“BIOS giant Phoenix, the company responsible for the pieces of code saved on flash memory and which initialize our computers and interfaces to the hardware at the low level – even before the OS loads, is giving us plenty of reasons to support the case of an open source BIOS.”
until the letters D, R and M show up to make the “Compelling” case an “Overwhelming, Set Up a PayPal Fund For a Reliable Development Team Right Now” necessity.
Flashing your BIOS is indeed a scary affair (during which I’ve made at least one motherboard useless..since noone seems to sell spare bios chips affordably). So I do hope the coming new version Intell is coming with (and shipping with the mactels) will be a significant improvement.
And it’s an open standard, right?
if it isn’t an open standard it’s in the process. I do know grub already has EFI support. heck Linux has already been booted on an EFI only system(the Intel imac, and itanium).
Open Source can support DRM. proper DRM doesn’t need security through obscurity no matter what MSFt says.
Proper use of DRM can be useful, as long as it doesn’t restrict fair use either. It’s a very fine line that is currently being crossed by both sides very often.
if it isn’t an open standard it’s in the process. I do know grub already has EFI support. heck Linux has already been booted on an EFI only system(the Intel imac, and itanium).
There is a working group right now; if one were to say ‘openstandard’ as in an organisation which a number of members are involved in its design and implementation; then EFI fits the same sort of definition that OpenGL and many other industry groups use. As for an ISO or ECMA, nope, it hasn’t yet.
There’s a trick for fixing a mis-flashed BIOS as long as the flash chip is socketed. You boot a working motherboard of the same model, hot-swap the flash chips (technically this is not allowed, but it works), and reflash the bad chip. I read about plenty of people doing this with dodgy SK8N BIOS updates.
Should be ‘mad geek with no knowledge of low level coding and hardware access uses phoenix as example on why he wants an open BIOS’
There are many many dozens of manufacturers of the many chips that make up a PC, USB, SATA, ATA, North Bridge, memory controller, etc, etc. and NONE of their chips work or interface the same way at the low level – if they did the same baseline core drivers would work on all operating systems without hassle. As such, the BIOS has to be tweaked/rewritten at least once for every mainboard combination of said chips. Even differences in how the chips are wired together can necessitate different code for boards using the same chips!
In order to do a open BIOS you’d either need to reduce functionality (truthfully I’m amazed we are allowed INTO BIOS anymore), or have a complete build source-tree akin to the linux shell (the latter appears to be the course the OpenBIOS project is taking). Even then, some companies (like gateway and dell) like to use proprietary chips for things like memory and cpu timing that they aren’t willing to share details on with the rest of the world.
Of course, the dirty hippy aspects of the article are the real meat. Things like:
>> oes Phoenix charge notebook and system builders to get access to newer versions of Winphlash?. Why not offer the Winphlash BIOS updater directly to end users through the phoenix.com site, so users always get the latest code with support for the newer flash memory chips out there?
OF COURSE THEY CHARGE MONEY FOR THEIR PRODUCT. I know with all the talk of open source these days that attitude may seem ‘evil’ or ‘unfair’ – WELCOME TO BUSINESS. They are a business, it costs MONEY to PAY programmers to write this stuff, especially when it comes to coding features to do things that were NOT part of the original design on something a few years old (like booting from USB) on multiple different chipsets.
God forbid programmers get PAID for their WORK – especially given the hoops the hardware manufacturers make them jump through when it comes to things like BIOS. You want to point a finger, point it at all the different hardware manufacturers who make devices that do the exact same thing entirely different ways on different ports with different interrupts with different communication protocols.
F COURSE THEY CHARGE MONEY FOR THEIR PRODUCT. I know with all the talk of open source these days that attitude may seem ‘evil’ or ‘unfair’ – WELCOME TO BUSINESS. They are a business, it costs MONEY to PAY programmers to write this stuff, especially when it comes to coding features to do things that were NOT part of the original design on something a few years old (like booting from USB) on multiple different chipsets
um the vast majority of open source coders get paid to code open source software. From sun to IBM, Novell, Red hat, and literally thousands of others pay coders to write Open Source software. Charging money for Software isn’t the problem. Even the GPL states you can sell the software for what ever you can get for it, as long as you include the source and redistribution rights along with it. Look at White Box linux. They take Red hat’s source tree remove Red Hat’s trademarked items compile it, and sell it for less than Red Hat, yet Red hat still does more business than White box.
Let’s see MSFt survive that way, or even Apple. They can’t. But Red Hat can, and Novell. Both still generate billions annually.
Open Source and Free software doesn’t mean the money stops flowing, it means you give rights instead of limiting people’s freedom and choices like MSFT and Apple does.
I don’t know how much of EFI is open but it’s a lot more so than Phoenix Bios has been of late. Hopefully with EFI we can finally move beyond the need for PS/2 keyboards, and 15 irq slots for all of our hardware.
>> They take Red hat’s source tree remove Red Hat’s trademarked items compile it, and sell it for less than Red Hat, yet Red hat still does more business than White box.
Which is the PROBLEM with the GPL and part of it’s naivete, you give up the right to distribute, you DESERVE to have some business major walk off with the programmers hard work. The ‘story’ of the GPL’s creation reads more like ‘you {censored} idealistic moron’. He gave away his work, then got pissed they wouldn’t give him the changes they made back. “You gave it away dumbass”
Free as in free speech is an idealism that has no place in business. Trade secrets and property rights are how you make money, and arguing against them rings of an naive idealism the likes of which hasn’t been seen since people were reading marx in the 60’s. As I’ve said before ‘programmers, stop devaluating your own work’ – it’s no wonder there’s no real money in programming anymore; what’s an entry level position pay now? 16-18K? You can make more than that working 3rd shift at a convenience store.
>> From sun to IBM, Novell, Red hat, and literally thousands of others pay coders to write Open Source software.
Which is funny since it always seems most of the true open source zealots fall into one of three categories: Professional Educator, student having life paid for by mommy and daddy, or hobbyist with a day job.
Even Linus has a day job at Transmeta – and if you think a hardware manufacturer REALLY has a vested interest in having him work on linux… Because of course we’re all running linux on Crusoe CPU’s. If that’s all he’s working on for them it’s a drain on their resources for no visible profit, and a likely contributor to the slow fade into obscurity. (Since the Crusoe introduction, when’s the last time you heard anything about Transmeta outside embedded circles?)
Again, I can hear the business majors laughing all the way to the bank as the programmers sit there eating ramen six nights a week.
Jesus….that had to be one of the smelliest brain farts I’ve ever read!
Edited 2006-03-01 01:48
I don’t know how much of EFI is open but it’s a lot more so than Phoenix Bios has been of late. Hopefully with EFI we can finally move beyond the need for PS/2 keyboards, and 15 irq slots for all of our hardware.
Welcome in 2000. Most chipsets support USB keyboards many CPUs got an APIC that can support up to 24 IRQs, some even more. Not like it matters since we got ACPI and some devices can even operate without using an IRQ.
EFI is nice, but it’s not like it’s going to change the world. Its openness doesn’t matter as much as software. Many modern OSes (like Linux) are not relying on a firmware/BIOS and writing quality code for these devices require some knowledge that many programmers don’t have, being used to an higher abstract level.
> Look at White Box linux. They take Red hat’s source tree remove Red Hat’s trademarked items compile it, and sell it for less than Red Hat, yet Red hat still does more business than White box.
Well, I suppose they do “sell it for less” because it’s free didn’t you know? As is what appears to be a better “fork” (CentOS) of RHEL – sometimes you wonder why all these free versions of RHEL don’t just merge into one…there is *no* point in having more than one near-identical free clone of RHEL (it’s one of the few areas of Linux where more “competition” is actually worse!).
You’ve got to remember that Red Hat’s reason for charging is that they have to maintain the distro for 5-7 years and provide support for it – the software itself is almost all free in the first place (except for some Red Hat artwork/trademark name issues), so they were always leaving themselves open to someone cloning it and releasing it for free. That’s the risk Red Hat were willing to take and it hasn’t impacted their business too badly. Of course, they do charge an arm and a leg for support (the world’s most expensive Linux distro?) which if you’re experienced with Linux admin, you don’t really need (but Red Hat won’t sell RHEL without support!) since CentOS + Red Hat’s Bugzilla is often good enough without spending a pennty.
I find a tech article a bit hard to take seriously if it contains the phrase:
“closed sauce code”
Whilst it’s all very well for the author to ask for an open-source version of the BIOS, it seems a bit pointless to me. To be brutally honest, something like Phoenix’s BIOS is really horribly backward, EFI provides a far superior solution
The Inquirer is a sort of tabloid for the IT world. They don’t take themselves too seriously, so it’s only natural that you don’t either. Many mainstream news sites pretend that they know what they’re talking about even if they don’t; the Inquirer doesn’t bother to pretend.
> Wouldn’t all that BIOS “spaghetti code”, in the hands
> of open source programmers, have produced a modular
> BIOS system by now […]
This shows a fundamental lack of understanding of the BIOS. It is correct that the BIOS is spaghetti code. Blaming it on the fact that there is no open source code for it is nonsense. First of all, the BIOS must support legacy programming interfaces which were simply not invented very well. This results in a worse code structure as well.
Secondly, many old programs bypass the actual programming interface and access internal BIOS data structures and code directly. This does not only mean that the BIOS *has* to be spaghetti code to work, it also means that various parts may not be based on high-level “source code” at all (i.e., neither open nor closed source code), because that would make it impossible to lay out the spaghetti in the right pattern (in the sense of memory addresses).
Sorry, but this is really idiotic. Telling 5 times BIOS 1 can’t boot, BIOS 2 can, well, I got that on the first time! What a useless read.
As a previous poster clearly explained, BIOS is quite chipset specific. So there is no easy way to make open source bios unless every bios writer (or modifier) releases the source with bios.
However, EFI from intel is open source and free. May be thats the way to go.
i think most of the heavy hitters are going to all move to EFI in the next few years anyway (or at least offer it as an option) so why fight it… lets embrase it.
>> However, EFI from intel is open source and free. May be thats the way to go.
Which would only be an option if manufacturers only used intel chipsets… No more VIA, nForce, ALi, ULi, etc, etc. unless those manufacturers take the open source and fork in the support for their chips – but at that point you are giving EFI all the problems inherent in BIOS – too many different versions that likely aren’t compatable with each-other, making the whole thing a wash.
Either that or those other manufacturers would have to agree to make their chipsets work the way the intel ones do – which is EXTREMELY unlikely.
The deal is that most of the EFI code is in C and i think it is well seperated from chipset specific stuff. So i feel there may be a way to use EFI efficiently and open source it without too much hassle. But then again, i may be wrong.
Also someone said that EFI C files can be compiled using intel’s EFI compiler but i think they can be simply compiled using Microsoft’s C compiler that comes with Visual C++ or their free platform SDK.
Anyways Mac is using EFI, Vista is using EFI so i think EFI is the way to go.
Aren’t you guys being a little too harsh on the author? He’s not claiming to be a BIOS programmer or a professional writer, but he wrote this article for a reason – he’s clearly got a problem, and Gateway is offering him no solution. I’d be really pissed if I couldn’t boot my notebook computer off a USB drive. This function has been available for years already, so any recently manufactured computer should have it. I’ll get Gateway didn’t put a sticker on the machine saying “USB mass storage devices not supported.”
Quite frankly, he should demand his money back. Or sue Gateway.
If commercial vendors like Phoenix can’t get it right, then an open source BIOS sounds like a good idea.
Quite frankly, he should demand his money back. Or sue Gateway.
Wow. What exactly would that do? As far as the article states, Gateway never said that the laptop does support USB 2.0 booting. The
So they have done nothing wrong.
and by business majors, I mean people like those who RUN IBM or Red Hat, take other programmers work, slap their name on it and make money off other peoples work without giving them a cent.
Because I’m SO sure Red Hat gives money to the people who wrote GiMP, Gnucash, etc. They’ll throw some money at the stuff that’s core, but have no qualms about letting 90% of what chews up the distro be stuff people were DUMB ENOUGH TO GIVE AWAY.
Cry me a river dude! If a programmer (or company) wants to release something under an open source license that’s THEIR choice, NOT yours.
Edited 2006-03-01 04:12
Many people seem to be unable to figure out how money is made off of open source and why it succeeds in general. It is really quite simple.
Open source is a way to combine R&D spending for certain products which companies desire to be available at a lower cost than what others are providing that product for. Such as at some point Linux will be powerful and full-featured enough that IBM will be able to eliminate AIX and switch from Windows to Linux on the desktop there by saving them R&D dollars, deployment costs within the company and providing a solid basis for more valuable non-commodity products such as software sold for vertical and specialized fields such as AutoCAD, medical software and etc.
Also Red Hat has proven that merely providing support can be very profitable for the millions who have no desire to get into the nitty gritty of applying patches manually, while also having someone to turn to if things go wrong.
Another reason that open-source succeeds is that if I’m good at writing network and TCP/IP code and you’re good at writing memory management code and someone else we know is good at device driver code in general, and so and so on then we can combine our efforts and create something excellent.
Many compare open-source to communism or socialism which is foolish. These operate by having design by the one or the committee similar to how most proprietary software is produced. Open-source on the other hand is very much a product of the free market. Someone voluntary produces a product and then agrees to give it away in exchange for you testing it for free and in the case of the GPL they must then return any changes and modifications of the code to you. This is the “viral” aspect of the GPL which isn’t viral at all and is merely an example of a contract or well written EULA at work.
Also many open-source products tend to be produced in a more decentralized manner than most proprietary software, although the project manager often decides the final pieces of code no one is forced to contribute unless paid to do so by voluntarily exchanging his services, programming, for money by his employer.
Also average wages in the technology sector have risen. The US Department of Labor Statistics reports that the average programmer in the US makes $75,000 a year. 6.4% higher than last year. Programmers are not making low wages as some on this board would have you believe. One of my friends who is graduating in May has a programming job in Central Arkansas and will be making $50,000 a year and he could probably make much more elsewhere but wants to stay close to home. Also $75,000 a year may be less than what programmers used to make on average but the late 1990’s were the result of an artificial bubble created due to too much capital (money and assets) being misdirected into the technology sector and therefore programmer’s where in much higher demand and when the demand of a good rises, the supply of that good falls, and so the price for that good in this case, computer programming, increases due to more limited availability. Average wages then fell because that service became less needed so the demand fell but the supply kept increasing thereby creating lower wages for programmers. Its really just simple economics.
“Also $75,000 a year may be less than what programmers used to make on average but the late 1990’s were the result of an artificial bubble created due to too much capital (money and assets) being misdirected into the technology sector and therefore programmer’s where in much higher demand and when the demand of a good rises, the supply of that good falls, and so the price for that good in this case, computer programming, increases due to more limited availability. Average wages then fell because that service became less needed so the demand fell but the supply kept increasing thereby creating lower wages for programmers. Its really just simple economics.”
Yeah, simple economics, but you got it wrong. In the short term, supply and demand are mostly independant of each other. When demand goes up, it doesn’t imply anything about supply. Increased demand will drive the price up, but not because “supply of that good falls.” In the short term, the supply is likely to stay the same. And in the long term, supply will almost assuredly increase as more people rush in to meet the high demand.
So, I think what you meant to say, was that in the late 90’s, demand for programmers increased drammatically faster than the number of programmers increased (the supply of programmers). So the price of programmers sky rocketed. But then, as time went by, a bunch of people got CS degrees, increasing the supply of programmers, and that lowered the price of programmers again. It’s also possible that as all those people were getting CS degrees the demand for programmers fell, which would also contribute to the lower price of programmers.
Thank You for the correction that is what I meant to say. I guess I should have looked it over once more.
What about open firmware?
I don’t know much about it, but it has open in it’s name.
What about linuxbios ?
What about open firmware?
I don’t know much about it, but it has open in it’s name.
the “open” in open firmware means open specs (in the form of the IEEE 1275 standard), not open source: actually, the PowerPc Mac’s, Pegasos boards, Sun’s machines and maybe some others, use commercial implementations of the standard made by some company wich may be less famous than the usual Ami or Phoenix, but doesnt use to give away the source code
AFAIK, openbios is the only opne source implementation of the open firmware spec
and, AFAIK again, they’re progressing slowly , but IMO, even in the case the code base was complete, mature and tested, they’ d have a hard time turning ALL those machines away from the legacy BIOS: too much software relying on bios calls and idiosyncrasies, but most of all, too many OPTION ROM’S (including PXE and netboot ), out there:
both open firmware and EFI aim to get away from option rom’s replacing them with interpreted drivers: but while EFI has opted for a C-like language (turned into bytecode by intel’s efi compiler) open firmware is stuck with forth and FCODE: yes, forth’s syntax allows for a very compact interpreter but how many people know how to write forth code? EFI on the other hand may leverage any developer with a minimum skill in kernel, or driver, programming, and then has a better time at becoming a mainstream standard
Edited 2006-03-01 14:17
So Gateway sold him a PC which can’t boot from USB devices and this smart-ass decides to blame Phoenix instead of Gateway. Nice! It isn’t Phoenix’s fault if Gateway doesn’t want to pay for USB booting support.
Tomorrow I’m going to write an article about how Microsoft sucks because most laptops don’t invlude floppy drives these days.
phoenix[1] started an open source project. it’s url is http://blahblahblah
[1] a BIOS manufacturer.