Re: How free is Debian

[Andrei POPESCU]

On Sb, 10 aug 19, 07:50:02, John Hasler wrote:
> I wrote:
> > I'd just flat out restrict them to five years.  Twenty is too long.
> 
> Andrei writes:
> > That would work against inventors as instead of buying useful patents
> > companies would just wait 5 years and then use it without any charge.
> 
> Anyone can practice an expired patent without any charge.  That's why
> they expire.  Companies could wait 20 years now but they usually don't:
> it's worth money to be in the market now.  If it was 5 years they'd pay
> less, but they'd still pay.

Like they did for the cyclonic vacuum cleaner?

Kind regards,
Andrei
-- 
http://wiki.debian.org/FAQsFromDebianUser


signature.asc
Description: PGP signature

Re: How free is Debian

[John Hasler]

I wrote:
> I'd just flat out restrict them to five years.  Twenty is too long.

Andrei writes:
> That would work against inventors as instead of buying useful patents
> companies would just wait 5 years and then use it without any charge.

Anyone can practice an expired patent without any charge.  That's why
they expire.  Companies could wait 20 years now but they usually don't:
it's worth money to be in the market now.  If it was 5 years they'd pay
less, but they'd still pay.

-- 
John Hasler 
jhas...@newsguy.com
Elmwood, WI USA

Re: How free is Debian

[Joe]

On Sat, 10 Aug 2019 10:52:39 +0300
Andrei POPESCU  wrote:

> On Jo, 08 aug 19, 19:05:03, John Hasler wrote:
> > deloptes writes:
> >   
> > > Entirely eliminate [patents] - no, but restrict if no commercial
> > > use to 5y.  
> > 
> > I'd just flat out restrict them to five years.  Twenty is too long.
> >  
> 
> That would work against inventors as instead of buying useful patents 
> companies would just wait 5 years and then use it without any charge.
> 

That was all that patents were ever intended to do: give the originator
a period of monopoly over that particular implementation of his idea.
It never covered different implementations.

That was in exchange for the inventor freely showing the world
('patent') how his device worked, precisely so that others could use
it later. It was never intended to grant a monopoly for half a working
lifetime, it was intended to discourage producers from keeping a
manufacturing process secret, which of course they have always been
entitled to do.

Closed-source code is secret anyway, definitely *not* patent, and ideas
cannot be patented, so the whole business is simply a demonstration of
the power of Microsoft, just as Disney dominates US copyright law.

-- 
Joe

Re: How free is Debian

[Andrei POPESCU]

On Jo, 08 aug 19, 09:20:29, The Wanderer wrote:
> 
> As far as I'm aware, there are significantly more projects out there to
> produce free motherboard firmware (BIOS / UEFI images) than there are to
> produce free firmware for any of those other things.

One has to start somewhere and the BIOS seems like a logical choice to 
me.

Kind regards,
Andrei
-- 
http://wiki.debian.org/FAQsFromDebianUser


signature.asc
Description: PGP signature

Re: How free is Debian

[Andrei POPESCU]

On Jo, 08 aug 19, 19:05:03, John Hasler wrote:
> deloptes writes:
> 
> > Entirely eliminate [patents] - no, but restrict if no commercial use
> > to 5y.
> 
> I'd just flat out restrict them to five years.  Twenty is too long.

That would work against inventors as instead of buying useful patents 
companies would just wait 5 years and then use it without any charge.

Kind regards,
Andrei
-- 
http://wiki.debian.org/FAQsFromDebianUser


signature.asc
Description: PGP signature

Re: How free is Debian

[John Hasler]

t writes:
> There are many countries where such (so-called "shrink-wrap licenses",
> because you have to tear the package open to discover it) aren't
> legally binding.

A true shrink-wrap "license" is one that is visible and readable through
the transparent shrink-wrap package.
-- 
John Hasler 
jhas...@newsguy.com
Elmwood, WI USA

Re: How free is Debian

[Alessandro Vesely]

On Thu 08/Aug/2019 13:50:40 +0200 John Hasler wrote:

> tomas writes:
>> This is one of those cases: if you're using a piece of non-free
>> software, you should know about it, and you should know which buy
>> decision led to it (so you can take that into account at your next buy
>> decision).
> 
> There is also a practical reason to keep non-free for the benefit of
> downstream distributions, CD makers, etc.  Some of the licenses on stuff
> in non-free make it ok for Debian to distribute the stuff but attempt to
> place restrictions on what recipients can do with it.  As long as you
> stick to Main you need only read the DFSG to know what your
> redistribution rights are.  As soon as you go into Non-free you have to
> study each license.
> 
> This can even hit end-users.  Non-free licenses can contain clauses
> barring "commercial use" (without defining the term) and other similar
> restrictions.  This package is not in Debian, but I recall a "free" text
> editor that was distributed on the Net back in the last century that
> barred use by the South African police.  It would have qualified for
> inclusion in Non-free.


Perhaps we'd need an Almost-free, or Semantically-free category in
order to distinguish blind redistributablity from actual
devil-may-be-here stuff.  I'm thinking GNU doc in particular...


Best
Ale

Re: How free is Debian

[John Hasler]

Alessandro Vesely writes:
> I had always considered those must-reply-yes questions akin to
> extortion.  My answer reflects the only possibility to use something I
> bought, not my free thought.

If the contract is presented before the sale is completed they are not
different from "You must agree to pay us the stated price".  If the
contract is presented afterward they are bluffing.

Determining at what point the sale is completed may not be trivial,
though.
-- 
John Hasler 
jhas...@newsguy.com
Elmwood, WI USA

Re: How free is Debian

[tomas]

On Fri, Aug 09, 2019 at 05:59:26PM +0200, Alessandro Vesely wrote:
> On Thu 08/Aug/2019 15:02:38 +0200 John Hasler wrote:
> 
> > These sorts of "licenses" are actually attempts at a civil contract.
> > They really have nothing to do with patent or copyright law.  A civil
> > contract requires agreement in advance, though.  
> 
> 
> I had always considered those must-reply-yes questions akin to
> extortion.  My answer reflects the only possibility to use something I
> bought, not my free thought.

There are many countries where such (so-called "shrink-wrap licenses",
because you have to tear the package open to discover it) aren't
legally binding.

Cheers
-- t


signature.asc
Description: Digital signature

Re: How free is Debian

[Alessandro Vesely]

On Thu 08/Aug/2019 15:02:38 +0200 John Hasler wrote:

> These sorts of "licenses" are actually attempts at a civil contract.
> They really have nothing to do with patent or copyright law.  A civil
> contract requires agreement in advance, though.  


I had always considered those must-reply-yes questions akin to
extortion.  My answer reflects the only possibility to use something I
bought, not my free thought.


Best
Ale

Re: How free is Debian

[Celejar]

On Fri, 09 Aug 2019 08:51:15 -0500
John Hasler  wrote:

> Celejar writes:
> > It does? Here's what the "Debian Position on Software Patents" says:
> 
> > Debian will not knowingly distribute software encumbered by patents;
> > Debian contributors should not package or distribute software they know
> > to infringe a patent.
> 
> The key word is "knowingly".  If you believe that a Debian package
> infringes a patent of yours you have to speak up.

You've snipped the context. You wrote:

> statutory damages as with copyright).  Thus while there is no statutory
> exemption for personal use there is an effective one.  Illustrative is
> how Debian deals with software patents: it ignores them.

So no, Debian does not "ignore" patents. Perhaps what you meant to say
was that Debian does not take the initiative of investigating whether a
given package infringes on any patents. And I still don't see how this
is illstrative of an effective exemption for personal use.

Celejar

Re: How free is Debian

[John Hasler]

Celejar writes:
> It does? Here's what the "Debian Position on Software Patents" says:

> Debian will not knowingly distribute software encumbered by patents;
> Debian contributors should not package or distribute software they know
> to infringe a patent.

The key word is "knowingly".  If you believe that a Debian package
infringes a patent of yours you have to speak up.
-- 
John Hasler 
jhas...@newsguy.com
Elmwood, WI USA

Re: How free is Debian

[Celejar]

On Thu, 08 Aug 2019 19:05:03 -0500
John Hasler  wrote:

...

> Also realize that patent infringement is not a crime (in the USA).
> Government won't enforce your patent for you.  It is a tort, and is
> grounds for the patent owner to sue for actual damages (there are no
> statutory damages as with copyright).  Thus while there is no statutory
> exemption for personal use there is an effective one.  Illustrative is
> how Debian deals with software patents: it ignores them.

It does? Here's what the "Debian Position on Software Patents" says:

> Debian will not knowingly distribute software encumbered by patents;
> Debian contributors should not package or distribute software they know
> to infringe a patent.

https://www.debian.org/legal/patent

Celejar

Re: How free is Debian

[Andrew McGlashan]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA256

Hi,

On 9/8/19 6:59 am, John Hasler wrote:
> Shahryar Afifi wrote:
>> Currently I have X61 with Middleton BIOS that claims to be free.
>> Is that also not the case?
> 
> We are talking about the microcode that is stored inside the cpu,
> not the BIOS which is x86 code stored in NVRAM on the motherboard.
> In Intel 64 bit cpus this currently includes an entire operating
> system.  It can completely bypass Linux without you even knowing
> that it is there.  Very useful for large organizations with
> thousands of machines that they need to centrally administer,
> but...

AIUI, when the machine boots, the BIOS loads the CPU microcode on
machine startup *before* Linux (or any other OS for that matter) boot
processes take over.  That is, every time the machine is booted, from
cold at least, the CPU microcode is updated from an area within the BIOS
.

If you have a newer microcode, then that /might/ be applied at the OS
level at some point.

Cheers
A.
-BEGIN PGP SIGNATURE-

iHUEAREIAB0WIQTJAoMHtC6YydLfjUOoFmvLt+/i+wUCXU0DPQAKCRCoFmvLt+/i
+8TMAP4xVQx0IPg4k3UHAcyHc8AaSKnOZTwaP677Wd5Xxa+6cwEAtxAopo7fzsJU
D0uXVWPk2Xp+01ucq3Jc3s5x3ffR3AU=
=wa3g
-END PGP SIGNATURE-

Re: How free is Debian

[Gene Heskett]

On Thursday 08 August 2019 21:17:25 Ben Caradoc-Davies wrote:

> On 09/08/2019 12:05, John Hasler wrote:
> > There is a lot wrong with the patent system.  Twenty years is too
> > long. Fees are too high.  Processing is too slow.  The language used
> > in the disclosures is arcane (the disclosure is supposed to teach
> > the invention to someone "skilled in the art", but the art one must
> > be skilled in is patent law).  Software patents not only exist, but
> > are described in *incredibly*  arcane language rather than in source
> > code.  Pharmaceutical patent law is even more bizarre than software
> > patent law (mostly due to the efforts of Congress to "fix" it).
>
> New Zealand banned pure software patents in 2013:
>
> How New Zealand banned software patents without violating
> international law
> https://qz.com/119419/how-new-zealand-banned-software-patents-without-
>violating-international-law/
>
> Kind regards,

Sounds like a great idea. Too bad Disney owns both offices today.

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
If we desire respect for the law, we must first make the law respectable.
 - Louis D. Brandeis
Genes Web page 

Re: How free is Debian

[Ben Caradoc-Davies]

On 09/08/2019 12:05, John Hasler wrote:

There is a lot wrong with the patent system.  Twenty years is too long.
Fees are too high.  Processing is too slow.  The language used in the
disclosures is arcane (the disclosure is supposed to teach the invention
to someone "skilled in the art", but the art one must be skilled in is
patent law).  Software patents not only exist, but are described in
*incredibly*  arcane language rather than in source code.  Pharmaceutical
patent law is even more bizarre than software patent law (mostly due to
the efforts of Congress to "fix" it).


New Zealand banned pure software patents in 2013:

How New Zealand banned software patents without violating international law
https://qz.com/119419/how-new-zealand-banned-software-patents-without-violating-international-law/

Kind regards,

--
Ben Caradoc-Davies 
Director
Transient Software Limited 
New Zealand

Re: How free is Debian

[John Hasler]

deloptes writes:
> So you mean after AMD reverse engineered Intel, they make it illegal
> for their own, by license agreement?

They can't make it illegal.  They can make it a breach of contract for
anyone who agreed to their terms to reverse engineer it and sue them for
damages if they do.

I can go over to UW Stout surplus, find a machine with an AMD cpu in it,
plunk down a fifty, and walk out owning it free and clear.  I then have no
contract with AMD and can reverse-engineer the my heart's content.

> Entirely eliminate [patents] - no, but restrict if no commercial use
> to 5y.

I'd just flat out restrict them to five years.  Twenty is too long.

> Few years ago when I was reading I had the impression as long as you
> can pay for the patent rights you are not forced to
> implement/productize your patent.

Of course not.  There is no fair way to do that.

> So the bastards are gathering tons of patents in their basements.

"The oil companies have bought up all the 100 mpg carburetor patents and
supressed them."  Bullshit.  First of all, patents are public.  Produce
a list of such patents (ones that are being *enforced*).  Secondly,
companies want to make money.  If there is a way to make money
exploiting, licensing, or selling a patent they will do it.

Companies such as IBM often own rafts of patents which they are *not*
enforcing (they are also often pretty iffy patents).  They use them to
strengthen their bargaining position should another company in the same
business sue them for patent enfringement.

> First of all it is very expensive to protect your rights worldwide...

In the USA you can get a provisional patent for $100.  This protects
your invention for one year, giving you time to find financing for a
full patent.  The filing requirements are fairly informal.  There are
also price breaks for individuals for utility patents and a sort of a
time payment plan rather than a large up-front fee.

> ...if you are a simple person with a great idea, there is no way to
> make money from it.

You can start a business to exploit it or you can sell all rights to the
highest bidder.  No, you won't make millions.  On the other hand,
there's a good chance that the buyer won't make anything.  Most patents
don't turn out to be worth anything.

I know of two local people who did very, very well from "basement"
patents, though.

> AFAIK it was different some 150y ago.

Yes.  The application fee (to be paid in cash at the time of
application) was more than a year's income for the average American.

There is a lot wrong with the patent system.  Twenty years is too long.
Fees are too high.  Processing is too slow.  The language used in the
disclosures is arcane (the disclosure is supposed to teach the invention
to someone "skilled in the art", but the art one must be skilled in is
patent law).  Software patents not only exist, but are described in
*incredibly* arcane language rather than in source code.  Pharmaceutical
patent law is even more bizarre than software patent law (mostly due to
the efforts of Congress to "fix" it).

Also realize that patent infringement is not a crime (in the USA).
Government won't enforce your patent for you.  It is a tort, and is
grounds for the patent owner to sue for actual damages (there are no
statutory damages as with copyright).  Thus while there is no statutory
exemption for personal use there is an effective one.  Illustrative is
how Debian deals with software patents: it ignores them.
-- 
John Hasler 
jhas...@newsguy.com
Elmwood, WI USA

Re: How free is Debian

[deloptes]

John Hasler wrote:

>> The irony here is that AMD started by reverse engineering Intel.
> 
> Legally.  Reverse-engineering is not illegal in the USA.
> 

>> And unfortunately the US has been protecting monopoly and fake
>> competition for years.  Such things as Microsoft, Apple and Google
>> should not exist, not to speak of Intel, IBM and many other
>> monsters. Amazon, Uber ... many many of them - it is cancer.  Some
>> time ago I read good article why the Patent Law should change, but I
>> forgot where I found the article. The problems are in the patent law,
>> as I understood the article
> 
> Patent law certainly needs to change (I'm not sure that it should not be
> entirely eliminated) but US patent law places no restrictions on reverse
> engineering, decompiling, or disassembling software.  US copyright law
> bars bypassing of copy protection schemes, but there is actually an
> exemption for reverse-engineering under some circumstances.  Printer and
> game manufacturers tried to use these "anti-piracy" provisions of the DMCA
> to block reverse-engineering for compatibility purposes and lost in
> court.
> 
> These sorts of "licenses" are actually attempts at a civil contract.
> They really have nothing to do with patent or copyright law.  A civil
> contract requires agreement in advance, though.

So you mean after AMD reverse engineered Intel, they make it illegal for
their own, by license agreement? Well here the word own is may be not
exactly in place :) and this is the irony I meant.

I think this is big topic and no attention and awareness in the public. IMO
it is even worse than climate change.

Entirely eliminate it - no, but restrict if no commercial use to 5y. Few
years ago when I was reading I had the impression as long as you can pay
for the patent rights you are not forced to implement/productize your
patent. So the bastards are gathering tons of patents in their basements. I
personally know few cases. First of all it is very expensive to protect
your rights worldwide or even in the major countries and second if you are
a simple person with a great idea, there is no way to make money from it.
AFAIK it was different some 150y ago.


 

Re: How free is Debian

[John Hasler]

Shahryar Afifi wrote:
> Currently I have X61 with Middleton BIOS that claims to be free. Is
> that also not the case?

We are talking about the microcode that is stored inside the cpu, not
the BIOS which is x86 code stored in NVRAM on the motherboard.  In Intel
64 bit cpus this currently includes an entire operating system.  It can
completely bypass Linux without you even knowing that it is there.  Very
useful for large organizations with thousands of machines that they
need to centrally administer, but...
-- 
John Hasler 
jhas...@newsguy.com
Elmwood, WI USA

Re: How free is Debian

[John Hasler]

Brad writes:
> A brief internet search resulted in me reading an article stating
> that, in the USA at least, the EULA supersedes the user's legal
> allowance for reverse engineering (case cited in the article) the
> software.

> So, when a person agrees to the terms of the EULA, they waive their
> legal right to reverse engineer.  If you wish to NOT waive your
> rights, then you don't accept the EULA.  Of course you then won't be
> able to install, never mind use, the software.  Catch-22.

An EULA is a civil contract having nothing to do with patent or
copyright law.  You agree to it as a condition of sale: in addition to
promising to pay the seller some money you also agree to refrain from
certain actions such as reverse engineering or selling your purchased
copy that copyright law alone would permit you to do[1].  Of course, a
contract cannot be imposed after the transaction is complete.  The
(shaky) theory of the shrinkwrap license is that you can read the
contract before you open the package and can then return the unopened
package for a refund if you don't like it.  Before downloading[2]
purchased software you usually have the opportunity to read the
contract.  However, when you download[2] a package from Debian Non-free
with full permission from the copyright owner and only later discover
a "license"[3] that purports to revoke some of the rights you would
otherwise have as a legitimate owner of a copy it's too late to
establish a contract (other elements of a contract are also missing).


[1]  You do *not* need any sort of license to use, sell, or give away a
copy that you legitimately own.  Note, however, that under copyright law
a "copy" is a tangible object   such as a printed book, a CD, or a
portion of your hard drive.  Copyright law limits the making of copies,
not the disposition of them.  It also grants owners of copies of
software limited rights to make copies (but not to distribute them).

[2] When you "download a copy" you create a physical, tangible copy that
you own in the same sense as you would by copying the material off the
screen with a pen and paper that you own.

[3] This is almost exactly the same as the original "first sale
doctrine" case in which a publisher printed a statement on the title
page of a book saying that in purchasing a copy of the book the buyer
agreed not to resell the copy.  That publisher lost in court.
-- 
John Hasler 
jhas...@newsguy.com
Elmwood, WI USA

Re: How free is Debian

[Andrew McGlashan]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA256



On 8/8/19 11:06 pm, Shahryar Afifi wrote:
> Thank you for this acknowledgment. Currently I have X61 with
> Middleton BIOS that claims to be free. Is that also not the case?

You can have a free BIOS, "Core boot, or similar?" ... but the CPU
itself has it's own non-free microcode.  Again, the display will have
it's own code, you might drive that display with free software, but
the display won't work fully and properly for protected content if it
doesn't validate the display as being compliant; heck even the display
cable can be a problem.

Again, if you want total free, then it isn't going to be easy to
achieve and it would be impossible with Intel CPUs unless you
don't consider the built in microcode to be a problem.  However, if
you do have an Intel CPU, then you almost certainly want the latest
microcode unless you don't trust it -- it may add more spyware or it
may fix security issues or both you will probably never know.

Cheers
A.

-BEGIN PGP SIGNATURE-

iHUEAREIAB0WIQTJAoMHtC6YydLfjUOoFmvLt+/i+wUCXUw19AAKCRCoFmvLt+/i
+xHbAP4gQatf0HhebBuFL99pPjgTishWOEiTGBxETXRCRbF2ywD9H+DZGQVnA00g
bxmvN5+La1l/ZsJrCeat+pbLiOtVUt4=
=lC0I
-END PGP SIGNATURE-

Re: How free is Debian

[Celejar]

On Thu, 08 Aug 2019 06:22:41 -0500
John Hasler  wrote:

>  Shahryar Afifi writes:
> > Very well said. If debian free is not using amd64 microcode, so what
> > kernel module runs my cpu as 64bit?
> 
> That microcode package contains bug fixes and updates for the microcode
> that the manufacturer shipped the cpu with.  The cpu will run without it
> but the spyware coded into it[1] might not work right.
> 
> [1] Or maybe there isn't any and they are really just fixing bugs that
> could be security risks.  Or maybe both.

I understand that you're being somewhat facetious, but it should be
understood that updating microcode should be taken seriously, according
to Debian developer and linux microcode expert Henrique de Maraes
Holschuh:

> Intel® 64 and IA-32 processors (x86_64 and i686 processors) are capable of
> field-upgrading their control program (microcode) as well as parameters
> for other on-chip subsystems (power management, interconnects, etc).
> These microcode updates correct processor errata, and are important for
> safe, stable and correct system operation.
> 
> While most of the microcode updates fix problems that happen extremely
> rarely, they also fix high-profile, high-hitting issues.  There are enough
> microcode updates fixing processor errata that would cause system lockup,
> memory corruption, or unpredictable system behavior, to warrant taking
> firmware updates and microcode updates seriously.

https://metadata.ftp-master.debian.org/changelogs/non-free/i/intel-microcode/unstable_README.Debian

Celejar

Re: How free is Debian

[Joe Pfeiffer]

John Hasler  writes:

> Joe Pfeiffer writes:
>> The LICENCE.amd-ucode file
>> includes the paragraph:
>
>>You may not reverse engineer, decompile, or disassemble this
>>Software or any portion thereof.
>
> Quite unenforceable, of course.

When discussing questions like "how free is this software", the question
is all about what's legal -- not what's practical or enforceable.

Re: How free is Debian

[Michael Stone]

On Thu, Aug 08, 2019 at 09:20:29AM -0400, The Wanderer wrote:

I'm not familiar with that particular product, but the BIOS isn't the
CPU firmware, although it may implicitly contain and apply files which
update the CPU firmware; the BIOS is the motherboard firmware, which is
a different kettle of fish.


In modern usage "BIOS" isn't as limited as it was on an IBM PC. E.g., on 
a modern AMD system the "BIOS" download from a vendor includes the 
AGESA, which is responsible for initializing the CPU. CPUs are simply 
more complex, so there are more bits involved with turning them on then 
there were in the old days. (At least for PCs--large systems in the old 
days had dedicated small computers to initialize the big computer; e.g., 
Cray systems used a sun workstation to manage initializiation. We've 
simply managed to miniaturize the small computer.)

Re: How free is Debian

[The Wanderer]

On 2019-08-08 at 09:06, Shahryar Afifi wrote:

> On Thu, 2019-08-08 at 17:37 +1000, Andrew McGlashan wrote:
> 
>> Hi,
>> 
>> On 8/8/19 2:27 pm, Shahryar Afifi wrote:
>> 
>>> Very well said. If debian free is not using amd64 microcode, so 
>>> what kernel module runs my cpu as 64bit?
>> 
>> Here's part of the problem.
>> 
>> The CPU has it's own microcode, when you buy it; the motherboard 
>> manufacturer may or may not update the BIOS and if they do, then
>> they may or may not provide an update to the microcode.
>> 
>> Now, you already have microcode, one way or another, using
>> non-free to get an update to the microcode is the best you can do
>> as you, theoretically know you can trust the source of the blob (as
>> much as you trust Debian AND Intel of course.
>> 
>> You can download microcode directly from Intel, it won't stop it 
>> being a binary blob though.
>> 
>> If you really need more TOTALLY FREE, then consider RISC-V now or
>> in the future; their may be other viable alternatives, but
>> something will always get you like for instance the blob of
>> firmware with your mouse or your monitor, even your keyboard.  And
>> yes there are open source keyboards, but I'm not sure about other
>> components.  Heck, even your HDD (of any type) has it's own
>> firmware blobs, some of which can only be sent to the device and
>> never read back.
>> 
>> So, I guess, totally free really means, give up on electronic gear 
>> altogether.

> Thank you for this acknowledgment. Currently I have X61 with
> Middleton BIOS that claims to be free. Is that also not the case?

I'm not familiar with that particular product, but the BIOS isn't the
CPU firmware, although it may implicitly contain and apply files which
update the CPU firmware; the BIOS is the motherboard firmware, which is
a different kettle of fish.

In addition to the separate objects which Andrew mentioned as
potentially each containing their own firmware blob, there tend to be
distinct firmware for each of the motherboard, CPU, GPU, network adapter
(especially if wireless), sound card, et cetera - many of which latter
may be integrated into the motherboard, but still have their own
firmware blobs, which are not under the control of the motherboard
manufacturer.

The motherboard firmware may well contain copies of the firmware blobs
for these integrated devices, and apply them as appropriate, but that
doesn't necessarily mean that the motherboard maker has any control over
what's in those blobs or what they do.

As far as I'm aware, there are significantly more projects out there to
produce free motherboard firmware (BIOS / UEFI images) than there are to
produce free firmware for any of those other things.

-- 
   The Wanderer

The reasonable man adapts himself to the world; the unreasonable one
persists in trying to adapt the world to himself. Therefore all
progress depends on the unreasonable man. -- George Bernard Shaw



signature.asc
Description: OpenPGP digital signature

Re: How free is Debian

[David Wright]

On Wed 07 Aug 2019 at 21:27:34 (-0700), Shahryar Afifi wrote:
> On Wed, 2019-08-07 at 23:11 -0500, David Wright wrote:
> > On Wed 07 Aug 2019 at 17:33:52 (-0700), Shahryar Afifi wrote:
> > > With respect to all the contributors, developers, hobbyist and
> > > users,
> > > who made GNU/Linux and Debian and all other distributions possible,
> > > here lies a humble, ignorance and yet curious question.
> > > 
> > > Are all binaries in the kernel code were writing from scratch? Are
> > > there any binary blobs in the kernel that it was given to
> > > developers?
> > > If amd64 license is not free, how is it that we have amd64
> > > microcode in
> > > the debian free?
> > 
> > It isn't free; look:

> > [ … snip …]

> > > and if they are not the same, are we using the full
> > > potential of our hardware?
> 
> Very well said. If debian free is not using amd64 microcode, so what
> kernel module runs my cpu as 64bit?

Here's an analogy:

You buy stretch on DVDs on 2017-06-17 and install just "main" on an
isolated machine to do a particular task. Two years later, the
installation is still working just as it was.

Over those two years, however, you could have upgraded the
installation with point-releases, and at any time you could have
chosen to add in non-free amd contrib. These would all have improved
the system (some would say polluted it in the case of non-free).
Most people would consider the point-releases as essential if you ever
connect the system to the Internet.

The first paragraph corresponds to your buying a CPU, which itself
contains what you could call an Operating System in its firmware,
and that OS will already be long in the tooth when you buy it
off the shelf. However, the CPU is still running today, and you expect
another ten years of solid performance.

The second paragraph corresponds to your downloading packages of
microcode and allowing them to be installed when you boot the CPU.
Some people would consider these microcode updates as essential if
the systems are to be connected to the Internet. Others do without.

Cheers,
David.

Re: How free is Debian

[tomas]

On Thu, Aug 08, 2019 at 06:06:24AM -0700, Shahryar Afifi wrote:
> On Thu, 2019-08-08 at 17:37 +1000, Andrew McGlashan wrote:
> > Hi,
> > 
> > On 8/8/19 2:27 pm, Shahryar Afifi wrote:
> > > Very well said. If debian free is not using amd64 microcode, so
> > > what kernel module runs my cpu as 64bit?
> > 
> > Here's part of the problem.
> > 
> > The CPU has it's own microcode [...]

> > So, I guess, totally free really means, give up on electronic gear
> > altogether.
> > 
> > Cheers
> > A.
> 
> Thank you for this acknowledgment.
> Currently I have X61 with Middleton BIOS that claims to be free. Is
> that also not the case?

It becomes more and more complicated the more you look into it.
Your harddisk? A complete subsystem with a microcontroller in
it, with some custom-made kind of operating system. Your WiFi
adapter? Same story. Sometimes they need firmware blobs, which
the operating system has to somehow make available to them.
Sometimes they do work without the blobs, sometimes the don't.
Sometimes they kind-of work, but have some security vulnerability.

Free is not an end state one can achieve, but a constant uphill
fight.

Or to put it another way, borrowing from somewhere else, "freedom
is not a product, it's a process".

Cheers
-- t


signature.asc
Description: Digital signature

Re: How free is Debian

[John Hasler]

Joe Pfeiffer wrote:
> You may not reverse engineer, decompile, or disassemble this
> Software or any portion thereof.

deloptes writes:
> The irony here is that AMD started by reverse engineering Intel.

Legally.  Reverse-engineering is not illegal in the USA.

> And unfortunately the US has been protecting monopoly and fake
> competition for years.  Such things as Microsoft, Apple and Google
> should not exist, not to speak of Intel, IBM and many other
> monsters. Amazon, Uber ... many many of them - it is cancer.  Some
> time ago I read good article why the Patent Law should change, but I
> forgot where I found the article. The problems are in the patent law,
> as I understood the article

Patent law certainly needs to change (I'm not sure that it should not be
entirely eliminated) but US patent law places no restrictions on reverse
engineering, decompiling, or disassembling software.  US copyright law
bars bypassing of copy protection schemes, but there is actually an
exemption for reverse-engineering under some circumstances.  Printer and
game manufacturers tried to use these "anti-piracy" provisions of the DMCA
to block reverse-engineering for compatibility purposes and lost in
court.

These sorts of "licenses" are actually attempts at a civil contract.
They really have nothing to do with patent or copyright law.  A civil
contract requires agreement in advance, though.  
-- 
John Hasler 
jhas...@newsguy.com
Elmwood, WI USA

Re: How free is Debian

[Curt]

On 2019-08-08, Brad Rogers  wrote:
>
> So, when a person agrees to the terms of the EULA, they waive their
> legal right to reverse engineer.  If you wish to NOT waive your rights,
> then you don't accept the EULA.  Of course you then won't be able to
> install, never mind use, the software.  Catch-22.
>

https://en.wikipedia.org/wiki/Sega_v._Accolade

...
 Sega v. Accolade has been an influential case in matters involving reverse
 engineering of software and copyright infringement, and has been cited in
 numerous cases since 1993.[3]:388[10] The case has redefined how reverse
 engineering with unlicensed products is seen in legal issues involving
 copyright. Legally, the decision concurred that the nature of Accolade's work
 in reverse engineering the Sega Genesis was to access ideas that were deemed
 unprotected by copyright law, and could only be accessed by decompiling.[12] By
 the verdict, the console's functional principles were established not to be
 protected by copyright,[5] and that when no other means were available, reverse
 engineering the copyrighted software to access information about the console's
 functional principles is protected by the fair use doctrine
...
 Sega v. Accolade also served to help establish that the functional principles
 of computer software cannot be protected by copyright law. Rather, the only
 legal protection to such principles can be through holding a patent or by trade
 secret.


-- 
“We are all in the gutter, but some of us are looking at the stars.” 
― Oscar Wilde, Lady Windermere's Fan

Re: How free is Debian

[Brad Rogers]

On Thu, 8 Aug 2019 07:51:00 -0400
rhkra...@gmail.com wrote:

Hello rhkra...@gmail.com,

>   * that statement / requirment is illegal (or not supported by (US?)
> law)

A brief internet search resulted in me reading an article stating that,
in the USA at least, the EULA supersedes the user's legal allowance for
reverse engineering (case cited in the article) the software.

See;


So, when a person agrees to the terms of the EULA, they waive their
legal right to reverse engineer.  If you wish to NOT waive your rights,
then you don't accept the EULA.  Of course you then won't be able to
install, never mind use, the software.  Catch-22.

-- 
 Regards  _
 / )   "The blindingly obvious is
/ _)radnever immediately apparent"
Kill joy, bad guy, big talking, small fry
Death On Two Legs - Queen


pgp_mBjMnUP5X.pgp
Description: OpenPGP digital signature

Re: How free is Debian

[John Hasler]

tomas writes:
> This is one of those cases: if you're using a piece of non-free
> software, you should know about it, and you should know which buy
> decision led to it (so you can take that into account at your next buy
> decision).

There is also a practical reason to keep non-free for the benefit of
downstream distributions, CD makers, etc.  Some of the licenses on stuff
in non-free make it ok for Debian to distribute the stuff but attempt to
place restrictions on what recipients can do with it.  As long as you
stick to Main you need only read the DFSG to know what your
redistribution rights are.  As soon as you go into Non-free you have to
study each license.

This can even hit end-users.  Non-free licenses can contain clauses
barring "commercial use" (without defining the term) and other similar
restrictions.  This package is not in Debian, but I recall a "free" text
editor that was distributed on the Net back in the last century that
barred use by the South African police.  It would have qualified for
inclusion in Non-free.
-- 
John Hasler 
jhas...@newsguy.com
Elmwood, WI USA

Re: How free is Debian

[rhkramer]

On Wednesday, August 07, 2019 11:03:46 PM John Hasler wrote:
> Joe Pfeiffer writes:
> > The LICENCE.amd-ucode file
> > 
> > includes the paragraph:
> >You may not reverse engineer, decompile, or disassemble this
> >Software or any portion thereof.
> 
> Quite unenforceable, of course.

When you say that, do you mean:

   * that statement / requirment is illegal (or not supported by (US?) law)
or
   * as a practical manner, somehow you don't think such actions could be 
detected 
or
   * something else (e.g., maybe that requirement is supported by US law, but 
could be done in some other country where it is not illegal?

Thanks!

Re: How free is Debian

[John Hasler]

 Shahryar Afifi writes:
> Very well said. If debian free is not using amd64 microcode, so what
> kernel module runs my cpu as 64bit?

That microcode package contains bug fixes and updates for the microcode
that the manufacturer shipped the cpu with.  The cpu will run without it
but the spyware coded into it[1] might not work right.

[1] Or maybe there isn't any and they are really just fixing bugs that
could be security risks.  Or maybe both.
-- 
John Hasler 
jhas...@newsguy.com
Elmwood, WI USA

Re: How free is Debian

[tomas]

On Thu, Aug 08, 2019 at 08:50:05AM +0200, john doe wrote:

[...]

> None-free is a repository that you enable if you need to, it is not the
> default in Debian if I am not mistaking.

This is correct, and I, for one am glad Debian has kept up to pressure
to changing it ("but... but... Ubuntu can do, why can't Debian?").

Sometimes freedom isn't to be had without knowledge. It can be sometimes
inconvenient, but we're here to help each other and try our best...

This is one of those cases: if you're using a piece of non-free software,
you should know about it, and you should know which buy decision led to
it (so you can take that into account at your next buy decision).

Soapbox free now. Next, please ;-)

Cheers
-- t


signature.asc
Description: Digital signature

Re: How free is Debian

[Andrew McGlashan]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA256

Hi,

On 8/8/19 2:27 pm, Shahryar Afifi wrote:
> Very well said. If debian free is not using amd64 microcode, so
> what kernel module runs my cpu as 64bit?

Here's part of the problem.

The CPU has it's own microcode, when you buy it; the motherboard
manufacturer may or may not update the BIOS and if they do, then they
may or may not provide an update to the microcode.

Now, you already have microcode, one way or another, using non-free to
get an update to the microcode is the best you can do as you,
theoretically know you can trust the source of the blob (as much as
you trust Debian AND Intel of course.

You can download microcode directly from Intel, it won't stop it being
a binary blob though.

If you really need more TOTALLY FREE, then consider RISC-V now or in
the future; their may be other viable alternatives, but something will
always get you like for instance the blob of firmware with your
mouse or your monitor, even your keyboard.  And yes there are open
source keyboards, but I'm not sure about other components.  Heck, even
your HDD (of any type) has it's own firmware blobs, some of which can
only be sent to the device and never read back.

So, I guess, totally free really means, give up on electronic gear
altogether.

Cheers
A.
-BEGIN PGP SIGNATURE-

iHUEAREIAB0WIQTJAoMHtC6YydLfjUOoFmvLt+/i+wUCXUvRTgAKCRCoFmvLt+/i
+yQ8AP9hVD8LeUMvRA2toVzn4GnyqKWMQfu6Ye1vC5Sgnh3w+wEA3qWXdubgbGeS
0UAt2uTQ/e55wjnS43qws+Da1+fbSTI=
=sWho
-END PGP SIGNATURE-

Re: How free is Debian

[john doe]

On 8/8/2019 6:11 AM, David Wright wrote:
> On Wed 07 Aug 2019 at 17:33:52 (-0700), Shahryar Afifi wrote:
>> With respect to all the contributors, developers, hobbyist and users,
>> who made GNU/Linux and Debian and all other distributions possible,
>> here lies a humble, ignorance and yet curious question.
>>
>> Are all binaries in the kernel code were writing from scratch? Are
>> there any binary blobs in the kernel that it was given to developers?
>> If amd64 license is not free, how is it that we have amd64 microcode in
>> the debian free?
>
> It isn't free; look:
>
> Package: amd64-microcode
> Version: 3.20160316.3
> Installed-Size: 68
> Maintainer: Henrique de Moraes Holschuh 
> Architecture: amd64
> Recommends: initramfs-tools (>= 0.113~) | dracut (>= 044) | tiny-initramfs
> Breaks: intel-microcode (<< 2)
> Description: Processor microcode firmware for AMD CPUs
> Description-md5: 093f190e183c7cfeca05b52ecd2116e3
> Section: non-free/admin
>  
> Priority: extra
> Filename: 
> pool/non-free/a/amd64-microcode/amd64-microcode_3.20160316.3_amd64.deb
>
> Size: 31116
> MD5sum: 7056e449d8bac87d85a4e434379d0e6e
> SHA256: f7bddaf712ffaa833ff65ef94bdd86720d55c2c56ae982c3db58181bbe70f147
>

None-free is a repository that you enable if you need to, it is not the
default in Debian if I am not mistaking.

--
John Doe

Re: How free is Debian

[deloptes]

Shahryar Afifi wrote:

> Very well said. If debian free is not using amd64 microcode, so what
> kernel module runs my cpu as 64bit?

I was thinking the CPU is running and not something else running the CPU.
I do not think you need something special to run 64bit CPU as such.

Re: How free is Debian

[deloptes]

Joe Pfeiffer wrote:

> You may not reverse engineer, decompile, or disassemble this
> Software or any portion thereof.

The irony here is that AMD started by reverse engineering Intel.

And unfortunately the US has been protecting monopoly and fake competition
for years.
Such things as Microsoft, Apple and Google should not exist, not to speak of
Intel, IBM and many other monsters. Amazon, Uber ... many many of them - it
is cancer.
Some time ago I read good article why the Patent Law should change, but I
forgot where I found the article. The problems are in the patent law, as I
understood the article 

regards

Re: How free is Debian

[David Wright]

On Wed 07 Aug 2019 at 17:33:52 (-0700), Shahryar Afifi wrote:
> With respect to all the contributors, developers, hobbyist and users,
> who made GNU/Linux and Debian and all other distributions possible,
> here lies a humble, ignorance and yet curious question.
> 
> Are all binaries in the kernel code were writing from scratch? Are
> there any binary blobs in the kernel that it was given to developers?
> If amd64 license is not free, how is it that we have amd64 microcode in
> the debian free?

It isn't free; look:

Package: amd64-microcode
Version: 3.20160316.3
Installed-Size: 68
Maintainer: Henrique de Moraes Holschuh 
Architecture: amd64
Recommends: initramfs-tools (>= 0.113~) | dracut (>= 044) | tiny-initramfs
Breaks: intel-microcode (<< 2)
Description: Processor microcode firmware for AMD CPUs
Description-md5: 093f190e183c7cfeca05b52ecd2116e3
Section: non-free/admin
 
Priority: extra
Filename: pool/non-free/a/amd64-microcode/amd64-microcode_3.20160316.3_amd64.deb
   
Size: 31116
MD5sum: 7056e449d8bac87d85a4e434379d0e6e
SHA256: f7bddaf712ffaa833ff65ef94bdd86720d55c2c56ae982c3db58181bbe70f147

> and if they are not the same, are we using the full
> potential of our hardware?

Cheers,
David.

Re: How free is Debian

[John Hasler]

Joe Pfeiffer writes:
> The LICENCE.amd-ucode file
> includes the paragraph:

>You may not reverse engineer, decompile, or disassemble this
>Software or any portion thereof.

Quite unenforceable, of course.
-- 
John Hasler 
jhas...@newsguy.com
Elmwood, WI USA

Re: How free is Debian

[Joe Pfeiffer]

Shahryar Afifi  writes:

> With respect to all the contributors, developers, hobbyist and users,
> who made GNU/Linux and Debian and all other distributions possible,
> here lies a humble, ignorance and yet curious question.
>
> Are all binaries in the kernel code were writing from scratch? Are
> there any binary blobs in the kernel that it was given to developers?
> If amd64 license is not free, how is it that we have amd64 microcode in
> the debian free? and if they are not the same, are we using the full
> potential of our hardware?
>
> I apologize in advance for my ignorance.
> Thank you.

Typically the binary blobs are not free.  If you get the source for a
package that includes a blob (for instance, amd64-microcode) you'll see
where the blob came from.  In the case of that package, it's all just
binary -- no source code for the microcode.  The LICENCE.amd-ucode file
includes the paragraph:

You may not reverse engineer, decompile, or disassemble this
Software or any portion thereof.

So... not free at all.

Re: How free is Debian

[Judah Richardson]

You don't need a license for an ISA to compile for it. You need a license
only if you're developing a CPU that uses that ISA.

On Wed, Aug 7, 2019, 19:34 Shahryar Afifi  wrote:

> With respect to all the contributors, developers, hobbyist and users,
> who made GNU/Linux and Debian and all other distributions possible,
> here lies a humble, ignorance and yet curious question.
>
> Are all binaries in the kernel code were writing from scratch? Are
> there any binary blobs in the kernel that it was given to developers?
> If amd64 license is not free, how is it that we have amd64 microcode in
> the debian free? and if they are not the same, are we using the full
> potential of our hardware?
>
> I apologize in advance for my ignorance.
> Thank you.
>
>