Talk:BIOS/Archive 0

From Wikipedia, the free encyclopedia
Jump to: navigation, search

BIOS as 'life' in Greek?

Is "While the name BIOS is an acronym, it is also a play on the Greek word βιος (bios) life. " intentional, or not? Dysprosia 08:15, 22 Mar 2004 (UTC)

It seems plausible to me, though of course I can't say for certain, that they thought of the meaning of the Greek word when they called the BIOS that. — Timwi 09:49, 22 Mar 2004 (UTC)
I had thought that the greek bios represented the connection that bonds your soul to your physical self. In a sense, that could be comparable to the computer bios, I guess. -Wisestfool, added July 30th
Actually to be more accurate.. the greek bios is the connection between the spirit and the body. Much like electricity (spirit) is applied to the element (body) in a light bulb, and the result is light (life,soul,bios). Life and soul are interchangable words for the greek bios. Spirit, soul and body are what make up our triune nature. - added November 17th, 2004
This is interesting, but still doesn't answer the question of intentionality? Suggest removing the sentence or changing unless a reference can be found? Chris Wood 03:27, 26 Jan 2005 (UTC)
I've changed the sentence to "may be" a play on the word Chris Wood 01:02, 5 Apr 2005 (UTC)

Thats not NPOV1

Thats not NPOV1

Read-Only Memory (ROM) chips are much slower to read than Random Access Memory (RAM). True. See later.

The storage capability of a ROM is also an issue, computer programs placed in ROM must be small and limited in features. In addition, BIOS code is a guarded trade secret, and the handful of vendors are often obfuscated their BIOS code to protect it from reverse-engineering. As technology advances, the features and functionalities built into the BIOS code begin to remove most of the limitations, or provide a workaround for them.

Most systems these days copy the BIOS code from the slow ROM into the faster RAM. Known as shadow RAM... done to speed up the BIOS routines.

The latest BIOS code support Plug and Play hardware components and Operating Systems, making system configuration much easier. Also, more and more manufacturer's license BIOS technology and optimize it to work with their components or systems.


My main reservation about the bits I cut out are: a) It views BIOS just a the PC BIOS everyone initially thinks of b) I'm not sure of its NPOV c) BIOS'es (in the PC ROM sense) are more like bootstrap roms than anything else, as most OS'es don't even use low level BIOS routines anymore.


Made several additions based on coding BIOSs for Phoenix Technologies. Though most people today equate BIOS = firmware = PC BIOS = boot, the first time I saw the term 'BIOS' was in 1981 source code from IBM's Technical Reference Manual. Thus, BIOS may be specific to the IBM PC, though it's meaning may have expanded in the last 20 years. Robert Keller


I think this requires some additional thought or clarification. "most OS'es don't even use low level BIOS routines anymore". That is, apparently you mean to say they don't use the DOS compatible BIOS routines, but they do. Also they use ASL routines which are provided by the BIOS - maybe this is a grey area though. On the first point, they still use the DOS BIOS routines for booting (INT 10, INT13, etc). Maybe you just mean the kernel doesn't use these routines anymore - on that we would basically agree.

On the second point, ASL/AML, it can be a grey area because there is a semi-valid argument that they are or are not BIOS routines. In my mind, they are provided by a BIOS vendor and written by a BIOS engineer and they express the same old BIOS functionality in a more modern way - they are BIOS routines. Of course, you could argue that a BIOS routine has to be DOS compatible or its not BIOS. I would counter that argument in that it is dated and unrealistic and in reality a DOS limitation and not a BIOS limitation.
--Riluve (talk) 20:52, 8 August 2009 (UTC)

EFI?

Can we please get some mention of the new EFI BIOS? -Unsigned I started an Extensible_Firmware_Interface like 4-5 years ago. Of course there is nothing left of what I started :-( --Riluve (talk) 20:55, 8 August 2009 (UTC)

BBS? Are we serious?

I think the whole BBS section needs to go. Its more of a suggestion than a specification. As far as I know the "BBS API" is just the PnP and PCI config space which the BBS suggests a BIOS read to get device information. Granted, its been 5 years since I read this "specification", but it was no specification the last time I saw it. It pointed out issues and problems but didn't have hard-line solutions for most of them. Furthermore, the issues it did try to solve, if implemented, no one would use/buy the result. Has there been an update since the original? If so, I haven't seen it. In short, BBS should not be taken seriously.

Some comments about the article: “which at a minimum drives the keyboard and provides primitive output to a display. ” This is incorrect. There is no need for keyboard support and the output to a display is misleading (especially on the primitive note). For example, A server BIOS has no need for keyboard support and for USB keyboard, the BIOS needs only to support USB HID’s and not act as a driver for the KBC. Rather, at a minimum, the BIOS initializes memory and programs the memory controller. Also, it discovers and initializes buses on the system (be it PCI, ISA, PCI-E or whatnot).

As for the primitive output display – it seems this “primitive” aspect should be something more about ASCII which could be on a local display or sent out say a serial port to a remote display. In the latter case, the display function, strictly speaking, is not supported. Maybe it would be better to say it would support ASCII console (at a minimum). That could be local or remote.

“most modern BIOS implementations” I would say all BIOSes are modern, maybe contemporary would put a finer point on it.

“My main reservation about the bits I cut out are: a) It views BIOS just a the PC BIOS everyone initially thinks of b) I'm not sure of its NPOV c) BIOS'es (in the PC ROM sense) are more like bootstrap roms than anything else, as most OS'es don't even use low level BIOS routines anymore.”

I am not sure what this is referencing, but it has an inaccurate characterization in it. That is to say, sure contemporary OSes no longer use the runtime services provided by the BIOS, but they are still provided. It is still a useful thing to be able to boot to DOS. If we are talking about this stuff, then it seems fair to mention ACPI as well.

As to EFI, I can draft up something if you would like. It is going to be the new PC standard. Starting with longhorn, and theoretically afterwards the only MS supported firmware interface. There are some introductory things about it on Intels website.

LMAO - ok see, this looks like me, I guess I can get a little crazy sometimes. It must be dusty on this talk page. No one cares about BIOS? --Riluve (talk) 20:57, 8 August 2009 (UTC)

Bios screenshot

I think the image of a BIOS on a payphone is not good enough for this article. It could be replaced by a screenshot (probably taken with a camera), clear, sharp, glare-free and only showing the BIOS screen (not the border of the monitor for example). An exact reproduction would be good as well. --Bernard François 16:22, 22 January 2006 (UTC)

Yeah, I agree. Does the PrtScrn/SysRq button work outside of the Windows operating system? If so, that would be great. If not, just use a camera to focus in on a monitor (preferably a large one with good resolution) and take a picture that way. --CanesOL79 17:36, 3 February 2006 (UTC)
No, it doesn't, and someone has posted a picture. It's not the best, but what someone could do is make a BIOS mock-up using Paint. Considering the limited graphics capabilities and extremely limited palette of the BIOS, it shouldn't be too hard. --SheeEttin 23:02, 23 March 2006 (UTC)

I think this is a better picrure: http://deepthought.ena.si/logo/bios/award.png - AWARD BIOS

http://www.ellak.gr/pub/OpenGuides/Debian/pics/app1/bios.png - PhoenixBIOS

or http://cdn.mirror.garr.it/mirrors/AppuntiLinux/immagini/bios-borg/ - many "screenshots"

Here's another one, made by me. Not an actual screenshot; I made it by extracting the OEM font from my video ROM and "drawing" the setup screen manually, with some help from self-written scripts. This is exactly, scanline for scanline, what appears on my screen when I enter BIOS setup (well, the palette may be a bit off. And yes, it is 10 years old - and still works!). You may replace File:AwardBIOS_CMOS_Setup_Utility.png with this, I don't claim any copyright on it (well, per Bridgeman v. Corel I actually can't claim it in the US). Why won't I simply register and do it myself? Good question. Well, I'm planning to, but for now I'd like to remain quasi-anonymous. Ian (77.254.8.139 (talk) 15:14, 8 September 2009 (UTC))

Address

The BIOS is located on address FFFF:0000 ? -- Frap 12:17, 12 December 2006 (UTC)
No
--Riluve (talk) 21:02, 8 August 2009 (UTC)

add these information

You may add information what is BIOS Release Number and BIOS Reference Number. Cya Jerry --195.113.141.199 (talk) 09:32, 5 November 2008 (UTC)

Definition

As it's long been a source of confusion, with little discussion had, I've amended the article to define BIOS specifically as the PC (IBM PC Compatible/IBM AT+/PC97+) firmware which provides the well known Interrupt interface found in the original IBM PC, that is well documented in Ralf Browns Interrupt list and appended to by BIOS manufacturers. While this leaves Wikipedia without an article, or at least information (could be added to the firmware article?) on boot firmware, or at least upon basic system ROM as opposed to generic firmware. The other issue lays in actually expressing the definition. I've currently used the term "de facto standard", however technically that's incorrect. The BIOS is a term that relates to every implementation of a BIOS that is built up on the original system, as a whole, but is compromised from numerous sources, some with specifications and not all implementations supporting all the functionality. Can anyone word the introduction to express this idea succinctly? I will work on it, but I'm sure someone far more adapt than me could do it first, and best. - Jimmi Hugh (talk) 21:11, 14 January 2009 (UTC)

Here's an example of what I am thinking, but it needs better writing, so I wouldn't want to add it to the article now.

In computing, the Basic Input/Output System (BIOS)[1] , also known as the System BIOS, refers to a component in IBM PC Compatible computers, based upon the original IBM PC firmware, and numerous continuing developments, that is an implementation specific system based upon a de facto standard boot firmware interface.[2]

Basically we need to word the development as a diverse production, with features constantly being borrowed and merged into an unmaintained standardisation. In the same way that E820 was brought from implementation into de facto status, or that the EDD specification was implemented across the board. Any issues? - Jimmi Hugh (talk) 22:17, 14 January 2009 (UTC)

CP/M systems had a BIOS that was not on a chip and not even (gasp!) IBM PC compatible. The defintiion is a little narrow as the article currently stands. --Wtshymanski (talk)
The CP/M had a software component called a BIOS, that represented the HAL, or architecture dependent half of its Kernel. That component was (gasp!) not a BIOS as detailed in this article. Infact the article makes clear that it was of a common, though not provalbly originating positiong, in name only. - Jimmi Hugh (talk) 18:14, 20 February 2009 (UTC)
The CP/M BIOS was as much a "basic input/output system" as the EPROM in an IBM PC. I think since the article only talks about the IBM-PC-compatible world that the first sentence should say so. If we're going to be narrow, we might as well be honest about it. --Wtshymanski (talk) 19:16, 20 February 2009 (UTC)
What? I'm completely confused... your point is identical to my own. This is specifically about the IBM PC-Compatible firmware, that's entirely what I stated, and what both introductions stated. The reason I reverted the edit had nothing todo with that, it was the blanket revert of incorrect information about purrpose and function. Note that both introductions make it abundantly clear that this is about the BIOS, and has nothing todo with the similarily named CP/M component. - Jimmi Hugh (talk) 01:07, 21 February 2009 (UTC)
The IBM pc rom bios is firmware, is it not? It's the first code the processor runs on a power up or reset, is it not? It detects, tests and does a rudimentary set-up of the hardware to boot the operating system, does it not? What's incorrect? Note that the same information is still in the article, without the windy generalizations 'in computing' and 'de facto'. Please explain which statement is incorrect. --Wtshymanski (talk) 15:15, 21 February 2009 (UTC)
I owe you an apology, in all but the conversation about the spread and defintion of BIOS, I wasn't aware how much of the correct text you had kept in the introduction, and was childish in assuming you'd reverted back to the claims that such things were the primary function without having read your changes into better detail. - Jimmi Hugh (talk) 15:27, 23 February 2009 (UTC)
Apologies are not necessary - it's all part of the "merciless editing" process. My concern was that we not say that something that is particular to the insular and isolated world of Windows-compatible computers is true for all "computing" everywhere and at all times. Let's not be any more pompous than we have to be. Now if only someone could straighten out the useless stubs at IBMDOS.COM IBMBIO.SYS, MSDOS.SYS and IO.SYS, etc. and put them all into an article called DOS boot files or something. I tried merging these into DOS (also a bad article title for other reasons) but got shot down. These stubs are repetitive and fragmentary - better to describe the processa all in one place. --Wtshymanski (talk) 16:34, 23 February 2009 (UTC)
I don't know what it has todo with Windows.... but it was still made clear in the first sentence that it's related to IBM compatible systems, that's quite early on to make that clear, how much more browsable does the text need to be? As for the structure, I don't think it pompous, but it does standardize articles and remove any tone they might have if we have contextual openings. Note also, my above comments on changing the introduction, I am aware we need a better way of expressing it's de facto status given the differences between implementations. - Jimmi Hugh (talk) 17:57, 23 February 2009 (UTC)
I really dislike this definition of BIOS as an IBM PC specific thing. There should be an article on the BIOS generically, and a separate article on the IBM PC BIOS. We can't even have a "History" section in this article with Gary Kildall and the like because all of that occurred before the IBM PC was created. Comet Tuttle (talk) 20:12, 23 October 2009 (UTC)
You know what to do. Find a reference to other uses. This article tips the hat at the CP/M world where BIOS was also a standard term, but I don't recall seeing this mentioned much in other contexts. Most of the home computers kept *everything* in ROM, and the way the Mac and Amiga boot ROMs worked isn't really conceptually different from the way the PC wakes itself up. --Wtshymanski (talk) 15:36, 24 October 2009 (UTC)
Google it. A quick search off "Atari 800 bios", "Commodore 64 bios", "Apple bios", "trs-80 bios" shows the term bios is far more generic then this article suggests. bios is basically the boot room of any computer. As someone who was there that was common usage far before the IBM PC came along.Gregg Tavares

Firmware on adapter cards

Last sentence points to dd, which is useless for accessing BIOS chips. I've changed dd to flashrom, which is the right UNIX utility for this, but it got reverted by user Wtshymanski. I disagree both with that revert, and with mentioning dd in this context. -- Cghost (talk) 13:36, 30 October 2009 (UTC)

SPI and booting out-of-the-box

If the BIOS has moved to buses such as SPI, how can the CPU still easily boot from a hardwired address such as CS:IP FFFF:0000 (see Booting#Boot_sequence_on_standard_PC_.28IBM-PC_compatible.29)? --Abdull (talk) 12:45, 27 January 2010 (UTC)

The chipset knows how to route that address to the SPI flash out of reset. Nothing changes from the CPU's perspective; the flash just moves from LPC to SPI behind the south bridge. --gribeco (talk) 19:05, 27 January 2010 (UTC)

IMAGE

Possibly replace generated image with real hardware version? http://img514.imageshack.us/img514/3301/realhw.jpg I made it myself Smeezekitty (talk) 22:44, 5 March 2010 (UTC)

Redundancy

A bit of redundancy can increase clarity. e.g. CD-ROM Disc (Compact Disc Read Only Memory Disc), DVD Video (Digital Versatile/Video Disc Video), SAT Test (Standardized Achievment Test Test), and BIOS System (Basic Input-Output System System).

Can the article retain "BIOS systems"? - Brewthatistrue 6 July 2005 20:50 (UTC)

Seriously off-target out of the gate

Just in reading the first two paragraphs, this article states incorrectly what BIOS is. Just want to share what I know; no time to do any serious editing on this article. First of all, BIOS is not a boot loader. BIOS includes a boot loader, but it is less than 5% of BIOS. BIOS also is not designed to load an operating system and then be forgotten. As the IBM Technical Reference manuals for the various PC-family machines, which originally defined the BIOS, clearly indicate, the BIOS is designed to be a low-level hardware abstraction layer, and IBM recommends (originally) its universal use. So, in other words, the BIOS is the earlier counterpart to things like the Windows API, which Windows requires all programs to use for hardware access, prohibiting (and preventing, as it can) direct manipulation of the hardware. The second paragraph goes on to say that the primary purpose of BIOS is to initialize hardware devices into some known minimal operative state and then. That may be the perspective of the programmers of large, ultra-sophisticated modern operating systems, but it is not objectively the purpose of BIOS. BIOS does not initialize devices so that other programs can directly manipulate them afterward, although it does allow for that in some cases (notably video display devices and serial UARTs). BIOS initializes devices so that it can provide services using them. BIOS is designed to allow other programs to go around it and even to essentially wipe out its control and capacity after its initialization is done, but this is largely because it is an Intel x86 real mode program (or, depending on perspective, a collection of programs) and there is no protection in real mode to prevent this; a program running in real mode essentially owns the machine, until it hands it back nicely to its caller. (It's sort of like renting your house to someone while you go on vacation--it's up to their good behavior to preserve your house the way you expect for when you get it back.) All operating systems that run purely in protected mode abandon BIOS, since they can't call it without switching to real mode. Therefore, for those operating systems, the only useful effect of BIOS is to initialize devices and perform the initial stage of booting. That BIOS is usually used that way today doesn't make that all that BIOS is.

I don't know much about the few "alternatives for Legacy BIOS in the x86 world" today, but I suspect they may be more focused on just initializing those devices necessary for booting (i.e. keyboard input, basic video, disk drives, and network adapters) than supporting every device in the machine through a standard interface. That is what BIOS aims to do, whether or not it is used by modern programs. Old programs (anything under DOS) still relies on it, and so DOS emulators must either thunk BIOS calls (through to real mode) or emulate BIOS completely.

The third paragraph is better, touching on this basic concept. However, it is not something else that BIOS "can be said" to be. The correct description of BIOS is that this is what it is, although that is how it's most often actually used in modern times. Feel free to use any phrases, sentences, or other parts of this message to improve the article.

People also commonly use the term BIOS to refer to CMOS Setup, or, more generally, a system configuration program in ROM, using sentences like "You have to enable the second IDE channel in BIOS." The BIOS is what primarily uses these settings, but the setup program is not BIOS, it is a program that configures BIOS and may, at most, be considered a small portion of BIOS. However, this usage is common in colloquial speech, and so we must acknowledge that it has come to be a second meaning of the word--BIOS is either the actual BIOS or the BIOS Setup Utility (program.)

I have a strong opinion on this subject, I admit. If anyone thinks I'm off base and wants to argue another position, please do so politely. I'm not attacking anyone, only some ideas; let's keep it that way. (I may not get back here to read it, but hopefully this will still start a discussion.)

Another problem is that the article text makes no mention of POST, which is the only operation of BIOS, other than the brief action of the boot loader, that is actually visible to the user and takes observable time.

71.242.27.236 (talk) 00:56, 19 October 2008 (UTC)

Agree, this article is mostly bollocks. Out of the 20+ BIOSes I've encountered in the last few decades, only about 4 of them have fitted within the bounds of this article. This is an article on *PC* BIOSes primarily. Almost nothing in the article applies to any BIOS that's ever run on any Alpha-, Sparc-, ARM-, or POWER-based machine I've ever used. (And probably the MIPS and HPPA ones too, but I never touched the BIOSes on those, so can't be sure). Everyone who's contributed to this article should feel ashamed of the depth of their ignorance, this is possibly one of the poorest quality wikipedia articles I've ever seen. — Preceding unsigned comment added by 87.119.183.129 (talk) 01:40, 1 December 2011 (UTC)

Built In Operating System

Does anyone have a source to show that it actually refers to BIOS as in Basic I/O System Built in Operating System? I've only ever heard the term used for the double meaning in relation to Operating System's built into firmware, and not to actual BIOS'. - Jimmi Hugh (talk) 22:07, 9 January 2009 (UTC)

One instant reference I have is Andy Johnson-Laird The Programmer's CP/M Handbook, Osborne-McGraw Hill, Berkeley, California, 1983 ISBN 0-88134-103-7, which refers to BIOS as "Basic Input/Output System" - and this term was old in the art in 1983. I expect it would pre-date CP/M. I've found the usage referring to I/O system to be a lot more common than "built-in operating system" which I'd never seen before it appeared in this article. --Wtshymanski (talk) 16:22, 10 January 2009 (UTC)
Oops, apologies, I didn't read my question before posting. I meant "Built in Operating System". There are a million sources for the BIOS. I've changed it in the original question, sorry about that. I have actually seen use of the second term, but my question is really if it refers to the same thing. The definition at current seems to contradict the very idea of a BIOS, and crosses over into boot firmware. - Jimmi Hugh (talk) 16:41, 11 January 2009 (UTC)
OK, that's a clearer question, but I still don't know the answer. I've seen, for example, one Tandy machine described as having "MS DOS in ROM", as does an HP 95 but these MS DOS ROMs I don't recall being referred to as "BIOS" as in "built-in operating system". Even in the murky world of personal computer nomenclature, this would be considered confusing. (An essay on the murkiness of MS DOS related terms would not be encylopediac, but would be entertaining.)--Wtshymanski (talk) 19:08, 11 January 2009 (UTC)
Yeah, I shudder at the thought of trying to sift through that mess. However, whether the term relates to that or not... is it still a BIOS? It may be firmware, and even part of the same chip, but the BIOS is meant to refer to a specific de-facto specification that features in IBM PC Compatible and later computers, is the term "Buillt in Operating system", really appropriate as simply another possible expansion? - Jimmi Hugh (talk) 19:51, 11 January 2009 (UTC)
I haven't seen "BIOS" expanded as "Built-in operating system" - and whoever stuck it in there initially didn't put in a reference for it. I'd be happy to see it gone. --Wtshymanski (talk) 20:21, 11 January 2009 (UTC)
Ok done. I'll be making some bigger edits soon hopefully, but hoping to compile a nice list of references to work from first. Unfortunately finding books that aren't oversimplified to the point they're technically incorrect (i.e. "How Computers Work" etc.) is quite difficult. But this is already quite set to being a good article. - Jimmi Hugh (talk) 20:32, 11 January 2009 (UTC)
It's months later, but I'd point out that I think the popularizer of "Built-In Operating System" (not that it's popular, but still) was the book Snow Crash; the author, Neal Stephenson, wrote a defiant comment in the foreword or afterword saying: I know this is wrong, but it's what it should be. Comet Tuttle (talk) 20:11, 23 October 2009 (UTC)
The term BIOS for Basic Input/Output System was coined by Gary Kildall, the inventor and developer of CP/M in the 1970s. Here is an PL/M excerpt from the original CP/M source code (BDOS.PLM):
|[...]
|
|/* C P / M   B A S I C   I / O    S Y S T E M    (B I O S)
|
|                    COPYRIGHT (C) GARY A. KILDALL
|                             JUNE, 1975
|
|                                                          */
|
|[...]
|
|/*  B A S I C   D I S K    O P E R A T I N G   S Y S T E M  (B D O S)
|
|                    COPYRIGHT (C) GARY A. KILDALL
|                            JUNE, 1975
|
|                                                                 */
|
|[...]

At this time, BIOS and BDOS were already logically seperate modules inside the same source code. The physical separation of the BIOS and the BDOS came somewhat later in 1975-1977 with CP/M 1.3 and 1.4, also inspired by input of Glenn Ewing, a developer at IMSAI, who adapted the CP/M BIOS to the IMSAI 8080

http://www.imsai.net/history/imsai_history/cp-m_history.htm

The ROM of a CP/M machine typically contained little more than a boot loader (and often a machine monitor), no BIOS. So, the vendor specific CP/M BIOS was developed by the OEM and loaded from disk. (CP/M BIOS example and template source code was typically provided by Digital Research.) The CP/M BDOS was completely machine indepedant and typically not provided as source code.

With the advent of the IBM PC in 1981, the BIOS became two parts, an operating system independant and an operating system dependant part. (They were not modelled after a CP/M BIOS, though.) The first is what makes up the ROM-BIOS (or System BIOS) of the PC, the second is the so called DOS-BIOS typically residing in a file named IBMBIO.COM, IO.SYS or DRBIOS.SYS.

MS-DOS/PC DOS had a BDOS as well, but Microsoft just named it differently, "DOS" kernel. This is what makes up IBMDOS.COM aka MSDOS.SYS. DR DOS continued to visibly use the BDOS designation (DRBDOS.SYS) up until 1989, but the internal kernel is still called BDOS up to today and the corresponding kernel version API still reports a BDOS version (which is different from the API emulation version).

In Digital Research terminology, the part of the BIOS and/or OS residing in ROM was also called ROS (either for Resident OS or for ROM OS, I'm not sure about that right now).

--Matthiaspaul (talk) 11:42, 1 December 2011 (UTC)

Regulated to bootstrapping

I removed the following line: As such, the BIOS was relegated to bootstrapping, at which point the operating system's own drivers can take control of the hardware.

This is misleading at best and I can not see an argument that it might be acurate. BIOS still does everything it use to do for DOS (you will notice you can still boot DOS) AND it provides ACPI for any contemporary MP OS. A key part of ACPI is the ASL/AML which is in fact a driver the OS uses all of the time. As evidence that BIOS is far more than just a little chunk of bootstrapping code, you must compare BIOSes of today with BIOes of yesteryear. In 1998, the largest you would see a BIOS would be 256kb - that was big in fact. Today, BIOS is usually at least 1mb in size and 4mb is not uncommon. This is a huge degree of additional funtionality squeezed in - with no functionality removed.

Additionally, you will see every x86 based computer manufacturer that supports windows license its BIOS from a BIOS vendor. Many of these companies have enough resources to write and maintain their own OS, but they have to buy the BIOS - to be competitive. Finally, free alternatives - U-Boot (an actual bootloader only for bootstapping) and OpenFirmware have made no significant impact on the BIOS "industry". This only makes financial sense if the BIOS is providing critical functionality that is difficult to maintain.
--Riluve (talk) 20:37, 8 August 2009 (UTC)

Kind of pointless though, since it's never used for that on 99% or more of PCs sold today. Even the few that do that generally load Windows 2000 or later or even Linux distributions which don't use BIOS for hardware interface support. They buy the BIOS because stuff like initializing the video card and detecting bootable devices on a half dozen different interfaces has to be done before booting to the OS and isn't trivial to program. This is not essential to the OS but rather to getting TO the OS. This is otherwise known as bootstrapping. Configuring DRAM to operate with a certain timing is also done before the BIOS even bothers initializing the video card. The same with CPU timing settings. It is something that is best programmed using modules licensed by the BIOS companies. No one "buys" a BIOS but rather use a set of modules that can be configured then packed (compressed) into a BIOS image file to write to the Flash ROM. They try to avoid rewriting the >95% of functionality that isn't unique to a specific motherboard or manufacturer. The ACPI support is actually more like tables that tell the OS what capabilities to use. I know your edit wasn't malicious but people need to let experts do wholesale deletions or research the issue (become expert) more before just removing information that novices might want to know. 71.196.246.113 (talk) 09:29, 3 February 2012 (UTC)

BIOS and PC compatibles

So sup, stupid question; a lot of people are calling the firmwares in a lot of stuff "BIOS", not just the PC; do we have any official source for this or is it just a case of people not having a clue, and on which page should the addition be done? Correcting the mistake at articles like PS2 or making a mention of the popular usage/synecdoche of the word BIOS to refer to firmwares in general... —Preceding unsigned comment added by 174.57.250.142 (talk) 11:50, 11 July 2010 (UTC)

The term BIOS does not refer to firmware in general, and its usage is not restricted to PC compatible computers, either. Other types of computers can have BIOSes as well. And BIOSes do not necessary need to reside in ROM (except for a ROM-BIOS), a BIOS can also be loaded from disk (as, for example, in the case of the CP/M BIOS or the DOS-BIOS - see further up) or a network (example: NETBIOS). BIOS refers to a certain type of software architecture, where it defines the lowest (or at least a low) level and hardware abstraction layer, the so called Basic Input/Output System.
Firmware is a more general term, refering to any kind of embedded software necessary to make a specific hardware device functional. Firmware may also include definitions for configurable hardware logic, a low level software library such as a BIOS, drivers, an operating system and application programs. A BIOS may be part of a device's firmware, but a firmware does not need to contain a BIOS. --Matthiaspaul (talk) 23:25, 5 December 2011 (UTC)

Way too PC-specific

It appears the authors of this article know nothing of other computers other than PCs.  They might be startled to learn that virtually all computers have a BIOS of some sort.  Even the venerable Commodore PET had one.  Really lousy article!

Bigdumbdinosaur (talk) 20:15, 12 August 2012 (UTC)

Perhaps the article should be renamed to PC BIOS but the bottom line is that the article is clearly describing a PC BIOS so your changes were personal POV and have been reverted. Asmpgmr (talk) 00:15, 13 August 2012 (UTC)
Wrong! A BIOS is present in all computer hardware except very small embedded systems.  My revisions amplified the fact that the essential purpose of the BIOS is the same in all systems—there is nothing unique about an x86 architecture machine in that regard.  Your rationale for undoing the revision suggests to me a limited knowledge of computer hardware in general and a POV that is clearly PC-centric.  Not everything with a microprocessor is an x86 machine, and unless an article specifically states that it addresses a particular MPU, architecture and/or brand-name, said article must, under WP guidelines, be written to be neutral.  The article as previously written, failed to meet that standard, since the title is BIOS, not PC BIOSBIOS is a generic computer hardware engineering term, whereas PC BIOS refers to a part of a specific (albeit widely used) computer architecture.
Now, if you want to create a PC-specific BIOS article you should have no difficulty copying one of the old revisions that contains all the PC specifics and edit it into a new article titled PC BIOS.  That would be the place to get into PC BIOS specifics.
Incidentally, may I suggest you read the WP policy on neutral point-of-view?  While you're at it, check out my credentials for writing about computers.  Then come on back and tell everyone about my supposedly personal viewpoints.
Bigdumbdinosaur (talk) 04:42, 20 August 2012 (UTC)
BIOS usually refers to PC BIOS just like DOS usually refers to MS-DOS/PC DOS. Adding a paragraph to that effect would be OK but removing significant portions of the article because you dislike this fact definitely comes across as point-of-view. The simplistic ROM firmware of any of the classic 6502-based systems (Apple II, Atari 800, Commodore 64, etc.) is definitely not comparable to even the original IBM PC BIOS as those systems had no formal API like the PC BIOS does (e.g. INT 10h-1Ch). Also if someone is a BIOS engineer, it is without a doubt someone who works on PC BIOS (or its successor UEFI). I have never heard that job title used in reference to any other computer's firmware, in the case of any other computer architecture the generic term firmware engineer is used. As for credentials, you really don't want to go there with me. Asmpgmr (talk) 05:07, 20 August 2012 (UTC)
How about finding some reliable sources about how the term is used? Verifiability is the policy, not credentials. Keφr (talk) 06:47, 20 August 2012 (UTC)
Well I'm not the one who started the talk about credentials. Anyway if you want reliable sources for the use of BIOS in referring to the firmware of PCs then how about Phoenix and AMI:
[Phoenix] [AMI]
Neither company specifies PC BIOS as this is assumed since this usage of the term BIOS is ubiquitous. Also these companies are referred to as BIOS companies and there is no need for further specificity. Terms like BIOS engineer, BIOS developer, BIOS company or BIOS API all refer to PC BIOS and this usage is ubiquitous so it can be safely assumed that if someone enters BIOS they mean the BIOS firmware of a PC just like if someone mentions DOS it can be safely assumed they mean MS-DOS/PC DOS (or compatibles) and not DOS/360 or something like Atari DOS. At most a paragraph mentioning that all computers have BIOS firmware should be added but the remainder of the article should be left alone. Asmpgmr (talk) 16:38, 20 August 2012 (UTC)
Not that I necessarily disagree with your point (that "BIOS" primarily refers to IBM PC compatible firmware), but these links count as primary and first-party sources (never mind that those specific pages do not seem to define "BIOS" anywhere). These may be sometimes considered reliable about what they claim, but are definitely inappropriate when establishing common usage; for one, first-party sources often insist on using "official" names, even where Wikipedia would rather use a WP:COMMONNAME. Secondary or even tertiary sources are more applicable for this purpose. Note that some sources already used in this article (though admittedly of questionable quality) are not very explicit about referring to IBM PC compatibles. And I have personally seen the term being used to refer to non-PC firmware (Game Boy Advance firmware, for one), so perhaps a usage note might be appropriate. Keφr (talk) 21:11, 20 August 2012 (UTC)
Why would Phoenix or AMI need to define BIOS ? They are BIOS companies, they certainly know what they are developing. This illustrates the point that 99.99% of the time BIOS refers to the firmware of PCs. Anyway what is the point of questioning this ? Anyone remotely familiar with computers since the 80s knows this. Adding a usage note or paragraph mentioning the generic use of the term and that BIOS generally refers to the firmware of PCs is fine but rewriting the article is not and the suggestion of doing so comes across as POV of someone who is unhappy that PCs are ubiquitous. Asmpgmr (talk) 23:37, 20 August 2012 (UTC)
Why would they? Perhaps to explain it to people interested in what these companies do. This is an article also for people unfamiliar with computing technology, and they probably want to read something based in established facts, and not opinions of individual editors. That is what sourcing is for. Whether to cite "obvious" (for a given definition of "obvious") facts is an ongoing dispute (see WP:BLUE and WP:NOTBLUE), but the consensus right now is that if a statement is challenged, it should be sourced. And I happen to find this objection valid, even if disagreeable. Keφr (talk) 10:04, 21 August 2012 (UTC)
Why would they? Perhaps to explain it to people interested in what these companies do.
Um, they develop BIOS (and UEFI) firmware for PCs. That's patently obvious from the fact that they're BIOS companies. Asmpgmr (talk) 03:34, 22 August 2012 (UTC)
Imagine for a while you have never heard the names "Award Software", "Phoenix Technologies", "American Megatrends". Do these names tell you anything? What do you think these companies do? How are you supposed to know that? Keφr (talk) 07:17, 22 August 2012 (UTC)
I've known what these companies do (did in the case of Award which was bought by Phoenix in 1998) for over 25 years. I'm pretty sure I knew what a BIOS was before I knew the company names so finding out that they were BIOS companies told me what they did. Anyway this is getting way off track. I have no objection to adding a paragraph that BIOS is commonly used to refer to the firmware of PCs but is a generic term but the article should not be rewritten since it is a fact that the term BIOS is almost always used to refer to PC BIOS just like DOS is almost always used to refer to MS-DOS/PC DOS or its compatibles. There's a definite expectation of what BIOS and DOS refer to. Asmpgmr (talk) 16:43, 22 August 2012 (UTC)
I am not arguing against that; in fact, I think this is the likely outcome of this dispute. What I am arguing for is that whatever the outcome be, it should be supported by reliable, third-party sources. Like this one: ISBN 1931769605. I once came upon an online article by the same author, which I found helpful and in my eyes gave some credibility to the author. I don't own the book itself, so I am unable to use it to cite anything, but I am willing to bet this is a reliable source for claims it contains. The kind of source we need. Though I think in this specific case, a WP:TERTIARY source would be more appropriate.
And you missed the point of my question. I was asking specifically how are you supposed to learn that these companies do what they do. If you infer it by yourself by looking at your computer's boot screen - well, your inferences could be wrong, couldn't they? And what if the companies also do something else, and developing BIOSes actually generates a tiny fraction of their revenue? So it's still about sourcing. Keφr (talk) 17:21, 22 August 2012 (UTC)
Why would that book possibly be a more reliable source than any of the following:
* IBM Personal System/2 and Personal Computer BIOS Interface Technical Reference, IBM, 1988, ISBN 999857739X
* System BIOS for IBM PCs, Compatibles, and EISA Computers, Phoenix Technologies, 1991, ISBN 0201577607
* Programmer's Guide to the AMIBIOS, American Megatrends, 1993, ISBN 0070015619
* Any of the IBM PC Technical Reference manuals (PC, XT, AT) or a technical reference for a PC-compatible which actually contained BIOS source code listings.
These are references from the actual BIOS manufacturers. Certainly a source code listing is absolutely definitive. Asmpgmr (talk) 18:02, 22 August 2012 (UTC)
I completely agree with Asmpgmr here:
  1. It is clear the topic of this article is the BIOS as found in PC-compatibles. If the scope is considered not to be broad enough then that should probably be dealt with by starting a new article, and possibly renaming this article, not by merely rewriting the lead section of the current article.
  2. However, while many non-PC computers have similar functionally to the BIOS, this is more often know as "ROM" or some elaboration on "firmware" (Open Firmware, Extensible Firmware Interface, etc.) I would be interested to see some reliable sources referring to the firmware on non-PCs as "BIOS" before considering whether this article should be renamed.
Ruud 19:56, 22 August 2012 (UTC)