Friday, September 28, 2007

What would it take to convince me Intel is Open-Source Friendly...

Currently I'm writing an entry about QT versus GTK and how the two different toolkits have an impact on the KDE vs. Gnome hostility. Part of the entry was driven by an Ubersoft comic, located here : http://www.ubersoft.net/comic/hd/2000/09/cant-win-losing

The basis of the comic is Richard Stallman talking down to TrollTech, makers of the QT toolkit. After reading it, it occurred to me that I might be acting like Richard Stallman in regards to Intel. Intel has done a lot for Open Source lately, so I can I continue my statements that Intel is Open Source hostile? Or do I have blinders on that prevent me from being objective about Intel?

Well, lets see.

For starters, Intel has open sourced their wireless drivers. Given that Intel Wireless chips are generally of a high quality with good signal range and transmission rates, this is a good step. Intel has open sourced their graphics drivers... sort of a middling step considering that most game developers get obscene when asked about support on Intel GPU's. Phoronix had a set of slides from IDF that painted Open Source driver development in a positive light.

So.. what has Intel done that is Anti-Open Source.

Well... we have TPM, or Trusted Platform Module, that effectively locks computers out from their users.

We have HDCP, which was developed by Intel, that effectively locks users out of their video and audio content.

We have the Intel Azelia audio specification which has seen no code releases or support from Intel. The current developer working on the Intel HDA audio driver is reported to have had significant hair loss due to the problems with the specification.

We have EMT64, which was a deliberate split in the x86-64 specification.

We have EFI, or Extensible_Firmware_Interface, which replaces the traditional BIOS and can also be used to lock users out of their own hardware.

Then last, but probably not least, there is LinuxBIOS / OpenBIOS. AMD has made significant code contributions to the LinuxBIOS project. The supported motherboard list is filled with AMD motherboards, and as far as I could read, no Intel boards are operational.


******

So, what would it take to convince me that Intel is serious about Open Source and that Intel is not Open Source hostile?

Well, get rid of TPM. Stop including it on motherboards. Make it clear to Microsoft that Intel will not assist in removing the end users controls on their hardware.

Open Source the HDCP specification. Make it public so that users can access their content. Tell the movie studios and RIAA executives that Intel will not help in their attempts to remove end-user control of content.

Provide a working 2 channel user-loadable firmware under a GPL style license, firmware examples for higher channel setups, and driver coding examples for Azelia Spec Audio devices.

Stop the x86-64 knockoff. Support the x86-64 standard in full and quit creating chips that have issues running x86-64 compiled code.

Open the source code and specification to EFI.

Support LinuxBIOS / OpenBIOS. Contribute code and BIOS examples for existing motherboards and insure that EFI is well supported.

These are the steps Intel would have to take before I would consider them Open Source friendly. Until these steps are taken? My statement stands. Intel is Open Source hostile, and end-users need to realize this when buying hardware.

Now, I'm not saying go out and buy AMD only, but there are not a whole lot of choices in the x86 processor market to choose from, although IBM's PowerPC based Cell would be desirable.
Post a Comment