Telemetry on by default kitty

Telemetry on by default kitty

[Bone Baboon]

Guix provides kitty a terminal emulator as a package.

Kitty has telemetry on by default.  See this issue on the kitty
repository for further information:
<https://github.com/kovidgoyal/kitty/pull/3544>

The issue was closed by the lead developer of the project without
addressing the concern raised.  It does not look like this is something
that is going to be fixed upstream.

The kitty telemetry is not a core part of kitty's functionality.  The
kitty lead developer said in that issue thread that the telemetry is to
notify users of available updates.  Further source code review would be
required to verify that is the only thing the telemetry is doing.  As
the lead developer did not provide much in the way of details when asked
about the telemetry in the issue thread.  It seems that the methods Guix
provides would be better suited for letting users know about updates. 

What do other people think about this in the context of the Free System
Distribution Guidelines?

```
No Malware
The distro must contain no DRM, no back doors, and no spyware.
```
<https://www.gnu.org/distros/free-system-distribution-guidelines.en.html>

How should the issue of kitty's telemetry on be default be addressed?

Re: Telemetry on by default kitty

[Tobias Geerinckx-Rice]

[-- Attachment #1: Type: text/plain, Size: 337 bytes --]

Bone Baboon 写道：
> What do other people think about this in the context of the Free 
> System
> Distribution Guidelines?

This is not a point of discussion.  Telemetry or ‘phoning home’ 
for updates must be opt-in if possible or disabled entirely 
otherwise.  Would you care to submit a patch?

Kind regards,

T G-R

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 247 bytes --]

Re: Telemetry on by default kitty

[Bone Baboon]

Tobias Geerinckx-Rice writes:

> Bone Baboon 写道：
>> What do other people think about this in the context of the Free
>> System
>> Distribution Guidelines?
>
> This is not a point of discussion.  Telemetry or ‘phoning home’ for
> updates must be opt-in if possible or disabled entirely otherwise.

The telemetry is not opt-in or disabled upstream.

> Would you care to submit a patch?

Should the patch be to remove the kitty package?

Re: Telemetry on by default kitty

[Tobias Geerinckx-Rice]

[-- Attachment #1: Type: text/plain, Size: 136 bytes --]

Hi Bone,

Bone Baboon 写道：
> Should the patch be to remove the kitty package?

No.  The telemetry.

Kind regards,

T G-R

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 247 bytes --]

Re: Telemetry on by default kitty

[Leo Prikler]

Hello everyone,

Am Samstag, den 12.06.2021, 23:44 +0200 schrieb Tobias Geerinckx-Rice:
> Hi Bone,
> 
> Bone Baboon 写道：
> > Should the patch be to remove the kitty package?
> 
> No.  The telemetry.
If I read terminals.scm, we already disable the telemetry in kitty:

> (invoke "python3" "setup.py" "linux-package"
>         ;; Do not phone home.
>         "--update-check-interval=0"

@Bone: Did you notice any other telemetry during your further code
review (or are you still in the process of reviewing the code)?  If
not, please try to cross-check Guix sources to see whether we already
disable telemetry, so as to not cause unwarranted panic :)

Regards,
Leo

Re: Telemetry on by default kitty

[Leo Prikler]

Am Sonntag, den 13.06.2021, 01:12 +0200 schrieb Leo Prikler:
> Hello everyone,
> 
> Am Samstag, den 12.06.2021, 23:44 +0200 schrieb Tobias Geerinckx-
> Rice:
> > Hi Bone,
> > 
> > Bone Baboon 写道：
> > > Should the patch be to remove the kitty package?
> > 
> > No.  The telemetry.
> If I read terminals.scm, we already disable the telemetry in kitty:
Oopsie-doopsie, I meant "If I read terminals.scm correctly".  Rest
still applies as-is.
> > (invoke "python3" "setup.py" "linux-package"
> >         ;; Do not phone home.
> >         "--update-check-interval=0"
> 
> @Bone: Did you notice any other telemetry during your further code
> review (or are you still in the process of reviewing the code)?  If
> not, please try to cross-check Guix sources to see whether we already
> disable telemetry, so as to not cause unwarranted panic :)
> 
> Regards,
> Leo

Re: Telemetry on by default kitty

[Mark H Weaver]

Hi Leo,

Leo Prikler <leo.prikler@student.tugraz.at> writes:

> If I read terminals.scm, we already disable the telemetry in kitty:
>
>> (invoke "python3" "setup.py" "linux-package"
>>         ;; Do not phone home.
>>         "--update-check-interval=0"

Indeed, it appears that Nicolas Goaziou addressed this issue last
December in commit 153b279e7baf1b23d9bb895ea2064faac0ed27c1.

> @Bone: Did you notice any other telemetry during your further code
> review (or are you still in the process of reviewing the code)?  If
> not, please try to cross-check Guix sources to see whether we already
> disable telemetry, so as to not cause unwarranted panic :)

I think that we should welcome reports like this one, even if the
reporter doesn't have the time/energy/skills to first investigate on
their own.  It would be unfortunate if we failed to find out about
spyware in Guix because those who knew about it had insufficient
confidence to file a bug report.

I, for one, appreciate that Bone Baboon brought this to our attention.

     Regards,
       Mark

-- 
Disinformation flourishes because many people care deeply about injustice
but very few check the facts.  Ask me about <https://stallmansupport.org>.

Re: Telemetry on by default kitty

[Bone Baboon]

Leo Prikler writes:
> Am Samstag, den 12.06.2021, 23:44 +0200 schrieb Tobias Geerinckx-Rice:
>> Bone Baboon 写道：
>> > Should the patch be to remove the kitty package?
>> 
>> No.  The telemetry.

While I appreciate the invitation to write a patch for kitty's telemetry
I am not going to write that patch.  My rational is explained below.

> If I read terminals.scm, we already disable the telemetry in kitty:
>
>> (invoke "python3" "setup.py" "linux-package"
>>         ;; Do not phone home.
>>         "--update-check-interval=0"
>
> @Bone: Did you notice any other telemetry during your further code
> review (or are you still in the process of reviewing the code)?  If
> not, please try to cross-check Guix sources to see whether we already
> disable telemetry, so as to not cause unwarranted panic :)

It appears problematic to patch kitty's telemetry for several reasons.

kitt's lead developer did not explaining kitty's telemetry when asked
for further information.  It is not clear if kitty performs other kinds
of telemetry as well.  It would seem that the only way to make sure that
all undesirable telemetry is removed would be a full review of kitty's
source code.

kitty's lead developer thinks that the telemetry in kitty is
acceptable.  So there would need to be an ongoing review of commits to
the kitty source code to ensure that undesirable telemetry is not added
in the future.

I have never been a kitty user and seeing the position kitty's lead
developer holds on telemetry I will not be a kitty user in the future.
I will not be reviewing the kitty source code or future commits to the
kitty source code.

Re: Telemetry on by default kitty

[Leo Prikler]

Am Samstag, den 12.06.2021, 22:03 -0400 schrieb Bone Baboon:
> Leo Prikler writes:
> > Am Samstag, den 12.06.2021, 23:44 +0200 schrieb Tobias Geerinckx-
> > Rice:
> > > Bone Baboon 写道：
> > > > Should the patch be to remove the kitty package?
> > > 
> > > No.  The telemetry.
> 
> While I appreciate the invitation to write a patch for kitty's
> telemetry
> I am not going to write that patch.  My rational is explained below.
> 
> > If I read terminals.scm, we already disable the telemetry in kitty:
> > 
> > > (invoke "python3" "setup.py" "linux-package"
> > >         ;; Do not phone home.
> > >         "--update-check-interval=0"
> > 
> > @Bone: Did you notice any other telemetry during your further code
> > review (or are you still in the process of reviewing the code)?  If
> > not, please try to cross-check Guix sources to see whether we
> > already
> > disable telemetry, so as to not cause unwarranted panic :)
> 
> It appears problematic to patch kitty's telemetry for several
> reasons.
> 
> kitt's lead developer did not explaining kitty's telemetry when asked
> for further information.  It is not clear if kitty performs other
> kinds
> of telemetry as well.  It would seem that the only way to make sure
> that
> all undesirable telemetry is removed would be a full review of
> kitty's
> source code.
> 
> kitty's lead developer thinks that the telemetry in kitty is
> acceptable.  So there would need to be an ongoing review of commits
> to
> the kitty source code to ensure that undesirable telemetry is not
> added
> in the future.
> 
> I have never been a kitty user and seeing the position kitty's lead
> developer holds on telemetry I will not be a kitty user in the
> future.
> I will not be reviewing the kitty source code or future commits to
> the
> kitty source code.
Are we reading the same issue?  As I understand it, there is the
ethical problem of kitty's telemetry being *enabled* by default rather
than *disabled*, but both parties seem to agree on how to disable it
(completely).  Of course, there's the added bonus of the lead developer
expressing their views in a… rather aggressive tone to put it mildly,
but that's a social problem.

As far as I can see, neither of those pose any technical problem to
disabling telemetry in Guix, which as far as I can see, we do.  It is
important to raise social and ethical issues, but I'm not sure whether
guix-devel is the right platform for that (in that change will be
achieved through discussion here).

Regards,
Leo

Re: Telemetry on by default kitty

[Tobias Geerinckx-Rice]

[-- Attachment #1: Type: text/plain, Size: 466 bytes --]

Mark, Bone,

Mark H Weaver 写道：
> I, for one, appreciate that Bone Baboon brought this to our 
> attention.

So do I, but I would have appreciated it even more if the subject 
and lede had been accurate:

> Guix provides kitty a terminal emulator as a package.
> Kitty has telemetry on by default.

is, at best, misleading.  This is a Guix mailing list, not 
Kitty's, so whatever upstream defaults to is not our concern.

Kind regards,

T G-R

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 247 bytes --]

Re: Telemetry on by default kitty

[Leo Famulari]

On Sun, Jun 13, 2021 at 11:32:01AM +0200, Leo Prikler wrote:
> Of course, there's the added bonus of the lead developer
> expressing their views in a… rather aggressive tone to put it mildly,
> but that's a social problem.

To be fair to the author, the bug report described a simple update
notifier as "surveillance" and said that Goyal is contributing "to the
normalization of surveillance in GNU/Linux software."

Those are opinions, not facts.

The bug reporter's social skills were also quite poor; you don't begin
to ask for something by using pejorative terms and accusing people of
doing something nefarious. It is insulting.

Because Kitty has an automatic update notifier built-in, it's obvious
that Kitty is intended to be distributed outside of the old-school
GNU/Linux distro model, and that means that automatic updates are
expected and valuable. Distros like Guix are the entities who are doing
something weird in this scenario.

Not to mention that telemetry about crashes and usage are extremely
valuable for developers. The free software community's resistance to
that kind of automated feedback is an important factor in the overall
low quality of free software relative to other non-free operating
systems, in my opinion. There are many reasons that "the masses" choose
not to use GNU/Linux despite extremely widespread popular
dissatisfaction with "big tech", and low quality and poor integration is
a big one, in my experience.

Re: Telemetry on by default kitty

[Leo Prikler]

Am Sonntag, den 13.06.2021, 13:57 -0400 schrieb Leo Famulari:
> On Sun, Jun 13, 2021 at 11:32:01AM +0200, Leo Prikler wrote:
> > Of course, there's the added bonus of the lead developer
> > expressing their views in a… rather aggressive tone to put it
> > mildly,
> > but that's a social problem.
> 
> To be fair to the author, the bug report described a simple update
> notifier as "surveillance" and said that Goyal is contributing "to
> the
> normalization of surveillance in GNU/Linux software."
> 
> Those are opinions, not facts.
Point taken, that too is a social problem.  The fact, that the patch in
question was like one line, does not help either.

> The bug reporter's social skills were also quite poor; you don't
> begin
> to ask for something by using pejorative terms and accusing people of
> doing something nefarious. It is insulting.
I read that in a somewhat different way, but I do agree with your
statement.

> Because Kitty has an automatic update notifier built-in, it's obvious
> that Kitty is intended to be distributed outside of the old-school
> GNU/Linux distro model, and that means that automatic updates are
> expected and valuable. Distros like Guix are the entities who are
> doing something weird in this scenario.
I see the same. Much software has built its own update channels,
because the ones supported by mainstream distros typically don't cut
it.  It's even worse if you want to supply other platforms as well. 
Even Emacs has its own package repo :P

> Not to mention that telemetry about crashes and usage are extremely
> valuable for developers. The free software community's resistance to
> that kind of automated feedback is an important factor in the overall
> low quality of free software relative to other non-free operating
> systems, in my opinion. There are many reasons that "the masses"
> choose
> not to use GNU/Linux despite extremely widespread popular
> dissatisfaction with "big tech", and low quality and poor integration
> is
> a big one, in my experience.
Perhaps it's valuable for developers, but as a user I often have next
to no information about what data gets collected and for which purpose,
both of which are important for *informed consent*.  If "the masses"
don't really care about the data being collected and would rather see
improvements on their software, they are free to enable telemetry –
that's what opt-in is for – but my personal opinion on this is that
you're going to have a hard time convincing people, that you actually
only collect reasonable amounts and use them with respect for privacy
rights.

Regards,
Leo

Re: Telemetry on by default kitty

[Leo Famulari]

On Sun, Jun 13, 2021 at 08:35:18PM +0200, Leo Prikler wrote:
> Perhaps it's valuable for developers, but as a user I often have next
> to no information about what data gets collected and for which purpose,
> both of which are important for *informed consent*.  If "the masses"
> don't really care about the data being collected and would rather see
> improvements on their software, they are free to enable telemetry –
> that's what opt-in is for – but my personal opinion on this is that
> you're going to have a hard time convincing people, that you actually
> only collect reasonable amounts and use them with respect for privacy
> rights.

Yeah, I agree that telemetry is a problem in addition to being valuable
for developers.

I think that making it opt-in doesn't really help very much. People use
defaults. I read that Firefox struggles with software quality on
GNU/Linux because almost nobody enables the telemetry.

I feel that, ultimately, we already trust most software authors
implicitly and totally, because we are not auditing their programs. So,
I am personally happy to enable the telemetry for most software I use —
especially if it is free software and especially for software that deals
with the network. I don't personally see the point of treating telemetry
as a special case in terms of trust or consent.

Re: Telemetry on by default kitty

[Ryan Prior]

On Sunday, June 13th, 2021 at 7:04 PM, Leo Famulari <leo@famulari.name> wrote:
> Yeah, I agree that telemetry is a problem in addition to being valuable
>
> for developers.

I've been encouraged by the recent progress in differential privacy that opt-out freedom respecting telemetry may be possible. I think this needs to be discussed and coordinated on a platform level and offered as a standard feature for packages on the platform, rather than something app developers have to implement on their own in each instance.

An easy API to discover whether you have any packages that could be upgraded would be very handy as well. Currently I find this information by running `guix time-machine --branch=master -- package -u --dry-run --no-grafts` but this is imperfect, takes a long time to run, and doesn't inform me whether there's upstream package updates that aren't packaged in Guix yet but that I could use with a package transform.

Re: Telemetry on by default kitty

[Leo Prikler]

Am Sonntag, den 13.06.2021, 23:54 +0000 schrieb Ryan Prior:
> An easy API to discover whether you have any packages that could be
> upgraded would be very handy as well. Currently I find this
> information by running `guix time-machine --branch=master -- package
> -u --dry-run --no-grafts` but this is imperfect, takes a long time to
> run, and doesn't inform me whether there's upstream package updates
> that aren't packaged in Guix yet but that I could use with a package
> transform.
`guix refresh PACKAGE` mostly works in this context, though obviously
its use is geared more towards development.

Re: Telemetry on by default kitty

[Ludovic Courtès]

Hi,

Leo Prikler <leo.prikler@student.tugraz.at> skribis:

> Am Sonntag, den 13.06.2021, 23:54 +0000 schrieb Ryan Prior:
>> An easy API to discover whether you have any packages that could be
>> upgraded would be very handy as well. Currently I find this
>> information by running `guix time-machine --branch=master -- package
>> -u --dry-run --no-grafts` but this is imperfect, takes a long time to
>> run, and doesn't inform me whether there's upstream package updates
>> that aren't packaged in Guix yet but that I could use with a package
>> transform.
> `guix refresh PACKAGE` mostly works in this context, though obviously
> its use is geared more towards development.

Don’t miss ‘--with-latest’!

  https://guix.gnu.org/manual/en/html_node/Package-Transformation-Options.html

Ludo’.

Re: Telemetry on by default kitty

[Giovanni Biscuolo]

[-- Attachment #1: Type: text/plain, Size: 5036 bytes --]

Hi Leo and Guix,

sorry for this long message but I would like to add my point of view to
the discussion about telemetry.

I apreciated the laconic statement by Tobias Geerinckx-Rice on Sat, 12
Jun 2021 22:35:40 +0200 [1]:

--8<---------------cut here---------------start------------->8---

 This is not a point of discussion.  Telemetry or ‘phoning home’ 
 for updates must be opt-in if possible or disabled entirely 
 otherwise.  Would you care to submit a patch?

--8<---------------cut here---------------end--------------->8---

AFAIU there is a general consensus above all GNU Guix maintainers (and
all FSDG compliant distros) on the above statement: am I wrong?

I'm using Guix (and other ditributions) primarily for this very reason,
for me this is the most important *feature* of a free software
distribution: no spyware ALSO means no opt-out telemetry.

To be clear: if Guix "only" had the fantastic features it has but was
not FSDG compliant, I'd use something else (and be very very sad).

Leo Famulari <leo@famulari.name> writes:

> On Sun, Jun 13, 2021 at 08:35:18PM +0200, Leo Prikler wrote:
>> Perhaps it's valuable for developers, but as a user I often have next
>> to no information about what data gets collected and for which purpose,
>> both of which are important for *informed consent*.

[...]

> Yeah, I agree that telemetry is a problem in addition to being valuable
> for developers.

No, telemetry is not just "a problem", it's A HUGE legal issue.

I don't want to have a too long privacy related discussion here, but
please consider in EU (I live in Italy) we have the GDPR [2] and we had
a LOT of issues with the "Privacy Shield", now invalidated by the
Schrems II [3] EU Court of Justice judgement, meaning that data
transfers abroad are... VERY problematic :-D

Just to give you one recent example, in Italy we have a public service
app called "IO App" (processing a lot of very sensitive data) that was
recently surveied by the italian Privacy Authority and it was a
*disaster* [4]:

--8<---------------cut here---------------start------------->8---

 the Authority, on general criticisms on the functioning of the IO App,
 has ordered, with a urgent measure, to PagoPA to temporally block the
 personal data processing by this App which require the interaction with
 Google’s services and Mixpanel, and which involve a transference to
 third countries (for example: USA, India, Australia) of personal
 sensitive data (like: cash back transactions, payments instruments,
 holydays bonus), carried out without the consent of the users.

--8<---------------cut here---------------end--------------->8---

So, the italian goverment is (still) tranfering a lot of personal data
to NOT (equivalent) GDPR compliant nations.

Please consider that much, if not all, of the personal data transferred
(and it's LOT of data) was allegedly for "telemetry" and "issue
tracking" purposes.

We are talking about this.  This is not for sure a kitty issue, but it
is a telemetry issue.

> I think that making it opt-in doesn't really help very much. People use
> defaults. I read that Firefox struggles with software quality on
> GNU/Linux because almost nobody enables the telemetry.

This is freedom n. 0 :-D

> I feel that, ultimately, we already trust most software authors
> implicitly and totally, because we are not auditing their
> programs. So, I am personally happy to enable the telemetry for most
> software I use — especially if it is free software and especially for
> software that deals with the network. I don't personally see the point
> of treating telemetry as a special case in terms of trust or consent.

I'm sorry you don't see the point, but please remember that in some
countries providing personal data to data processors needs informed
consent on what, why, by whom and where the data is processed (please
consider this as an executive-summary, it's a complex matter).

Please also consider I'm not willing to provide data to the developers
of software I use simply because I don't want to exchange data for the
permission to use the software... and I'm not the only one: this is the
most important reason telemetry must be disabled by default (opt-in) if
possible or completely disabled otherwhise.

Privacy is valuable, developers must respect their users.

Thank you! Giovanni.


[1] Message-Id:87eed695yb.fsf@nckx

[2] https://en.wikipedia.org/wiki/GDPR

[3] https://en.wikipedia.org/wiki/Max_Schrems#Schrems_II

[4] https://www.privacy365.eu/en/by-the-italian-data-protection-authority-green-certification-the-green-light-of-the-authority-but-with-specific-guarantees-it-has-been-disposed-the-block-of-io-app/

https://www.privacy365.eu/en/by-the-italian-data-protection-authority-app-io-the-authority-implements-the-technical-relation/
(unfortunately the relation is in italian only, it's very very interesting!)

-- 
Giovanni Biscuolo

Xelera IT Infrastructures

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 849 bytes --]

Re: Telemetry on by default kitty

[Leo Prikler]

Am Dienstag, den 15.06.2021, 19:24 +0200 schrieb Giovanni Biscuolo:
> Hi Leo and Guix,
Wrong Leo here, I hope you don't mind me responding.

> sorry for this long message but I would like to add my point of view
> to
> the discussion about telemetry.
> 
> I apreciated the laconic statement by Tobias Geerinckx-Rice on Sat,
> 12
> Jun 2021 22:35:40 +0200 [1]:
> 
> --8<---------------cut here---------------start------------->8---
> 
>  This is not a point of discussion.  Telemetry or ‘phoning home’ 
>  for updates must be opt-in if possible or disabled entirely 
>  otherwise.  Would you care to submit a patch?
> 
> --8<---------------cut here---------------end--------------->8---
> 
> AFAIU there is a general consensus above all GNU Guix maintainers
> (and all FSDG compliant distros) on the above statement: am I wrong?
This depends on how you interpret "The distro must contain no DRM, no
back doors, and no spyware.".  The consensus (at least among Guix and
also when we consider EU law) is that you need to opt in to any
collection of data.

> I'm using Guix (and other distributions) primarily for this very
> reason, for me this is the most important *feature* of a free
> software distribution: no spyware ALSO means no opt-out telemetry.
> 
> To be clear: if Guix "only" had the fantastic features it has but was
> not FSDG compliant, I'd use something else (and be very very sad).
> 
> Leo Famulari <leo@famulari.name> writes:
> 
> > On Sun, Jun 13, 2021 at 08:35:18PM +0200, Leo Prikler wrote:
> > > Perhaps it's valuable for developers, but as a user I often have
> > > next
> > > to no information about what data gets collected and for which
> > > purpose,
> > > both of which are important for *informed consent*.
> 
> [...]
> 
> > Yeah, I agree that telemetry is a problem in addition to being
> > valuable
> > for developers.
> 
> No, telemetry is not just "a problem", it's A HUGE legal issue.
> 
> I don't want to have a too long privacy related discussion here, but
> please consider in EU (I live in Italy) we have the GDPR [2] and we
> had
> a LOT of issues with the "Privacy Shield", now invalidated by the
> Schrems II [3] EU Court of Justice judgement, meaning that data
> transfers abroad are... VERY problematic :-D
Legally speaking, you might be able to claim legitimate interest
according to §6.1.f for your telemetry (I really hate that one).  It'd
be interesting to see what happens in court if you do, but it's out
there.  As a EU citizen myself, I really wish the GDPR was stricter in
statements.

> Just to give you one recent example, in Italy we have a public
> service
> app called "IO App" (processing a lot of very sensitive data) that
> was
> recently surveied by the italian Privacy Authority and it was a
> *disaster* [4]:
> 
> --8<---------------cut here---------------start------------->8---
> 
>  the Authority, on general criticisms on the functioning of the IO
> App,
>  has ordered, with a urgent measure, to PagoPA to temporally block
> the
>  personal data processing by this App which require the interaction
> with
>  Google’s services and Mixpanel, and which involve a transference to
>  third countries (for example: USA, India, Australia) of personal
>  sensitive data (like: cash back transactions, payments instruments,
>  holydays bonus), carried out without the consent of the users.
> 
> --8<---------------cut here---------------end--------------->8---
> 
> So, the italian goverment is (still) tranfering a lot of personal
> data
> to NOT (equivalent) GDPR compliant nations.
> 
> Please consider that much, if not all, of the personal data
> transferred
> (and it's LOT of data) was allegedly for "telemetry" and "issue
> tracking" purposes.
Which is exactly why I said what I said about informed consent.

> We are talking about this.  This is not for sure a kitty issue, but
> it
> is a telemetry issue.
> 
> > I think that making it opt-in doesn't really help very much. People
> > use
> > defaults. I read that Firefox struggles with software quality on
> > GNU/Linux because almost nobody enables the telemetry.
> 
> This is freedom n. 0 :-D
You could equivalently say, that freedom 0 is guaranteed through an
opt-out mechanism.  Opt-in vs. opt-out is a different ethical
conundrum, I fear.

> > I feel that, ultimately, we already trust most software authors
> > implicitly and totally, because we are not auditing their
> > programs. So, I am personally happy to enable the telemetry for
> > most
> > software I use — especially if it is free software and especially
> > for
> > software that deals with the network. I don't personally see the
> > point
> > of treating telemetry as a special case in terms of trust or
> > consent.
> 
> I'm sorry you don't see the point, but 
Might be just me, but this phrasing appears a little aggressive given
the overall tone of the message being… a little less so.

> Please remember that in some countries providing personal data to
> data processors needs informed consent on what, why, by whom and
> where the data is processed (please consider this as an executive-
> summary, it's a complex matter).
> 
> Please also consider I'm not willing to provide data to the
> developers of software I use simply because I don't want to exchange
> data for the permission to use the software... and I'm not the only
> one: this is the most important reason telemetry must be disabled by
> default (opt-in) if possible or completely disabled otherwhise.
> 
> Privacy is valuable, developers must respect their users.
Agree completely.

> [1] Message-Id:87eed695yb.fsf@nckx
> 
> [2] https://en.wikipedia.org/wiki/GDPR
> 
> [3] https://en.wikipedia.org/wiki/Max_Schrems#Schrems_II
> 
> [4] 
> https://www.privacy365.eu/en/by-the-italian-data-protection-authority-green-certification-the-green-light-of-the-authority-but-with-specific-guarantees-it-has-been-disposed-the-block-of-io-app/
> 
> https://www.privacy365.eu/en/by-the-italian-data-protection-authority-app-io-the-authority-implements-the-technical-relation/
> (unfortunately the relation is in italian only, it's very very
> interesting!)
The green certification is another ethic conundrum, not only on the
basis of it being a data protection nightmare.  But alas, politicians
will take any reason they can get to weaken restrictions so that "the
economy may prosper".

Re: Telemetry on by default kitty

[Mark H Weaver]

Hi Leo,

Leo Famulari <leo@famulari.name> writes:
> I feel that, ultimately, we already trust most software authors
> implicitly and totally, because we are not auditing their programs.

Agreed.

> So, I am personally happy to enable the telemetry for most software I
> use — especially if it is free software and especially for software
> that deals with the network.

That's your personal decision, and I agree that telemetry functionality
should be permissible in Guix, as long as it's opt-in.

> I don't personally see the point of treating telemetry as a special
> case in terms of trust or consent.

One problem is that telemetry involves trusting more than just the
developer.  Telemetry also reveals information to the user's internet
service provider, the network operators between user and the server, the
company that controls the hardware that the server runs on, and any
intelligence agencies or other hostile actors that have infiltrated
those networks or servers.  Moreover, if the server keeps logs,
governments may coerce the developer into surrendering those logs.

Therefore, when a program generates unsolicited and unexpected network
traffic -- and I certainly do *not* expect a terminal program to
generate network traffic -- it is effectively leaking some of your
private information to all of those other actors.  That, in itself, is
arguably a breach of trust, regardless of the developer's presumably
good intentions.

I understand that many people have given up on protecting their privacy,
or simply don't care.  Kitty's developer seems to be of that mindset.

However, I strongly believe that each Guix user should be given the
opportunity to make that decision for themselves, i.e. that telemetry,
auto-update checks, and more generally unsolicited network traffic
should be disabled until the user has given informed consent.

What do other people think?

      Regards,
        Mark

-- 
Disinformation flourishes because many people care deeply about injustice
but very few check the facts.  Ask me about <https://stallmansupport.org>.

Re: Telemetry on by default kitty

[Jack Hill]

[-- Attachment #1: Type: text/plain, Size: 2357 bytes --]

On Tue, 15 Jun 2021, Mark H Weaver wrote:

[…]

> However, I strongly believe that each Guix user should be given the
> opportunity to make that decision for themselves, i.e. that telemetry,
> auto-update checks, and more generally unsolicited network traffic
> should be disabled until the user has given informed consent.
>
> What do other people think?

I'm not sure I have too much to add to the discussion, but since I once 
submitted a patch to disable this type of telemetry⁰, I support the notion 
that programs should not generate network traffic unless they are asked to 
do so. As Mark says, it's more than just the two endpoints that can 
observe the traffic. Even encrypted traffic provides some information.

Perhaps opting-in can be another use case for parameterized packages. We 
could have our cake and still allow folks to opt-in without having to 
tediously configure or modify their packages.

On the note of trusting software authors, for me a lot of it is 
understanding the development process and analyzing if my interests are 
aligned with those the authors. However, that can be a complicated thing. 
In general, I'm much more trusting of community projects than ones with 
corporate sponsors. Track record also counts too, so I'm glad that Bone 
referred us to the upstream discussion. I'll probably spend more of my 
time looking for problems in future releases of projects like kitty and 
audacity¹ than more trusted (to me) projects like goffice.

Even if we're not able to catch everything, auditing source can still be 
useful. I found an information leak in innernet (not packaged for Guix 
yet) in part because the authors where kind enough to point it out in a 
comment². Perhaps auditing/patching is a test that is well suited to 
combining efforts with folks beyond Guix. That can be either in dedicated 
projects like Icecat or ungoogled-chromium, or simply by looking at what 
patches and configuration options other package distributions apply. Of 
course we can also share anything that we learn.

⁰ https://issues.guix.gnu.org/40360
¹ https://www.theregister.com/2021/05/14/audacity_telemetry/
² https://github.com/tonarino/innernet/blob/46d97831094d04fe3ad802a4bf2ac645e09d568c/publicip/src/lib.rs#L3-L4

Well, I guess I ended up adding more comments than I thought I would. Hope 
they're helpful!

Jack

Re: Telemetry on by default kitty

[Leo Famulari]

On Tue, Jun 15, 2021 at 11:39:59PM +0200, Leo Prikler wrote:
> Am Dienstag, den 15.06.2021, 19:24 +0200 schrieb Giovanni Biscuolo:
> > I'm sorry you don't see the point, but 

Good grief...

> Might be just me, but this phrasing appears a little aggressive given
> the overall tone of the message being… a little less so.

Yeah, indeed.

Giovanni, please consider that I might "see the point" but I instead
reached a different conclusion, because I have different values and
preferences than you.

Re: Telemetry on by default kitty

[Mark H Weaver]

Hi Leo,

Leo Famulari <leo@famulari.name> writes:

> On Tue, Jun 15, 2021 at 11:39:59PM +0200, Leo Prikler wrote:
>> Am Dienstag, den 15.06.2021, 19:24 +0200 schrieb Giovanni Biscuolo:
>> > I'm sorry you don't see the point, but 
>
> Good grief...
>
>> Might be just me, but this phrasing appears a little aggressive given
>> the overall tone of the message being… a little less so.
>
> Yeah, indeed.
>
> Giovanni, please consider that I might "see the point" but I instead
> reached a different conclusion, because I have different values and
> preferences than you.

I guess you feel offended that Giovanni implied that you don't "see the
point".  There's just one problem: those were the _exact_ words that
_you_ used to describe yourself.

Leo Famulari <leo@famulari.name> wrote:
> I don't personally see the point
> of treating telemetry as a special case in terms of trust or consent.

Giovanni Biscuolo <g@xelera.eu> replied (after quoting the words above):
> I'm sorry you don't see the point, but please remember [...]

and now you respond with indignation: "please consider that I might
'see the point'".

You are contradicting yourself.  First you say that you don't "see the
point", then you are offended that Giovanni implies that you don't "see
the point" and ask him to consider that you _might_ "see the point".
You can't have it both ways.

In my opinion, here you are being far more aggressive in your tone
policing than Giovanni has ever been on this list.

       Mark

--
Disinformation flourishes because many people care deeply about injustice
but very few check the facts.  Ask me about <https://stallmansupport.org>.

my apoligies (was Re: Telemetry on by default kitty)

[Giovanni Biscuolo]

[-- Attachment #1: Type: text/plain, Size: 1064 bytes --]

Dear Leo F. and Leo P.

Leo Famulari <leo@famulari.name> writes:

> On Tue, Jun 15, 2021 at 11:39:59PM +0200, Leo Prikler wrote:
>> Am Dienstag, den 15.06.2021, 19:24 +0200 schrieb Giovanni Biscuolo:
>> > I'm sorry you don't see the point, but 
>
> Good grief...
>
>> Might be just me, but this phrasing appears a little aggressive given
>> the overall tone of the message being… a little less so.
>
> Yeah, indeed.

In my sentence:

>>> I'm sorry you don't see the point, but please remember that [...]

I was just repeating a few words from this sentence by Leo F.:

>>>> [...] I don't personally see the point of treating telemetry as a
>>>> special case in terms of trust or consent.

Please can you (both) explain me what's aggressive in my above phrasing?

I'm not a native speaker and I'm pretty sure my english needs much
improvement, so I'll appreciate I you explain me what I did wrong and
why you are calling my sentence "grief".

[...]

All apologies, Giovanni.

-- 
Giovanni Biscuolo

Xelera IT Infrastructures

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 849 bytes --]

Re: my apoligies (was Re: Telemetry on by default kitty)

[Leo Prikler]

Hello Giovanni,

Am Mittwoch, den 16.06.2021, 19:32 +0200 schrieb Giovanni Biscuolo:
> Dear Leo F. and Leo P.
> 
> Leo Famulari <leo@famulari.name> writes:
> 
> > On Tue, Jun 15, 2021 at 11:39:59PM +0200, Leo Prikler wrote:
> > > Am Dienstag, den 15.06.2021, 19:24 +0200 schrieb Giovanni
> > > Biscuolo:
> > > > I'm sorry you don't see the point, but 
> > 
> > Good grief...
> > 
> > > Might be just me, but this phrasing appears a little aggressive
> > > given the overall tone of the message being… a little less so.
> > 
> > Yeah, indeed.
> 
> In my sentence:
> 
> > > > I'm sorry you don't see the point, but please remember that
> > > > [...]
> 
> I was just repeating a few words from this sentence by Leo F.:
> 
> > > > > [...] I don't personally see the point of treating telemetry
> > > > > as a
> > > > > special case in terms of trust or consent.
> 
> Please can you (both) explain me what's aggressive in my above
> phrasing?
I'm not going to argue as Leo F. did, that they in fact did see the
point when they claimed differently.  If that discussion has any value,
it should be done under Mark's reply.

I simply want to draw your attention to the phrasing "I'm sorry you
VERB/ACTION".  This phrasing can be used to deliberately provoke, but
it's more commonly used when
- people are forced to apologize out of social pressure without
actually understanding what they are apologizing for
- people generally have troubles expressing themselves in English.
This is something, that English native speakers grasp intuitively, but
can be a bit alien towards non-native speakers.

Consider the following dialogue:

Bob: *says something misogynistic*
Alice: I feel offended by that.
Bob: I'm sorry you feel offended.

Here, Bob is not apologizing for the fact, that he hurt Alice through
his misogynistic statement, but rather for the fact, that Alice felt
hurt.  It is a subtle difference, that shows how Bob does not see the
error of his way.  Note, that I'm using male-on-female verbal
aggression here as an example, but it's by far not the only one.

In your case, this general negative connotation of the phrasing "I'm
sorry you X" is amplified by you having stated it when joining the
conversation, i.e. without even having done anything any of us could
feel offended by.  This proactive use of "I'm sorry you X" can be
argued to be even more aggressive.

Imagine instead of typing out a detailed reply explaining why people
read your message as somehow containing anger, I said something along
the lines of "I'm sorry your English is not as good as mine". Not only
would this make me appear extremely pompous to any third party
observing, but more importantly you as my conversation partner would be
attacked for your lack in skill, empathy, …, pick something you like. 
This is somewhat comparable to how people will misread your reply when
condensed to that one line.

Going back to the full quote:
> > > Might be just me, but this phrasing appears a little aggressive
> > > given the overall tone of the message being… a little less so.
I singled out this very phrase, because it appears to be in stark
contrast to the rest of your message.  I also voiced my agreement with
the statement that followed, so it's really just a minor complaint
about this particular part that makes up a minor portion of your reply.

Do not under any circumstances take this as somehow invalidating your
whole point. It doesn't. It's just a tiny case of bad optics.

> I'm not a native speaker and I'm pretty sure my english needs much
> improvement, so I'll appreciate I you explain me what I did wrong and
> why you are calling my sentence "grief".
Leo F. is not labelling your sentence grief, it is rather an expression
of dismay.  It may alternatively (in different contexts) also be
surprise, shock or disgust and probably other things Wiktionary doesn't
tell you about.  There is no direct Italian equivalent here – the
translation depends on context.

You may take all of the above with a grain of salt, as I too am not a
native English speaker.  In some sense, I am projecting a German view
of language (as that's my native tongue) onto English, as the same
applies here to "Tut mir leid, dass du …", but I believe those two
languages are close enough for that inference to be made.  There might
perhaps also be cultural differences in English as a first language
within the UK, the US, Canada and Australia, not to mention English as
a second language in various parts of the world, including Italy.  In
any case, there are contexts in which the thing you've written appears
rather aggressive and you might want to learn how to pick up such
nuances.

I hope this clears up the point I was trying to make.
Regards, Leo P.

Re: my apoligies (was Re: Telemetry on by default kitty)

[Leo Famulari]

On Wed, Jun 16, 2021 at 07:32:34PM +0200, Giovanni Biscuolo wrote:
> Please can you (both) explain me what's aggressive in my above phrasing?
> 
> I'm not a native speaker and I'm pretty sure my english needs much
> improvement, so I'll appreciate I you explain me what I did wrong and
> why you are calling my sentence "grief".

I'm sorry for my role in making this conversation discordant.

The phrases "I don't see the point of ..." and "good grief" are
colloquial English idioms that may be difficult to translate [0].

I'll try to speak more precisely in the future, because these idioms
make it harder for people who speak other languages to communicate
within the project.

[0] The first phrase is a way to say that I think the position being
described is incoherent. The second phrase is an exclamation of negative
emotions.

Re: Telemetry on by default kitty

[Ludovic Courtès]

Hi,

Giovanni Biscuolo <g@xelera.eu> skribis:

> I apreciated the laconic statement by Tobias Geerinckx-Rice on Sat, 12
> Jun 2021 22:35:40 +0200 [1]:
>
>
>  This is not a point of discussion.  Telemetry or ‘phoning home’ 
>  for updates must be opt-in if possible or disabled entirely 
>  otherwise.  Would you care to submit a patch?
>
>
> AFAIU there is a general consensus above all GNU Guix maintainers (and
> all FSDG compliant distros) on the above statement: am I wrong?

I think you’re right; I think phoning home should be either disabled
opt-in.

As Leo Prikler wrote, it’s arguably a legal obligation in some
jurisdictions.

Ludo’.

Re: Telemetry on by default kitty

[Bone Baboon]

Leo Prikler writes:
> If I read terminals.scm, we already disable the telemetry in kitty:
>
>> (invoke "python3" "setup.py" "linux-package"
>>         ;; Do not phone home.
>>         "--update-check-interval=0"

> @Bone: Did you notice any other telemetry during your further code
> review (or are you still in the process of reviewing the code)?  If
> not, please try to cross-check Guix sources to see whether we already
> disable telemetry, so as to not cause unwarranted panic :)

I have followed up with kitty's use of telemetry by opening an issue on
the kitty repository.

https://github.com/kovidgoyal/kitty/issues/3802

I suggested that all the uses of telemetry be documented and that there
be a configuration to disable all telemetry.

My interpretation of the response is that the update checking telemetry
is the only telemetry and that the way to disable it is how it is being
done in terminals.scm.