Although you have send me this question 3 times now. (within one day!)
(by email)

I still don't understand the question. What requirements are you talking about:

1. requirements needed to run dosbox.
2. requirements in the specs dosbox would have to had (e.g. what were the goals of dosbox). (what we want dosbox to be able to do)

Ps: sending html email is not going to help as I got a text based email client.

I think he was trying to refer to the latter. -- The requirements/goals of what DosBox needs to satisfy when developing this software system.

For example, before you guys started developing DosBox, did you guys write down stuff like:
* DosBox needs to run on Linux, Mac & Windows platforms.
* DosBox needs to support the following sound device emulations:
- Soundblaster
- Adlib
- Roland
- PC Speaker
* DosBox needs to support the following video setting emulations:
- CGA
- EGA
- VGA
- Monochrome
* DosBox needs to support a minimum latency of 200 milliseconds for a keystroke.
* A release/version (alpha or not 😉 ) of DosBox needs to be completed within a one year time frame/product cycle.
* DosBox needs to support Gravis Gamepad input devices.
... etc, etc...

Pretty much what ur 2nd point was: "what were the goals of dosbox" and in developing DosBox, just written out and expressed in some detail.

I think hes referring to the requirements of dosbox itself. Just like when you buy a computer game or program it tells you the minimum system requirements (pentium 2, 512mb ram etc).

Hi Qbix,
Actually what my teammate is asking is not the minimum system requirements. The minimum system requirements are specified in the source files. I dont know if you guys are familiar with software engineering principles, but just incase, requirements are the intials constraints in which a software must be developed. These requirements must be well understood before moving into the implementation phase. This technique is generally applied to large scale software systems, but as learning students we wanted to apply this to a decent sized software system.

For example, if a customer asks: can you please build me a game emulator. And here are the requirements:

* must run on platforms linux, windows, MacOS X
* when special keys are pressed it must have a latency time of, lets say 2ms or less
* inorder for you to get the full payment for the development of the emulator, it must be completed by August 5th 2004, because this is when our company wishes to deploy the emulator for retail sales -- i know this is open source, but just an example of time constraints that effect the development of software systems.
* The system must support full screen mode
* the system must be stable or if there is an exception then it must be handled elegantly. ie, the thread the emulator is running on is killed instead of locking up (freezing)

The requirements we are asking for, is like when software for airplanes are built, initially there are basic requirements. ie the software system is fault tolerance ( if the software system fails then the software system shall allow the pilot to manually navigate the plane, so 200 people wont plummet 2000ft to their death)

My team is wondering if there was anything like the above specified before the initial implementation of the DOSBOX, kind of like a list of things the DOSBOX had to satisfy. If there is then this would help our research project tremendously.

Thanks

I dont know if you guys are familiar with software engineering principles, but just incase, requirements are the intials constraints in which a software must be developed.

This is fine in theory but may be not appliccable to a non-commercial-one-man-project (one-man-project refering to how it started out if I'm not mistaken).

Imran & comp804 - I think it is great that you have picked DOSBos as the subject of your study. I look forward the read the results 😀

I don't know which 'itch' that prompted the original developer(s?) to embark on the journey of creating DOSBox. If I was to venture a guess, I would say a combination of an interesting read about how to emulate an Intel 286 CPU in a magazine (Byte? Dr. Dobbs?), and a favourite DOS game gathering dust on a bookshelf.

Anyway, the 'itch' was sufficient to get the ball rolling.

You might be able to reconstruct part of the development process by looking through the official release notes, starting with v. 0.50, July 2002. They (should) summarise all the changes to the program as it has progressed to the current v. 0.61.

But that will not tell you what happended before the first release (v. 0.50). You have to play the role of archeologist. E.g. I dived into the CVS, and found an old file in one of the Attics: instructions.h. It contained this interesting comment embedded in the code:

/* let's hope bochs has it correct with the higher than 16 shifts */

The bochs being refered to here, is probably another Open Source Intel emulator, officially known as Bochs (currently v. 2.1), so perhaps it was not a magazine article that inspired the authors to start DOSBox, but the existence of Bochs?

Do not be surprised, if no great masterplan existed when DOSBox was initially conceived. The problem with such great plans are, that they run the risk of creating a system that is outdated, matches the requirements but not the users needs, etc.

To balance software engineering principles you mention, have a look at this classical text on the development of Open Source Software: The Cathedral and the Bazaar.

Hey everyone,

Hi, Dominus. Well you can actually apply the software engineering process to a non-commercial-one-man-project, but who would be crazy enough to do this. I for sure wouldn't. It's alot of dicipline that is unnecessary for non-commercial-one-man or small scale projects.

Hi, MiniMax. I will take a look at BOCHS, and see if I can dig up some rudimentary requirements from there, since I have extrapolated some high level requirements from the previous versions. Upon Analyzing the DOS BOX more throughly, I think this was a project that just started with no initial constraints and modified during the progression of implementation. But I will definately take a look at the .h file and the article you mentioned. Oh yea, when we are done with the reverse and forward architecture, I will post a link to where you can view what we have learned. Thanks again.

Tony

you are quite right.
The only requirement was that we wanted to play old dosgames on win2k (and linux for me).

All features were added because they were handy or neat or needed.

hi imran!

our company is also interested in dosbox architecture and
how it can be modified to fit our specific needs for dos emulators.

our 2 primary objectives in studying dosbox is allowing it
to use japanese fonts and rs232 communication.

version .61 has modem support but we have not tested this
yet. i've tried to code-trace the source to see if it will
fit our needs but have been unsuccessful compiling from source.

anybody care to help me how to compile the codes?

its hard code-tracing when you cant compile from codes.

anyway, whats your company's interest in dosbox?

please share to us your updates. or email me directly
at this address : jbillena@tspi.com.ph

good luck!

both should be possible.

Compiling the code...
What os ?

my pc specs:

intel 815 chipset motherboard
p3 800
128 SDRAM
Windows 98SE
S3 inc. Trio3D/2X (8mb)

dosbox 0.61 wont run on my pc but dosbox 0.60 runs well.

0.61 has problems with certain opengl implementations.
either get the current cvs source or wait till 0.62 arives.
Or update videocard drivers.

i noticed .61 has lots of problems with my pc. .60 works well.
but i couldn't compile from source the ver .60

right now, its linking error. i cant find a decent libpng.lib for windows.

but even after that, i'm still not sure if it will work. can anybody
just send me ALL THE NEEDED FILES TO COMPILE DOSBOX .60?

i already got the source code of .60. I NEED THE COMPILE INCLUDE
FILES AND LIBRARIES

more power to all of you, dudes!