want to perfom and set rules for its presentation and come back to it whenever I want. Like sometimes I want to add extra variables like date or position or whatever the case may require. Anyway. I had most of the trouble with the flow chart because of the complexity I wanted and the writing was a little tricky cause I typo a lot. Damn it! :) Anyway.

That was the hardest one done. Right now I'm working on a toy encryption program that creates keys and such and such. I'm still working out the general theory and some of the flow for cetain aspects. Fun stuff, but also very much a headache.

Alcides.

[cosi]

I see a wonderful mix of hard programs here. :)

My hardest would have to be work on creating a low-level type-safe language similar in spirit, but more expressive and powerful than Microsoft's Vault Language.

The key to this language was that you could annotate code with a set of invariants that each function and data structure must maintain throughout the program. These invariants are statically checked at compile time; they incur no run-time cost, and yet, if you compiled the code, you were guaranteed that the code would abide by that set of invariants.

This was great because at compile time you could do things like... prove that your code will never overflow a buffer, and you didn't have to implement the run-time array bounds check that Java does to be safe!!

Best part of this language is that the compiled result was safe C code (though using the AST, we could just as easily generate other languages). You could plug it into any existing system components as safe code.

Here were the drawbacks to the language:
1) Programming required intense expertise---writing a program involved finding a logical progression of constraints.
2) Extra annotated code (30% extra code... though we were working on gettting this number down)
3) Compilation took exponential time. We used the Omega test to solve multiple Pressburger equations for every new constraints... Oh how painful it was when a program didn't compile after 3 hours due to some constraint equation I got wrong!! (Garbage collector: 5 hours distributed compilation on 3 machines. Though there are many tricks at optimation we have yet to try.)

We made a safe firewall and fileserver that in conjunction with an IBM secure coprocessor should be a very secure system.

If you are interested, you can check out the most recent workshop publications at Space 2004.

Ed

[Toba]

Cosi, that's pretty impressive... I've never heard of Pressburger equations or the Omega test (and I'm going into Calculus III this semester). That must be something in Calculus XIV or something like that.

Anyway, that sounds similar to the DOD's language, Ada, which allows coders to mathematically prove that their code is bug free or something like that. Would it be possible to get a copy of your language or is it an in-house specialty that nobody else can use?

[z80-2-Risc]

Wow! What a long thread! I've done the 2-byte program too. I saved the binary to "int ?" something to call the boot ROM - can't remember to a DOS .com file. The 1 byte would be a 0x90 which is handy for padding out hacker shell code but doesn't do much.
Most complicated code was 8086 assembly to run in high memory and process credit card message via SNA/SDLC (LU0). The ABSOLUTELY most complicated thing I've done is what I am doing right now -- trying to install Red Hat in a high partition without moving W2K and coming up with SOME way of booting it! Is there no end to having to learn stuff!

[Pimp Daddy]

Interesting question. In terms of programming coursework at university there was a GUI program called Dress Shop which, using extended user defined classes stored specific information about each item of clothing and used a mixture of both public and private variables. It required full Javadoc which was an absolute nightmare to put together. Another piece of coursework which was an absolute nightmare was a console program based around singly linked lists. To advanced programmers it probably won't sound very hard but I found it so difficult to understand the concept. I understand most of it now but not all of it.

As for personal work there is Longhaul. This is an IDE I'm currently creating which can be used to type, compile, execute and debug Java code. I started typing it in Java but I needed to type so much code I eventually switched to C# and am now using Windows Forms. It's as tough as the pieces of uni coursework that I mentioned because I'm learning complex new concepts on the fly and there are so many different sections to the program. The main editor window is also by far the biggest code file I've ever created, which is not necessarily a good thing!

[nanosani]

The hardest program I wrote was a compiler of C language ( not a full one, but it did have many features) ... and the only reason I did code that was that it was a compiler construction lab assignment ... LOL ...

[mister-fett]

Hardest program I ever wrote...
that would probably be the attached two dimensional two playered game,
written in G, which is NOT an object based language. It took me six hours of jury rigging to set it up, and then I had to tweak out all the bugs and add more weapons, which took an additional forty or so hours. :mrgreen:
Try it! (WARNING: FILE IS FROM 5-10 MB)
www.freewebs.com/netbt

(Edit: ARGH! Freewebs doesnt allow more than 750k uploads! $#%#@$$@#$%#%@(*^(%%^%$36546 :evil: :evil: :evil: )

[cosi]

Cool. Have you seen Eclipse? It's a Java IDE written in Java... well mostly Java. Why did you think that the Java required more code than C#? From my experience, they are roughly equivalent. The only code saving option for projects is finding similar open source projects's code to leverage. Though... the only problem with that is in the licensing of the code. I think Apple has had to reinvent the wheel for a number of open sourced projects (think RPM) because of incompatible licensing.

As for personal work there is Longhaul. This is an IDE I'm currently creating which can be used to type, compile, execute and debug Java code. I started typing it in Java but I needed to type so much code I eventually switched to C# and am now using Windows Forms. It's as tough as the pieces of uni coursework that I mentioned because I'm learning complex new concepts on the fly and there are so many different sections to the program. The main editor window is also by far the biggest code file I've ever created, which is not necessarily a good thing!

[Mahen]

i would like the ASM code of the restart computer program

[Mahen]

i would like to have the code for the restart prg

[z80-2-Risc]

i would like to have the code for the restart prg

Hi Mahen. It has been awhile since I visited - sorry for the wait. The old real-mode DOS interrupt to restart is int 19h:
Run DEBUG from a DOS box, type a (for enter assembly), and enter "int 19".
Press enter again to get out of coding mode and type "d 100"
The dump will give you "CD 19" in the first two bytes. Now the tricky part:

debug has a w (write) instruction that can write to any unprotected disk space. You need to adjust a couple of registers before writing:
enter "n reboot.com"
-- that sets a file name for you
enter "rcx"
-- that displays the cx value (forget that you are on a 32-bit system)
enter "2"
-- that puts a 2 in cx (write size)
enter "w"
-- that [shudder] writes from your current ds:100
Enter "q"
-- that gets you out of that deadly program

There you have it. A 2-byte program that uses real-mode DOS ints.

[Narue]

>What's the HARDEST program you've written?
Operating system. Well, technically I didn't write it, but I did patch it into behaving and that consisted of replacing about 90% of the code.

[jwenting]

tricky...

Once wrote a complete middleware system to link store terminals (running some fancy software we would not be writing) with a Cobol backend we maintained.
Didn't write it alone of course, was part of a team of 3.
During internal beta testing (using a dummy frontend to generate the messages from textfiles) the customer decided they didn't want the product and the subproject was scrapped. But it did work.

Or the most tricky app I wrote that actually made it to production was a C program written under OS/2 to swap the assignments of the mouse buttons.
For security reasons the customer didn't want their users to access the existing control panel applet. Added complexity was that the settings would have to be stored on the network and applied whenever the user logged on to any machine in the network.
Sounds simple (that's what I thought too when I started), until you find out that the calls required are not part of the published API for the operating system.
In the end I spent about 3 days writing the program (most of it on the network code) and 3 weeks trying to figure out what calls to use and what the heck those parameters meant by using a kernel debugger from a remote terminal and hacking into the operating system code for OS/2 on a running machine.
Start debugger, launch applet, change setting, and see what calls are generated...

[Toba]

tricky...

Once wrote a complete middleware system to link store terminals (running some fancy software we would not be writing) with a Cobol backend we maintained.
Didn't write it alone of course, was part of a team of 3.
During internal beta testing (using a dummy frontend to generate the messages from textfiles) the customer decided they didn't want the product and the subproject was scrapped. But it did work.

Or the most tricky app I wrote that actually made it to production was a C program written under OS/2 to swap the assignments of the mouse buttons.
For security reasons the customer didn't want their users to access the existing control panel applet. Added complexity was that the settings would have to be stored on the network and applied whenever the user logged on to any machine in the network.
Sounds simple (that's what I thought too when I started), until you find out that the calls required are not part of the published API for the operating system.
In the end I spent about 3 days writing the program (most of it on the network code) and 3 weeks trying to figure out what calls to use and what the heck those parameters meant by using a kernel debugger from a remote terminal and hacking into the operating system code for OS/2 on a running machine.
Start debugger, launch applet, change setting, and see what calls are generated...

Wow. Just wow.

[herdi]

mine don't know :o

[tgreer]

I invented the printing industry's first true eCommerce site, complete with dynamic, real-time PDF production. I wrote the first version single-handedly, before any PDF libraries existed, in CGI with a 2GL business language called PL/B. I had to work directly with the language vendor to get CGI to work.

The next version was also CGI, but I had a team: myself and two great freelance contractors who had previously worked for me.

The 3rd and current version is ASP, and is being used by the largest commercial printing company in the U.S.

I called it "COIN", but after it was finished they fired me and all of the other developers, turned it over to the marketing staff, and renamed it. It's still at http://www.cgxcoin.com.

Since then they have purchased several other off-the-shelf systems but still use the one I created.

It was extremely difficult because of, mainly, the interstate commerce and taxation laws. It was used by over 60 companies, in nearly every state. Any single order could be produced in multiple states, each with their own taxation rules, and shipped to a different state again, with ITS rules.

The most difficult part, though, wasn't the coding. It was working for a brutally awful company!

[1o0oBhP]

A 3d FPS Using a choice of DirectX 8 or software rendering... only got through 90% then gave up in the end as i have just started uni and dont have time to finish it. I also lost a bit after a serious error when messing about with asm code :-| needed file compression, some sound code and a lot more data to finish. The engine also was modified for isometric cnc type games. took a fair few weeks of solid code, not forgetting the years of tutorials and learning / error fixing before and after coding it... :!:

[mrwhy]

The hardest programs I've ever written

Most trouble with:

[mrwhy]

The hardest programs I've ever written

Most trouble with:
On the Altair (1965 a.d.) we hade to KEY in 8-bit code watching leds and when you went home the computer woiped it all out (not even tape memory)
On the Ferranti Pegasus (1963 a.d.) getting you punched cards in before the computer blew up!

Most difficult (complex) prog
A basic prog that played ANY notes on ANY scale imitating ANY musical instrument (even those beyond those allowed by physics)

Most interesting:
A set of progs that teach themselves how to solve "impossible" tasks (like the 100-city travelling salesman task)
Progs that teach themselves how to alter themselves to get better and better
Defining "better"

[Phaelax]

Either my iTunes clone in Java, or my RTS game in a non-OOP language. Neither are finished yet.

[chound]

Mine is "Hello World" when I was first introduced to C

[vegaseat]

A jpeg picture encrypter/decrypter I wrote for a friend, after his little brother spread the somewhat compromising digital pictures he took of his fiancee amongst his play buddies. Sorry, I don't have those pictures!

It wasn't too difficult, but had to be done right now!

[herdi]

I know now. The hardest program was making my own stupid darn web! Thre procceser wasn't working. So Ihave to buy an anti virus systum...
:evil: :mrgreen:

[herdi]

Java too. I always have problems with it. :evil: