Guix on a microkernel

Guix on a microkernel

[mikadoZero]

# Appreciation

I appreciate:

* many of Guix's design decisions.  The one that is relevant to this
  discussion is the kernel.  I like that Guix uses the linux-libre (no
  binary blobs) instead of the linux kernel.

* that work is underway to get Guix to work with GNU Hurd.  I like that
  a microkernel is a potential kernel option. 

  http://lists.gnu.org/archive/html/guix-devel/2016-12/msg00857.html

* the effort that has been put into GNU Hurd to get it to where it is.

* the bootsrapping initiative.

  https://bootstrappable.org/
  https://fosdem.org/2019/schedule/event/gnumes/

# Intent

* I would like to understand why GNU Hurd is being focused
  on (my perception) given other microkernel options. 
* I want to share what I have found after doing some looking into
  microkernels.
* I am curious what others think of microkernels.

I mention the appreciations above because I am aiming for a tone of
appreciation and curiosity and not a critical one.  The tone can be a
challenge for written communication. 

# My microkernel experience

Currently I do not have any practical experience using any microkernel.
I have just spent time looking into the topic as it is interesting to
me.

# Why microkernels?

I think Andrew Tanenbaum explains benefits of microkernel entertainingly
in this talk:
http://www.youtube.com/watch?v=bx3KuE7UjGA

The talks has a focus on Minix but I think the benefits are transferable
to other microkernels.

# GNU Hurd

## Perceived focus

I looks to me like there is a effort (which I appreciate) to get Guix
working on Hurd.  I get this perception from:

http://lists.gnu.org/archive/html/guix-devel/2016-12/msg00857.html

These comments from this thread:

https://lists.gnu.org/archive/html/help-guix/2019-03/msg00158.html

Ricardo Wurmus:  "Let’s work on the Hurd, people!  It’s beautiful!"

Jan Nieuwenhuizen: "FWIW the Mes port to the Hurd is ongoing and mes now
runs, next thing up is fork which we need for running mescc."

## Critiques of Hurd

I would be curious what people think about these third party critiques
(not mine) of Hurd.

### From X15

https://www.sceen.net/x15/

"Although the design of the Hurd is promising and attractive, its
implementation has a number of severe issues. X15 takes the approach of
the complete rewrite to make sure that key ideas are kept in mind at all
times during development. Since it’s not meant to be compatible with the
Hurd, critical interfaces such as IPC and signals can be re-implemented
completely differently. There is a lot of emphasis on code quality and
ease of maintenance, obtained from disciplined application of best
practices."

### From HelenOS

http://www.helenos.org/wiki/FAQ#HowisHelenOSdifferentfromGNUHurd

### Why Hurd?

Why the focus on Hurd given other microkernel options?  I ask this
question out of curiosity and a lack of practical experience with
microkernels.

# Microkernel wish list

These are things that I see as desirable in a microkernel.

## Free software

It should be completely free software.  No binary blobs included.  It
looks like all of the microkernel listed here are:
http://www.microkernel.info/

## RISC-V

RISC-V a free and open instruction set architecture is a nice complement
to a free operating system.  It is nice if a mircokernel already has
plans to run on RISC-V. 

Intel security issues:
https://libreboot.org/faq.html#intel

ARM security issues:
https://libreboot.org/faq.html#amd

### Entirely free RISC-V computers

These two initiatives are entirely free hardware based on RISC-V.

* HiFive Unleashed
  https://www.sifive.com/boards/hifive-unleashed

* lowRISC
  https://www.lowrisc.org

## Formal verification

An application of the minimality principle in the design of microkernel
leads to smaller code bases which are amenable to formal verification. 

https://en.wikipedia.org/wiki/Microkernel#Essential_components_and_minimality

I see the extra security assurance that formal verification provide as
desirable.

# Alternative microkernels

I used http://www.microkernel.info/ as the starting point when I began
looking into microkernels. 

## Summary of interesting microkernels

This is a high level summary based on the "Microkernel wish list" above.
All of these are free software.  I am likely missing some other
interesting microkernel projects.

| projects     | RISC-V efforts | Formal verification |
|--------------+----------------+---------------------|
| sel4.systems | Yes            | Yes                 |
| genode.org   | Yes            | Yes                 |
| helenos.org  | Yes            | No                  |
| muen.sk      | ?/No           | Yes                 |
| minix3.org   | ?/No           | No                  |
| hurd.gnu.org | ?/No           | No                  |

Note:

* ?/No is where (to me without asking) there does not look like there
  have been efforts to make the project work with RISC-V. 

* Genode is different than the others as it is not just a microkernel.
  I have given Genode Yes for both RISC-V and Formal verification
  because it can use the seL4 microkernel.  It can also use other
  microkernels beyond just seL4.

## Other interesting projects

robigalia.org:  based on seL4 microkernel which is formally verified and
has RISC-V efforts underway.  It is using Rust to build the parts that
would normally be part of a monolithic kernel in user space.  It looks
like a young project.

redox-os.org:  Rust based microkernel project.  It looks like a young
project. 

## Projects I have not looked into

I have not looked at the following projects which were also listed on
http://www.microkernel.info/ 

* github.com/Nils-TUD/Escape
* github.com/f9micro
* l4re.org
* github.com/TUD-OS/M3
* hypervisor.org

Re: Guix on a microkernel

[znavko]

[-- Attachment #1: Type: text/plain, Size: 7421 bytes --]

Thanks for your compilation.
Do you have found actual benchmark tests? 
https://www.gnu.org/software/hurd/faq/slow.html <https://www.gnu.org/software/hurd/faq/slow.html>
"The Hurd is currently slower than Linux, yes. But not very much, so it is completely usable."
Vulnerabilities of processors is also sensitive task. Maybe RISC-V will not have such bugs? Need to know in a practice.

Mar 31, 2019, 12:05 AM by mikadozero@yandex.com:

> # Appreciation
>
> I appreciate:
>
> * many of Guix's design decisions.  The one that is relevant to this
>  discussion is the kernel.  I like that Guix uses the linux-libre (no
>  binary blobs) instead of the linux kernel.
>
> * that work is underway to get Guix to work with GNU Hurd.  I like that
>  a microkernel is a potential kernel option. 
>
>  > http://lists.gnu.org/archive/html/guix-devel/2016-12/msg00857.html <http://lists.gnu.org/archive/html/guix-devel/2016-12/msg00857.html>
>
> * the effort that has been put into GNU Hurd to get it to where it is.
>
> * the bootsrapping initiative.
>
>  > https://bootstrappable.org <https://bootstrappable.org/>
>  > https://fosdem.org/2019/schedule/event/gnumes <https://fosdem.org/2019/schedule/event/gnumes/>
>
> # Intent
>
> * I would like to understand why GNU Hurd is being focused
>  on (my perception) given other microkernel options. 
> * I want to share what I have found after doing some looking into
>  microkernels.
> * I am curious what others think of microkernels.
>
> I mention the appreciations above because I am aiming for a tone of
> appreciation and curiosity and not a critical one.  The tone can be a
> challenge for written communication. 
>
> # My microkernel experience
>
> Currently I do not have any practical experience using any microkernel.
> I have just spent time looking into the topic as it is interesting to
> me.
>
> # Why microkernels?
>
> I think Andrew Tanenbaum explains benefits of microkernel entertainingly
> in this talk:
> http://www.youtube.com/watch?v=bx3KuE7UjGA <http://www.youtube.com/watch?v=bx3KuE7UjGA>
>
> The talks has a focus on Minix but I think the benefits are transferable
> to other microkernels.
>
> # GNU Hurd
>
> ## Perceived focus
>
> I looks to me like there is a effort (which I appreciate) to get Guix
> working on Hurd.  I get this perception from:
>
> http://lists.gnu.org/archive/html/guix-devel/2016-12/msg00857.html <http://lists.gnu.org/archive/html/guix-devel/2016-12/msg00857.html>
>
> These comments from this thread:
>
> https://lists.gnu.org/archive/html/help-guix/2019-03/msg00158.html <https://lists.gnu.org/archive/html/help-guix/2019-03/msg00158.html>
>
> Ricardo Wurmus:  "Let’s work on the Hurd, people!  It’s beautiful!"
>
> Jan Nieuwenhuizen: "FWIW the Mes port to the Hurd is ongoing and mes now
> runs, next thing up is fork which we need for running mescc."
>
> ## Critiques of Hurd
>
> I would be curious what people think about these third party critiques
> (not mine) of Hurd.
>
> ### From X15
>
> https://www.sceen.net/x15 <https://www.sceen.net/x15/>
>
> "Although the design of the Hurd is promising and attractive, its
> implementation has a number of severe issues. X15 takes the approach of
> the complete rewrite to make sure that key ideas are kept in mind at all
> times during development. Since it’s not meant to be compatible with the
> Hurd, critical interfaces such as IPC and signals can be re-implemented
> completely differently. There is a lot of emphasis on code quality and
> ease of maintenance, obtained from disciplined application of best
> practices."
>
> ### From HelenOS
>
> http://www.helenos.org/wiki/FAQ#HowisHelenOSdifferentfromGNUHurd <http://www.helenos.org/wiki/FAQ#HowisHelenOSdifferentfromGNUHurd>
>
> ### Why Hurd?
>
> Why the focus on Hurd given other microkernel options?  I ask this
> question out of curiosity and a lack of practical experience with
> microkernels.
>
> # Microkernel wish list
>
> These are things that I see as desirable in a microkernel.
>
> ## Free software
>
> It should be completely free software.  No binary blobs included.  It
> looks like all of the microkernel listed here are:
> http://www.microkernel.info <http://www.microkernel.info/>
>
> ## RISC-V
>
> RISC-V a free and open instruction set architecture is a nice complement
> to a free operating system.  It is nice if a mircokernel already has
> plans to run on RISC-V. 
>
> Intel security issues:
> https://libreboot.org/faq.html#intel <https://libreboot.org/faq.html#intel>
>
> ARM security issues:
> https://libreboot.org/faq.html#amd <https://libreboot.org/faq.html#amd>
>
> ### Entirely free RISC-V computers
>
> These two initiatives are entirely free hardware based on RISC-V.
>
> * HiFive Unleashed
>  > https://www.sifive.com/boards/hifive-unleashed <https://www.sifive.com/boards/hifive-unleashed>
>
> * lowRISC
>  > https://www.lowrisc.org <https://www.lowrisc.org>
>
> ## Formal verification
>
> An application of the minimality principle in the design of microkernel
> leads to smaller code bases which are amenable to formal verification. 
>
> https://en.wikipedia.org/wiki/Microkernel#Essential_components_and_minimality <https://en.wikipedia.org/wiki/Microkernel#Essential_components_and_minimality>
>
> I see the extra security assurance that formal verification provide as
> desirable.
>
> # Alternative microkernels
>
> I used > http://www.microkernel.info <http://www.microkernel.info/>>  as the starting point when I began
> looking into microkernels. 
>
> ## Summary of interesting microkernels
>
> This is a high level summary based on the "Microkernel wish list" above.
> All of these are free software.  I am likely missing some other
> interesting microkernel projects.
>
> | projects     | RISC-V efforts | Formal verification |
> |--------------+----------------+---------------------|
> | sel4.systems | Yes            | Yes                 |
> | genode.org   | Yes            | Yes                 |
> | helenos.org  | Yes            | No                  |
> | muen.sk      | ?/No           | Yes                 |
> | minix3.org   | ?/No           | No                  |
> | hurd.gnu.org | ?/No           | No                  |
>
> Note:
>
> * ?/No is where (to me without asking) there does not look like there
>  have been efforts to make the project work with RISC-V. 
>
> * Genode is different than the others as it is not just a microkernel.
>  I have given Genode Yes for both RISC-V and Formal verification
>  because it can use the seL4 microkernel.  It can also use other
>  microkernels beyond just seL4.
>
> ## Other interesting projects
>
> robigalia.org:  based on seL4 microkernel which is formally verified and
> has RISC-V efforts underway.  It is using Rust to build the parts that
> would normally be part of a monolithic kernel in user space.  It looks
> like a young project.
>
> redox-os.org:  Rust based microkernel project.  It looks like a young
> project. 
>
> ## Projects I have not looked into
>
> I have not looked at the following projects which were also listed on
> http://www.microkernel.info <http://www.microkernel.info/>>  
>
> * github.com/Nils-TUD/Escape
> * github.com/f9micro
> * l4re.org
> * github.com/TUD-OS/M3
> * hypervisor.org
>


[-- Attachment #2: Type: text/html, Size: 16718 bytes --]

Re: Guix on a microkernel

[Pjotr Prins]

On Sat, Mar 30, 2019 at 08:05:40PM -0400, mikadoZero wrote:
> * I am curious what others think of microkernels.

Microkernels are of great interest from a security point of view. I
think they will become popular once there is a usable alternative. We
need free software, free software and proper isolation at the OS
level to improve security.

> ### Why Hurd?
> 
> Why the focus on Hurd given other microkernel options?  I ask this
> question out of curiosity and a lack of practical experience with
> microkernels.

Call it an accident of history ;). Being a GNU project we have a stake
in getting the Hurd to a usable level and get people to start using
it for daily work. 

I attended a talk at FOSDEM this year which gave an overview

https://fosdem.org/2019/schedule/event/roadmap_for_the_hurd/attachments/slides/3270/export/events/attachments/roadmap_for_the_hurd/slides/3270/2019_02_01_fosdem_roadmap.pdf

It looks like not a fat lot is needed to make this usable for many
people. Most free software will actually run on Hurd today.

Pj.

Re: Guix on a microkernel

[mikadoZero]

I do not have benchmark information.

Thank you for sharing the link.

znavko@tutanota.com writes:

> Thanks for your compilation.
> Do you have found actual benchmark tests? 
> https://www.gnu.org/software/hurd/faq/slow.html <https://www.gnu.org/software/hurd/faq/slow.html>
> "The Hurd is currently slower than Linux, yes. But not very much, so it is completely usable."
> Vulnerabilities of processors is also sensitive task. Maybe RISC-V will not have such bugs? Need to know in a practice.
>
> Mar 31, 2019, 12:05 AM by mikadozero@yandex.com:
>
>> # Appreciation
>>
>> I appreciate:
>>
>> * many of Guix's design decisions.  The one that is relevant to this
>>  discussion is the kernel.  I like that Guix uses the linux-libre (no
>>  binary blobs) instead of the linux kernel.
>>
>> * that work is underway to get Guix to work with GNU Hurd.  I like that
>>  a microkernel is a potential kernel option. 
>>
>>  > http://lists.gnu.org/archive/html/guix-devel/2016-12/msg00857.html <http://lists.gnu.org/archive/html/guix-devel/2016-12/msg00857.html>
>>
>> * the effort that has been put into GNU Hurd to get it to where it is.
>>
>> * the bootsrapping initiative.
>>
>>  > https://bootstrappable.org <https://bootstrappable.org/>
>>  > https://fosdem.org/2019/schedule/event/gnumes <https://fosdem.org/2019/schedule/event/gnumes/>
>>
>> # Intent
>>
>> * I would like to understand why GNU Hurd is being focused
>>  on (my perception) given other microkernel options. 
>> * I want to share what I have found after doing some looking into
>>  microkernels.
>> * I am curious what others think of microkernels.
>>
>> I mention the appreciations above because I am aiming for a tone of
>> appreciation and curiosity and not a critical one.  The tone can be a
>> challenge for written communication. 
>>
>> # My microkernel experience
>>
>> Currently I do not have any practical experience using any microkernel.
>> I have just spent time looking into the topic as it is interesting to
>> me.
>>
>> # Why microkernels?
>>
>> I think Andrew Tanenbaum explains benefits of microkernel entertainingly
>> in this talk:
>> http://www.youtube.com/watch?v=bx3KuE7UjGA <http://www.youtube.com/watch?v=bx3KuE7UjGA>
>>
>> The talks has a focus on Minix but I think the benefits are transferable
>> to other microkernels.
>>
>> # GNU Hurd
>>
>> ## Perceived focus
>>
>> I looks to me like there is a effort (which I appreciate) to get Guix
>> working on Hurd.  I get this perception from:
>>
>> http://lists.gnu.org/archive/html/guix-devel/2016-12/msg00857.html <http://lists.gnu.org/archive/html/guix-devel/2016-12/msg00857.html>
>>
>> These comments from this thread:
>>
>> https://lists.gnu.org/archive/html/help-guix/2019-03/msg00158.html <https://lists.gnu.org/archive/html/help-guix/2019-03/msg00158.html>
>>
>> Ricardo Wurmus:  "Let’s work on the Hurd, people!  It’s beautiful!"
>>
>> Jan Nieuwenhuizen: "FWIW the Mes port to the Hurd is ongoing and mes now
>> runs, next thing up is fork which we need for running mescc."
>>
>> ## Critiques of Hurd
>>
>> I would be curious what people think about these third party critiques
>> (not mine) of Hurd.
>>
>> ### From X15
>>
>> https://www.sceen.net/x15 <https://www.sceen.net/x15/>
>>
>> "Although the design of the Hurd is promising and attractive, its
>> implementation has a number of severe issues. X15 takes the approach of
>> the complete rewrite to make sure that key ideas are kept in mind at all
>> times during development. Since it’s not meant to be compatible with the
>> Hurd, critical interfaces such as IPC and signals can be re-implemented
>> completely differently. There is a lot of emphasis on code quality and
>> ease of maintenance, obtained from disciplined application of best
>> practices."
>>
>> ### From HelenOS
>>
>> http://www.helenos.org/wiki/FAQ#HowisHelenOSdifferentfromGNUHurd <http://www.helenos.org/wiki/FAQ#HowisHelenOSdifferentfromGNUHurd>
>>
>> ### Why Hurd?
>>
>> Why the focus on Hurd given other microkernel options?  I ask this
>> question out of curiosity and a lack of practical experience with
>> microkernels.
>>
>> # Microkernel wish list
>>
>> These are things that I see as desirable in a microkernel.
>>
>> ## Free software
>>
>> It should be completely free software.  No binary blobs included.  It
>> looks like all of the microkernel listed here are:
>> http://www.microkernel.info <http://www.microkernel.info/>
>>
>> ## RISC-V
>>
>> RISC-V a free and open instruction set architecture is a nice complement
>> to a free operating system.  It is nice if a mircokernel already has
>> plans to run on RISC-V. 
>>
>> Intel security issues:
>> https://libreboot.org/faq.html#intel <https://libreboot.org/faq.html#intel>
>>
>> ARM security issues:
>> https://libreboot.org/faq.html#amd <https://libreboot.org/faq.html#amd>
>>
>> ### Entirely free RISC-V computers
>>
>> These two initiatives are entirely free hardware based on RISC-V.
>>
>> * HiFive Unleashed
>>  > https://www.sifive.com/boards/hifive-unleashed <https://www.sifive.com/boards/hifive-unleashed>
>>
>> * lowRISC
>>  > https://www.lowrisc.org <https://www.lowrisc.org>
>>
>> ## Formal verification
>>
>> An application of the minimality principle in the design of microkernel
>> leads to smaller code bases which are amenable to formal verification. 
>>
>> https://en.wikipedia.org/wiki/Microkernel#Essential_components_and_minimality <https://en.wikipedia.org/wiki/Microkernel#Essential_components_and_minimality>
>>
>> I see the extra security assurance that formal verification provide as
>> desirable.
>>
>> # Alternative microkernels
>>
>> I used > http://www.microkernel.info <http://www.microkernel.info/>>  as the starting point when I began
>> looking into microkernels. 
>>
>> ## Summary of interesting microkernels
>>
>> This is a high level summary based on the "Microkernel wish list" above.
>> All of these are free software.  I am likely missing some other
>> interesting microkernel projects.
>>
>> | projects     | RISC-V efforts | Formal verification |
>> |--------------+----------------+---------------------|
>> | sel4.systems | Yes            | Yes                 |
>> | genode.org   | Yes            | Yes                 |
>> | helenos.org  | Yes            | No                  |
>> | muen.sk      | ?/No           | Yes                 |
>> | minix3.org   | ?/No           | No                  |
>> | hurd.gnu.org | ?/No           | No                  |
>>
>> Note:
>>
>> * ?/No is where (to me without asking) there does not look like there
>>  have been efforts to make the project work with RISC-V. 
>>
>> * Genode is different than the others as it is not just a microkernel.
>>  I have given Genode Yes for both RISC-V and Formal verification
>>  because it can use the seL4 microkernel.  It can also use other
>>  microkernels beyond just seL4.
>>
>> ## Other interesting projects
>>
>> robigalia.org:  based on seL4 microkernel which is formally verified and
>> has RISC-V efforts underway.  It is using Rust to build the parts that
>> would normally be part of a monolithic kernel in user space.  It looks
>> like a young project.
>>
>> redox-os.org:  Rust based microkernel project.  It looks like a young
>> project. 
>>
>> ## Projects I have not looked into
>>
>> I have not looked at the following projects which were also listed on
>> http://www.microkernel.info <http://www.microkernel.info/>>  
>>
>> * github.com/Nils-TUD/Escape
>> * github.com/f9micro
>> * l4re.org
>> * github.com/TUD-OS/M3
>> * hypervisor.org
>>

Re: Guix on a microkernel

[mikadoZero]

Pjotr Prins writes:

> ... Being a GNU project we have a stake
> in getting the Hurd to a usable level and get people to start using
> it for daily work. 

I do not understand this.  Could someone elaborate on this "stake"
idea further.

Does Guix give preferential treatment to GNU projects over other free
software projects?

> I attended a talk at FOSDEM this year which gave an overview
>
> https://fosdem.org/2019/schedule/event/roadmap_for_the_hurd/attachments/slides/3270/export/events/attachments/roadmap_for_the_hurd/slides/3270/2019_02_01_fosdem_roadmap.pdf
>
> It looks like not a fat lot is needed to make this usable for many
> people. Most free software will actually run on Hurd today.

Thanks for sharing this link.  I will watch that talk.

Re: Guix on a microkernel

[Ludovic Courtès]

Hi!

Pjotr Prins <pjotr.public12@thebird.nl> skribis:

> Call it an accident of history ;). Being a GNU project we have a stake
> in getting the Hurd to a usable level and get people to start using
> it for daily work. 

Maybe I shouldn’t write such a thing today, or maybe writing it today is
more fun because it lets people interpret it the way they want, but
anyway: there’s this idea of us Guix hackers getting to work on the Hurd
once Guix 1.0 is out!

Ludo’.

Re: Guix on a microkernel

[Joshua Branson]

Ludovic Courtès <ludo@gnu.org> writes:

> Hi!
>
> Pjotr Prins <pjotr.public12@thebird.nl> skribis:
>
>> Call it an accident of history ;). Being a GNU project we have a stake
>> in getting the Hurd to a usable level and get people to start using
>> it for daily work.
>
> Maybe I shouldn’t write such a thing today, or maybe writing it today is
> more fun because it lets people interpret it the way they want, but
> anyway: there’s this idea of us Guix hackers getting to work on the Hurd
> once Guix 1.0 is out!

That would be pretty cool!   The GNU/Hurd could use some more device
drivers, modern filesystem support, etc.

>
> Ludo’.
>

--
Joshua Branson
Sent from Emacs and Gnus

Re: Guix on a microkernel

[Vasilii Kolobkov]

[-- Attachment #1: Type: text/plain, Size: 313 bytes --]

My .2 cryptocoins, but you might find it interesting in evaluating HURD
as a kernerl for Guix system.  Frankly, I haven't been following HURD
development and don't know how much's changed since.

[[http://walfield.org/papers/200707-walfield-critique-of-the-GNU-Hurd.pdf]]
-- 
v, pgp key 0xFF3C7272972E83A5

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 832 bytes --]

Re: Guix on a microkernel

[Pronaip]

Not a microkernel, but Plan 9's file system semantics might be worth playing around with. The POSIX emulation probably wouldn't be able to support a full-blown Guix system, but the OS is very well worth taking hints from, at the very least. Overlays would imho be much nicer to work with than path variables.


Sent with ProtonMail Secure Email.

‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
On Sunday, April 28, 2019 9:54 PM, Vasilii Kolobkov <vasilii@orangeshoelaces.net> wrote:

> My .2 cryptocoins, but you might find it interesting in evaluating HURD
> as a kernerl for Guix system. Frankly, I haven't been following HURD
> development and don't know how much's changed since.
>
> [[http://walfield.org/papers/200707-walfield-critique-of-the-GNU-Hurd.pdf]]
>
> -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> v, pgp key 0xFF3C7272972E83A5

Re: Guix on a microkernel

[mikadoZero]

Thank you for sharing this.

Vasilii Kolobkov writes:

> My .2 cryptocoins, but you might find it interesting in evaluating HURD
> as a kernerl for Guix system.  Frankly, I haven't been following HURD
> development and don't know how much's changed since.
>
> [[http://walfield.org/papers/200707-walfield-critique-of-the-GNU-Hurd.pdf]]