Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
How does Linux update compare to the other common consumer operating systems?
10 views
Skip to first unread message
paul
May 2, 2021, 9:53:54 PM5/2/21
to
In the Windows newsgroup we were discussing how Windows differs from iOS and
Android in terms of (a) frequency of update (b) length of update (c)
component separation and (d) mandatory versus voluntary (see below).
Can someone who knows Linux explain how it is most updated by consumers?
(You can pick the most common Linux consumer OS which I suspect is Ubuntu).
See the context of the question in the discussion included in the sig.
--
John Doe wrote on 02.05.2021 21:38
> That's how operating systems are, generally. You might be able to choose,
> Apple or Android, but you may as well get used to whatever they feed you.
What's really critical is to figure out HOW each operating system updates.
(I think most people still do not understand what I am about to say below.)
Windows 10 for example, seems to have (at least) 4 major components
(1) The operating system (updated about twice yearly, is voluntary I think)
(2) The security patches (updated about monthly, I'm not sure if mandatory)
(3) Microsoft key apps (updated asynchronously I think, voluntary I think)
(4) All the rest of your (non Microsoft core) apps (updated asynchronously)
Within reason all four of these main parts seem to be updated frequently for
(almost) as long as you can practically own the device (at least in terms of
pragmatic typical electronic system lifetimes I think).
In contrast Android has a different update model with different lifetimes.
(1) The operating system (updated about yearly, is voluntary)
(2) The security patches (updated about monthly, I think are also voluntary)
(3) Google key apps (updated many times a month I think, voluntary I think)
(4) Core framework of currently about 25 functions (updated often I think)
(5) All the rest of your (non Google core) apps (updated asynchronously)
The difference is parts #1 and #2 are pushed out by the device manufacturer.
And they push them out for three or four (how long depends on the OEM).
Parts #3 and #4 (eg Projects Treble & Mainline) are not only far more
frequent than most people realize but they both completely bypass the
manufacturer of the phone and they both occur at that frequency for
essentially the life of the Android device using the Google Play update
mechanism (at least within general practical electronic terms).
Part #5 also lasts essentially for the life of the device (as with Windows).
In huge contrast, currently iOS is completely different in terms of updates.
(1) The operating system (updated roughly about monthly or so, is voluntary)
(2) All the rest of your (non Apple core) apps (updated asynchronously)
And iOS is completely different in that there is a cutoff life that is
longer for some things and shorter for others.
Anyone who claims iOS is updated either longer or shorter than for the other
consumer operating system is just ignorant because for some things iOS is
updated longer and for other things it's updated for shorter lifetimes.
Likewise anyone who claims iOS is updated either more frequently or less
frequently is also ignorant because for some things iOS is updated more
frequently and for others iOS is updated far less frequently (mostly due to
the nature that everything is currently lumped into a single OS release).
Notice how hugely different these three operating system update models are?
They're NOT the same.
Not in length of support.
Nor in frequency.
Mostly due to how Android & Windows & iOS break components down into parts.
BTW, if someone can list how Linux and macOS break components into parts it
would be instructive for all of us in terms of learning the differences.
Android in terms of (a) frequency of update (b) length of update (c)
component separation and (d) mandatory versus voluntary (see below).
Can someone who knows Linux explain how it is most updated by consumers?
(You can pick the most common Linux consumer OS which I suspect is Ubuntu).
See the context of the question in the discussion included in the sig.
--
John Doe wrote on 02.05.2021 21:38
> That's how operating systems are, generally. You might be able to choose,
> Apple or Android, but you may as well get used to whatever they feed you.
What's really critical is to figure out HOW each operating system updates.
(I think most people still do not understand what I am about to say below.)
Windows 10 for example, seems to have (at least) 4 major components
(1) The operating system (updated about twice yearly, is voluntary I think)
(2) The security patches (updated about monthly, I'm not sure if mandatory)
(3) Microsoft key apps (updated asynchronously I think, voluntary I think)
(4) All the rest of your (non Microsoft core) apps (updated asynchronously)
Within reason all four of these main parts seem to be updated frequently for
(almost) as long as you can practically own the device (at least in terms of
pragmatic typical electronic system lifetimes I think).
In contrast Android has a different update model with different lifetimes.
(1) The operating system (updated about yearly, is voluntary)
(2) The security patches (updated about monthly, I think are also voluntary)
(3) Google key apps (updated many times a month I think, voluntary I think)
(4) Core framework of currently about 25 functions (updated often I think)
(5) All the rest of your (non Google core) apps (updated asynchronously)
The difference is parts #1 and #2 are pushed out by the device manufacturer.
And they push them out for three or four (how long depends on the OEM).
Parts #3 and #4 (eg Projects Treble & Mainline) are not only far more
frequent than most people realize but they both completely bypass the
manufacturer of the phone and they both occur at that frequency for
essentially the life of the Android device using the Google Play update
mechanism (at least within general practical electronic terms).
Part #5 also lasts essentially for the life of the device (as with Windows).
In huge contrast, currently iOS is completely different in terms of updates.
(1) The operating system (updated roughly about monthly or so, is voluntary)
(2) All the rest of your (non Apple core) apps (updated asynchronously)
And iOS is completely different in that there is a cutoff life that is
longer for some things and shorter for others.
Anyone who claims iOS is updated either longer or shorter than for the other
consumer operating system is just ignorant because for some things iOS is
updated longer and for other things it's updated for shorter lifetimes.
Likewise anyone who claims iOS is updated either more frequently or less
frequently is also ignorant because for some things iOS is updated more
frequently and for others iOS is updated far less frequently (mostly due to
the nature that everything is currently lumped into a single OS release).
Notice how hugely different these three operating system update models are?
They're NOT the same.
Not in length of support.
Nor in frequency.
Mostly due to how Android & Windows & iOS break components down into parts.
BTW, if someone can list how Linux and macOS break components into parts it
would be instructive for all of us in terms of learning the differences.
Mike Easter
May 2, 2021, 11:00:45 PM5/2/21
to
paul wrote:
> Can someone who knows Linux explain how it is most updated by consumers?
> (You can pick the most common Linux consumer OS which I suspect is Ubuntu).
The fact that there are so many distro/s and so many different
> Can someone who knows Linux explain how it is most updated by consumers?
> (You can pick the most common Linux consumer OS which I suspect is Ubuntu).
versions/spins of so many distro/s makes for a lot of variability. Your
Ub Gnome example is just one little slice, and Ub has its various
repositories and .ppa/s and numerous alternate desktop environment
flavors, and Ub chooses an update model of 6 month point releases,
periodic long term support releases, and/but not rolling.
Underneath there is the kernel which has its own release cycles, some of
which discussion was elucidated in the recent '30 years of Linux' w/
Linus Torvalds at Tag1.
The various distro/s, using one example of Mint which is based on the Ub
repo/s (and a separate spin based on Debian) have their own system of
'relating' to the user in terms of how the user wants to handle Security
Updates vs Software updates vs system snapshots to be able to revert back.
Your other proprietary examples of MS & Apple don't march to the same
drummer as linux users; the /external/ 'proprietor' is in complete
control of the 'situation' and the user is in the position of 'take it
or leave it' much more so than linux.
This says nothing of the other big families; Debian, Arch, RedHat,
Slackware. SUSE, Gentoo, and kin like BSD.
--
Mike Easter
J.O. Aho
May 3, 2021, 2:03:36 AM5/3/21
to
On 03/05/2021 03.53, paul wrote:
> In the Windows newsgroup we were discussing how Windows differs from iOS and
> Android in terms of (a) frequency of update (b) length of update (c)
> component separation and (d) mandatory versus voluntary (see below).
>
> Can someone who knows Linux explain how it is most updated by consumers?
> (You can pick the most common Linux consumer OS which I suspect is Ubuntu).
Lets just look at the official part, skipping ppa (third party
> In the Windows newsgroup we were discussing how Windows differs from iOS and
> Android in terms of (a) frequency of update (b) length of update (c)
> component separation and (d) mandatory versus voluntary (see below).
>
> Can someone who knows Linux explain how it is most updated by consumers?
> (You can pick the most common Linux consumer OS which I suspect is Ubuntu).
repositories), for *buntu it's a twice a year OS update (not mandatory),
The OS update includes upgrading kernel to a new major/minor version,
this also applies to software too.
There is no frequency for updated on minor/patch level of applications
and kernel between the OS updates, those are pushed out as soon as a
package has been verified to work. This leas to that security updates
are pushed out as soon as possible.
The major thing is that everything is up to the administrator of the
system to pick if packages are auto installed, manually installed or not
installed at all.
Most Linux distributions have a similar thought of releases, the OS
release cycle may differ and on smaller distributions it may be some
delay on package updates.
There is of course the more and more popular rolling release which you
can see in distributions like Gentoo/Funtoo, Arch/Artix, and otheres
where you don't have OS release cycles, things are updated during the
normal patch cycle, so that major/minor versions of kernel and
applications are done all the time, so you have the latest version of
software all the time. This model has the disadvantage that sometimes
there can be updates that can break things on your system and require
some extra hands on to fix it (this tend to be notified ahead of time
and with a how to fix it).
This is kind of like you installed ms-windows 98 back in the 1990's and
with your normal system updates you are today running ms-windows 10
without doing a system upgrade or installed a new version.
--
//Aho
Melzzzzz
May 3, 2021, 3:20:46 AM5/3/21
to
On 2021-05-03, paul <nos...@nospam.invalid> wrote:
> In the Windows newsgroup we were discussing how Windows differs from iOS and
> Android in terms of (a) frequency of update (b) length of update (c)
> component separation and (d) mandatory versus voluntary (see below).
>
> Can someone who knows Linux explain how it is most updated by consumers?
> (You can pick the most common Linux consumer OS which I suspect is Ubuntu).
>
> See the context of the question in the discussion included in the sig.
Windows update is horror, mildly speaking. My Manjaro updates in 1-2
> In the Windows newsgroup we were discussing how Windows differs from iOS and
> Android in terms of (a) frequency of update (b) length of update (c)
> component separation and (d) mandatory versus voluntary (see below).
>
> Can someone who knows Linux explain how it is most updated by consumers?
> (You can pick the most common Linux consumer OS which I suspect is Ubuntu).
>
> See the context of the question in the discussion included in the sig.
minutes, most time spent is downloading.
--
current job title: senior software engineer
skills: x86 aasembler,c++,c,rust,go,nim,haskell...
press any key to continue or any other to quit...
Paul
May 3, 2021, 4:40:04 AM5/3/21
to
Melzzzzz wrote:
> On 2021-05-03, paul <nos...@nospam.invalid> wrote:
>> In the Windows newsgroup we were discussing how Windows differs from iOS and
>> Android in terms of (a) frequency of update (b) length of update (c)
>> component separation and (d) mandatory versus voluntary (see below).
>>
>> Can someone who knows Linux explain how it is most updated by consumers?
>> (You can pick the most common Linux consumer OS which I suspect is Ubuntu).
>>
>> See the context of the question in the discussion included in the sig.
>
> Windows update is horror, mildly speaking. My Manjaro updates in 1-2
> minutes, most time spent is downloading.
It depends.
> On 2021-05-03, paul <nos...@nospam.invalid> wrote:
>> In the Windows newsgroup we were discussing how Windows differs from iOS and
>> Android in terms of (a) frequency of update (b) length of update (c)
>> component separation and (d) mandatory versus voluntary (see below).
>>
>> Can someone who knows Linux explain how it is most updated by consumers?
>> (You can pick the most common Linux consumer OS which I suspect is Ubuntu).
>>
>> See the context of the question in the discussion included in the sig.
>
> Windows update is horror, mildly speaking. My Manjaro updates in 1-2
> minutes, most time spent is downloading.
Windows works at a "package" level. Just like Linux.
When there is an OS Upgrade, that's a thousand packages, separately downloaded.
Sound familiar ?
Microsoft has been crippled by their package management, since WinXP.
The scheme doesn't scale well, especially if not curated properly.
If a software has updated 50 times since origin, the computation to
figure out version 50 is better than version 49, is lengthy.
DotNET updates take a long time. Part of that is NGEN,
the recompiling of .NET assemblies. Just like Linux
rebuilds the GRUB menu and initrd too many times during
updates, the .NET installs kick off NGEN recompiles a bit
too much. So those suck.
Windows uses "Jumbo Updates", which package together
many security fixes. Some of those are lengthy, because they
involve booting into a temporary profile while fixing
something. That's not good.
But on OS Upgrades, Microsoft has you beat. On Windows 10, if
the upgrade process goes tits up, the installer can revert to
original conditions. Without the user having to make a backup.
The process is now so good, it also includes protection
against the "reboot into new OS" failing. If the reboot into
new OS does not work, it reverts to the old OS. No Timeshift
or Clonezilla needed. It may be bloody slow reverting, but,
it works. It can take a whole hour to revert, as it unwinds itself.
But generally, at package level and security updates, it feels
"stodgy". Best practice is to delay the updates until the
smoke clears and it's safe to install them. You do your
updates, the day before Patch Tuesday. The day before the
next "batch of one" arrives.
Patch Tuesday is not the only day that updates come in. There
are other popular weeks for certain activities. But Patch Tuesday
is something everyone has heard of.
*******
Ubuntu has the annoyance, that when the Software Updater is
working out what to do (after the system boots), it blocks
Synaptic, such that the user cannot install applications
while that goes on in the background. Now, that's annoying.
And an example of how not to run a ship.
Windows has subsystems that asynchronously update. Metro applications
are upgraded separately from the Windows Update cycle.
Ubuntu does something similar, with Snap updates.
You can find all sorts of "similar analogies" if you
look carefully.
If you want hell on earth, try and figure out all the
knobs and levers on the Gentoo package management. The first
time I tried Gentoo, I could actually do stuff. Not any
more. What they did, is technically more sophisticated,
but it left me (a user) behind.
Paul
Peter Köhlmann
May 3, 2021, 7:20:24 AM5/3/21
to
Am 03.05.21 um 10:40 schrieb Paul:
It is OK that you don't know your ass from a hole in the ground when
linux is compared.
But simply don't assume that linux users take your word on limux
matters. They know that you are a dimbulb, when you talk about idiocy
like "Updates going wrong, but MS has the advantage".
Who do you think you are kidding? There were several examples of windows
being broken after an update, and no magical recovery anywhere in sight.
And simply ignoring that linux has snapshots too, to revert to an older
version, does not help your agenda.
In short: You are a babbling dimbulb
linux is compared.
But simply don't assume that linux users take your word on limux
matters. They know that you are a dimbulb, when you talk about idiocy
like "Updates going wrong, but MS has the advantage".
Who do you think you are kidding? There were several examples of windows
being broken after an update, and no magical recovery anywhere in sight.
And simply ignoring that linux has snapshots too, to revert to an older
version, does not help your agenda.
In short: You are a babbling dimbulb
Melzzzzz
May 3, 2021, 7:24:47 AM5/3/21
to
On 2021-05-03, Paul <nos...@needed.invalid> wrote:
> Melzzzzz wrote:
>> On 2021-05-03, paul <nos...@nospam.invalid> wrote:
>>> In the Windows newsgroup we were discussing how Windows differs from iOS and
>>> Android in terms of (a) frequency of update (b) length of update (c)
>>> component separation and (d) mandatory versus voluntary (see below).
>>>
>>> Can someone who knows Linux explain how it is most updated by consumers?
>>> (You can pick the most common Linux consumer OS which I suspect is Ubuntu).
>>>
>>> See the context of the question in the discussion included in the sig.
>>
>> Windows update is horror, mildly speaking. My Manjaro updates in 1-2
>> minutes, most time spent is downloading.
>
> It depends.
>
> Windows works at a "package" level. Just like Linux.
>
> When there is an OS Upgrade, that's a thousand packages, separately downloaded.
> Sound familiar ?
>
> Microsoft has been crippled by their package management, since WinXP.
> The scheme doesn't scale well, especially if not curated properly.
> If a software has updated 50 times since origin, the computation to
> figure out version 50 is better than version 49, is lengthy.
>
> DotNET updates take a long time. Part of that is NGEN,
> the recompiling of .NET assemblies. Just like Linux
> rebuilds the GRUB menu and initrd too many times during
> updates, the .NET installs kick off NGEN recompiles a bit
> too much. So those suck.
I examined why updates takes so long? Problem is that they
> Melzzzzz wrote:
>> On 2021-05-03, paul <nos...@nospam.invalid> wrote:
>>> In the Windows newsgroup we were discussing how Windows differs from iOS and
>>> Android in terms of (a) frequency of update (b) length of update (c)
>>> component separation and (d) mandatory versus voluntary (see below).
>>>
>>> Can someone who knows Linux explain how it is most updated by consumers?
>>> (You can pick the most common Linux consumer OS which I suspect is Ubuntu).
>>>
>>> See the context of the question in the discussion included in the sig.
>>
>> Windows update is horror, mildly speaking. My Manjaro updates in 1-2
>> minutes, most time spent is downloading.
>
> It depends.
>
> Windows works at a "package" level. Just like Linux.
>
> When there is an OS Upgrade, that's a thousand packages, separately downloaded.
> Sound familiar ?
>
> Microsoft has been crippled by their package management, since WinXP.
> The scheme doesn't scale well, especially if not curated properly.
> If a software has updated 50 times since origin, the computation to
> figure out version 50 is better than version 49, is lengthy.
>
> DotNET updates take a long time. Part of that is NGEN,
> the recompiling of .NET assemblies. Just like Linux
> rebuilds the GRUB menu and initrd too many times during
> updates, the .NET installs kick off NGEN recompiles a bit
> too much. So those suck.
do lot of overwrited of previous updates. They don't simply
apply lsat update rather they do 10 updates in a row till
last one.
>
> Windows uses "Jumbo Updates", which package together
> many security fixes. Some of those are lengthy, because they
> involve booting into a temporary profile while fixing
> something. That's not good.
Another problem is that they perform updates in background,
> Windows uses "Jumbo Updates", which package together
> many security fixes. Some of those are lengthy, because they
> involve booting into a temporary profile while fixing
> something. That's not good.
which makes computer more unusable then to apply on boot or shutdown.
So SSD is a must...
>
> But on OS Upgrades, Microsoft has you beat. On Windows 10, if
> the upgrade process goes tits up, the installer can revert to
> original conditions. Without the user having to make a backup.
> The process is now so good, it also includes protection
> against the "reboot into new OS" failing. If the reboot into
> new OS does not work, it reverts to the old OS. No Timeshift
> or Clonezilla needed. It may be bloody slow reverting, but,
> it works. It can take a whole hour to revert, as it unwinds itself.
Well, at least something...
> But on OS Upgrades, Microsoft has you beat. On Windows 10, if
> the upgrade process goes tits up, the installer can revert to
> original conditions. Without the user having to make a backup.
> The process is now so good, it also includes protection
> against the "reboot into new OS" failing. If the reboot into
> new OS does not work, it reverts to the old OS. No Timeshift
> or Clonezilla needed. It may be bloody slow reverting, but,
> it works. It can take a whole hour to revert, as it unwinds itself.
>
> But generally, at package level and security updates, it feels
> "stodgy". Best practice is to delay the updates until the
> smoke clears and it's safe to install them. You do your
> updates, the day before Patch Tuesday. The day before the
> next "batch of one" arrives.
yoi just delay final step.
>
> Patch Tuesday is not the only day that updates come in. There
> are other popular weeks for certain activities. But Patch Tuesday
> is something everyone has heard of.
>
> *******
>
> Ubuntu has the annoyance, that when the Software Updater is
> working out what to do (after the system boots), it blocks
> Synaptic, such that the user cannot install applications
> while that goes on in the background. Now, that's annoying.
> And an example of how not to run a ship.
I do from command line, if needed.
>
> Windows has subsystems that asynchronously update. Metro applications
> are upgraded separately from the Windows Update cycle.
>
> Ubuntu does something similar, with Snap updates.
>
> You can find all sorts of "similar analogies" if you
> look carefully.
>
> If you want hell on earth, try and figure out all the
> knobs and levers on the Gentoo package management. The first
> time I tried Gentoo, I could actually do stuff. Not any
> more. What they did, is technically more sophisticated,
> but it left me (a user) behind.
Don't have time for Gentoo. Tried once and quickly get bored.
> If you want hell on earth, try and figure out all the
> knobs and levers on the Gentoo package management. The first
> time I tried Gentoo, I could actually do stuff. Not any
> more. What they did, is technically more sophisticated,
> but it left me (a user) behind.
>
> Paul
Melzzzzz
May 3, 2021, 7:28:48 AM5/3/21
to
>
> In short: You are a babbling dimbulb
>
> In short: You are a babbling dimbulb
>
Carlos E.R.
May 3, 2021, 7:32:07 AM5/3/21
to
For example, on openSUSE the default is to have the system in btrfs with
snapshots enabled. So you update the OS, you reboot, it goes tits up, so
you reboot again to the previous update and consolidate. You are back to
the previous status in minutes.
...
--
Cheers, Carlos.
Carlos E.R.
May 3, 2021, 7:36:07 AM5/3/21
to
On 03/05/2021 03.53, paul wrote:
> In the Windows newsgroup we were discussing how Windows differs from iOS and
> Android in terms of (a) frequency of update (b) length of update (c)
> component separation and (d) mandatory versus voluntary (see below).
>
> Can someone who knows Linux explain how it is most updated by consumers?
> (You can pick the most common Linux consumer OS which I suspect is Ubuntu).
>
> See the context of the question in the discussion included in the sig.
>
It is impossible to answer, as there are a dozen or two of major
> Android in terms of (a) frequency of update (b) length of update (c)
> component separation and (d) mandatory versus voluntary (see below).
>
> Can someone who knows Linux explain how it is most updated by consumers?
> (You can pick the most common Linux consumer OS which I suspect is Ubuntu).
>
> See the context of the question in the discussion included in the sig.
>
distributions and hundreds of other distribution, so that the upgrade
methods have a lot of variance. Even inside one distribution there can
be several strategies.
I can only explain how it is done in the distribution I use.
--
Cheers, Carlos.
paul
May 3, 2021, 12:01:57 PM5/3/21
to
Carlos E.R. wrote on 03.05.2021 13:32
> It is impossible to answer, as there are a dozen or two of major
> distributions and hundreds of other distribution, so that the upgrade
> methods have a lot of variance. Even inside one distribution there can
> be several strategies.
I've read the Linux responses to date and I admit it's confusing, especially
as I own Windows, Android, and iOS so I only know how _they_ update.
a- I think I know the major component parts of those OSs.
b- I think I know the frequency of update of those component parts.
c- And for what period of time those component parts are typically updated.
I'm simply trying to ascertain the answer to those questions - for Linux.
(Which is the most common consumer desktop Linux distribution?)
As an example, when I compared Windows to Android to iOS, I found the
differences shockingly huge in terms of component parts, frequency of
update, period of time, who supplies the updates to the consumer, and how
the consumer obtains those updates (pull or push for example).
>
> I can only explain how it is done in the distribution I use.
I saw your explanation of how openSUSE updates - which I appreciate.
It's clear each consumer OS (Android, iOS, Windows at least) updates quite
differently in terms of components, frequencies, and support lifetimes.
But is there any way we can generalize how "most" consumer Linux's update?
a- What are the component parts and who supplies them to the consumer?
b- What is the frequency of update of those component parts?
c- For what period of time are those component parts typically updated?
d- How does the consumer obtain those component parts (pull or push)?
> It is impossible to answer, as there are a dozen or two of major
> distributions and hundreds of other distribution, so that the upgrade
> methods have a lot of variance. Even inside one distribution there can
> be several strategies.
as I own Windows, Android, and iOS so I only know how _they_ update.
a- I think I know the major component parts of those OSs.
b- I think I know the frequency of update of those component parts.
c- And for what period of time those component parts are typically updated.
I'm simply trying to ascertain the answer to those questions - for Linux.
(Which is the most common consumer desktop Linux distribution?)
As an example, when I compared Windows to Android to iOS, I found the
differences shockingly huge in terms of component parts, frequency of
update, period of time, who supplies the updates to the consumer, and how
the consumer obtains those updates (pull or push for example).
>
> I can only explain how it is done in the distribution I use.
It's clear each consumer OS (Android, iOS, Windows at least) updates quite
differently in terms of components, frequencies, and support lifetimes.
But is there any way we can generalize how "most" consumer Linux's update?
a- What are the component parts and who supplies them to the consumer?
b- What is the frequency of update of those component parts?
c- For what period of time are those component parts typically updated?
d- How does the consumer obtain those component parts (pull or push)?
Mike Easter
May 3, 2021, 12:12:22 PM5/3/21
to
paul wrote:
> But is there any way we can generalize how "most" consumer Linux's update?
This is beginning to sound very much like a classroom or 'homework'
> But is there any way we can generalize how "most" consumer Linux's update?
assignment.
> a- What are the component parts and who supplies them to the consumer?
> b- What is the frequency of update of those component parts?
> c- For what period of time are those component parts typically updated?
> d- How does the consumer obtain those component parts (pull or push)?
as the particular distro release is upgradable.
Alternatively, the user/consumer may obtain an entirely 'new'
distribution for a fresh install, almost always free.
--
Mike Easter
Soviet_Mario
May 3, 2021, 12:42:30 PM5/3/21
to
Il 03/05/21 03:53, paul ha scritto:
years) on Debian STABLE 80%, MX 10%, Mint LMDE, Manjaro,
Mageia and Fedora (the remaining 10%)
On debian I had adopted various layers on APT, initially
many gnome services like "Discover / Software", now Apper,
always synaptic.
If set up as autorun at startup, such tend to scan for
updates and notify. Synaptic defers to manual updates.
Since I notices the STABLE branch of Debian have relatively
little frequent updates, now I tend to use only MANUAL
update mode. Within the options, I chosed to scan for EVERY
KIND of update available, and automatically download, but to
let me choose to apply manually.
Linux MX (19) has a very handy updater/notifier, that
downloads in the backgroun and shows in the status part of
the default panel, and gets GREEN if some update is pending
to be applied.
Rather often it does not manage to apply all updates (then I
use text mode with --fix-missing option, which almost always
solves the problem).
When I use synaptic, normally I also make a further scan for
updates.
I have a lot of problems in fluid integration with pCloud
AppImage package, but I use it too seldom to dedicate enough
time to understand the error messages about "FUSE".
Mint LMDE has a really handy and rock solid gUI updater (but
I have no strange stuff installed on it)
Manjaro (I gave up using it somewhat more than a year ago).
Yes, it is rolling, so frequent updates are expected. But I
found that really too much frequent and too heavy. Many
heavy SW required updating.
Also the updater seemed really fragile for such a model of
deployment. I would have assumed a rolling release to be
really unbreakable. But often the updater failed to resolve
conflicts, and also the other front ends (Octopi ? Not sure
I recall it right) got stuck.
I have had the sensation that the system strictly required
timely update, or else sth would get broken, and I got
annoied with this.
Mageia Updater (I have a fresh and very standard system) is
again ROCKY, never fails even after a long time down. It
patiently downloads a lot of stuff, dependencies, and slowly
grinds all and goes through.
Fedora : I had used just 28 and 29. Fedora worked well. With
29 I noticed same problems as Manjaro : DNF-DRAGORA updater
ran frequently in problems I were not able to cope with.
Apart from Mageia (which does everything by itself with no
intervention), on Fedora I don't feel comfortable as it is
the only RPM package manager based system I had tried. Maybe
my poor confidence was the problem, with dnf-dragora
updater. Dunno.
Now I am relying only on .DEB based systems.
I also find it easier to locate programs by name, and to
find extrnal tools for DEB than for other formats.
My Penny
--
1) Resistere, resistere, resistere.
2) Se tutti pagano le tasse, le tasse le pagano tutti
Soviet_Mario - (aka Gatto_Vizzato)
> In the Windows newsgroup we were discussing how Windows differs from iOS and
> Android in terms of (a) frequency of update (b) length of update (c)
> component separation and (d) mandatory versus voluntary (see below).
>
> Can someone who knows Linux explain how it is most updated by consumers?
> (You can pick the most common Linux consumer OS which I suspect is Ubuntu).
>
> See the context of the question in the discussion included in the sig.
>
I have most of my relatively scarce experience (about 4
> Android in terms of (a) frequency of update (b) length of update (c)
> component separation and (d) mandatory versus voluntary (see below).
>
> Can someone who knows Linux explain how it is most updated by consumers?
> (You can pick the most common Linux consumer OS which I suspect is Ubuntu).
>
> See the context of the question in the discussion included in the sig.
>
years) on Debian STABLE 80%, MX 10%, Mint LMDE, Manjaro,
Mageia and Fedora (the remaining 10%)
On debian I had adopted various layers on APT, initially
many gnome services like "Discover / Software", now Apper,
always synaptic.
If set up as autorun at startup, such tend to scan for
updates and notify. Synaptic defers to manual updates.
Since I notices the STABLE branch of Debian have relatively
little frequent updates, now I tend to use only MANUAL
update mode. Within the options, I chosed to scan for EVERY
KIND of update available, and automatically download, but to
let me choose to apply manually.
Linux MX (19) has a very handy updater/notifier, that
downloads in the backgroun and shows in the status part of
the default panel, and gets GREEN if some update is pending
to be applied.
Rather often it does not manage to apply all updates (then I
use text mode with --fix-missing option, which almost always
solves the problem).
When I use synaptic, normally I also make a further scan for
updates.
I have a lot of problems in fluid integration with pCloud
AppImage package, but I use it too seldom to dedicate enough
time to understand the error messages about "FUSE".
Mint LMDE has a really handy and rock solid gUI updater (but
I have no strange stuff installed on it)
Manjaro (I gave up using it somewhat more than a year ago).
Yes, it is rolling, so frequent updates are expected. But I
found that really too much frequent and too heavy. Many
heavy SW required updating.
Also the updater seemed really fragile for such a model of
deployment. I would have assumed a rolling release to be
really unbreakable. But often the updater failed to resolve
conflicts, and also the other front ends (Octopi ? Not sure
I recall it right) got stuck.
I have had the sensation that the system strictly required
timely update, or else sth would get broken, and I got
annoied with this.
Mageia Updater (I have a fresh and very standard system) is
again ROCKY, never fails even after a long time down. It
patiently downloads a lot of stuff, dependencies, and slowly
grinds all and goes through.
Fedora : I had used just 28 and 29. Fedora worked well. With
29 I noticed same problems as Manjaro : DNF-DRAGORA updater
ran frequently in problems I were not able to cope with.
Apart from Mageia (which does everything by itself with no
intervention), on Fedora I don't feel comfortable as it is
the only RPM package manager based system I had tried. Maybe
my poor confidence was the problem, with dnf-dragora
updater. Dunno.
Now I am relying only on .DEB based systems.
I also find it easier to locate programs by name, and to
find extrnal tools for DEB than for other formats.
My Penny
--
1) Resistere, resistere, resistere.
2) Se tutti pagano le tasse, le tasse le pagano tutti
Soviet_Mario - (aka Gatto_Vizzato)
William Unruh
May 3, 2021, 12:51:45 PM5/3/21
to
On 2021-05-03, paul <nos...@nospam.invalid> wrote:
> Carlos E.R. wrote on 03.05.2021 13:32
>
>> It is impossible to answer, as there are a dozen or two of major
>> distributions and hundreds of other distribution, so that the upgrade
>> methods have a lot of variance. Even inside one distribution there can
>> be several strategies.
>
> I've read the Linux responses to date and I admit it's confusing, especially
> as I own Windows, Android, and iOS so I only know how _they_ update.
> a- I think I know the major component parts of those OSs.
> b- I think I know the frequency of update of those component parts.
> c- And for what period of time those component parts are typically updated.
Any you have been answered. It depends on the distribution. For major
>
>> It is impossible to answer, as there are a dozen or two of major
>> distributions and hundreds of other distribution, so that the upgrade
>> methods have a lot of variance. Even inside one distribution there can
>> be several strategies.
>
> I've read the Linux responses to date and I admit it's confusing, especially
> as I own Windows, Android, and iOS so I only know how _they_ update.
> a- I think I know the major component parts of those OSs.
> b- I think I know the frequency of update of those component parts.
> c- And for what period of time those component parts are typically updated.
updates it will be between a few days for rolling distributions, to 2
years (for the one I use Mageia). However, for each major updates, there
will be updates to components which will arrive conginuously as they are
generated and tested. Thus in my case, even new kernel updates will come
out about once a month.
Security updates will come out within a few days that a security fix for
a Linux component will come out.
The updates are posted on distribution sites. Your distribution will
probably have a process running which checks for updates, and notifies
you when they are available, so you can initiate the download and
installation. Almost all distributions will leave it up to you whether
or not you want to install them, or any particular update.
So, it depends.
Any particular release will continue to receive support at least until
the next release comes out and about 6 months thereafter. But the length
of support will depend on the distribution. Thus the Redhat reelases
will give you support for something like 5 years after release. The
Redhat "community" releases (Fedora) will offer support for about 6
months after the next release comes out. Rolling release distributions
could be said to offer continuous support, but if you say do not keep
with the rolling you may find that a recent update to some program will
no longer work with your 5-year-not-updated system.
>
> I'm simply trying to ascertain the answer to those questions - for Linux.
> (Which is the most common consumer desktop Linux distribution?)
>
> As an example, when I compared Windows to Android to iOS, I found the
> differences shockingly huge in terms of component parts, frequency of
> update, period of time, who supplies the updates to the consumer, and how
> the consumer obtains those updates (pull or push for example).
>
>>
>> I can only explain how it is done in the distribution I use.
>
> I saw your explanation of how openSUSE updates - which I appreciate.
>
> It's clear each consumer OS (Android, iOS, Windows at least) updates quite
> differently in terms of components, frequencies, and support lifetimes.
>
> But is there any way we can generalize how "most" consumer Linux's update?
> a- What are the component parts and who supplies them to the consumer?
Teh "component parts" are every program, or library in the system.
> As an example, when I compared Windows to Android to iOS, I found the
> differences shockingly huge in terms of component parts, frequency of
> update, period of time, who supplies the updates to the consumer, and how
> the consumer obtains those updates (pull or push for example).
>
>>
>> I can only explain how it is done in the distribution I use.
>
> I saw your explanation of how openSUSE updates - which I appreciate.
>
> It's clear each consumer OS (Android, iOS, Windows at least) updates quite
> differently in terms of components, frequencies, and support lifetimes.
>
> But is there any way we can generalize how "most" consumer Linux's update?
> a- What are the component parts and who supplies them to the consumer?
> b- What is the frequency of update of those component parts?
years.
> c- For what period of time are those component parts typically updated?
No idea what that means. updates are not withdrawn.
> d- How does the consumer obtain those component parts (pull or push)?
notification that component parts have updates available. It is then up
to the consumer to initiate the updates. So it is announced pull.
William Unruh
May 3, 2021, 2:05:00 PM5/3/21
to
On 2021-05-03, Soviet_Mario <Sovie...@CCCP.MIR> wrote:
....
problems.
> patiently downloads a lot of stuff, dependencies, and slowly
> grinds all and goes through.
>
> Fedora : I had used just 28 and 29. Fedora worked well. With
> 29 I noticed same problems as Manjaro : DNF-DRAGORA updater
> ran frequently in problems I were not able to cope with.
>
> Apart from Mageia (which does everything by itself with no
> intervention), on Fedora I don't feel comfortable as it is
It can, but usually the user needs to initiate it. After initiation it
can replace all of the needed packages on its own. Ie, it does not carry
out updating on its own. You can also specify programs that you do not
want to have updated.
> the only RPM package manager based system I had tried. Maybe
Mageia is an rpm based system as well. It has a user layer, urpmi which
handles things like dependencies. (they are also using dnf but I do not
know much about it, but it is still rpm based.)
>
....
>
> Mageia Updater (I have a fresh and very standard system) is
> again ROCKY, never fails even after a long time down. It
I presume you mean rock solid. Rocky means the opposite-- lots of
> Mageia Updater (I have a fresh and very standard system) is
> again ROCKY, never fails even after a long time down. It
problems.
> patiently downloads a lot of stuff, dependencies, and slowly
> grinds all and goes through.
>
> Fedora : I had used just 28 and 29. Fedora worked well. With
> 29 I noticed same problems as Manjaro : DNF-DRAGORA updater
> ran frequently in problems I were not able to cope with.
>
> Apart from Mageia (which does everything by itself with no
> intervention), on Fedora I don't feel comfortable as it is
can replace all of the needed packages on its own. Ie, it does not carry
out updating on its own. You can also specify programs that you do not
want to have updated.
> the only RPM package manager based system I had tried. Maybe
handles things like dependencies. (they are also using dnf but I do not
know much about it, but it is still rpm based.)
> my poor confidence was the problem, with dnf-dragora
> updater. Dunno.
...
> updater. Dunno.
>
Soviet_Mario
May 3, 2021, 2:30:54 PM5/3/21
to
Il 03/05/21 20:04, William Unruh ha scritto:
solid :D
>
>> patiently downloads a lot of stuff, dependencies, and slowly
>> grinds all and goes through.
>>
>> Fedora : I had used just 28 and 29. Fedora worked well. With
>> 29 I noticed same problems as Manjaro : DNF-DRAGORA updater
>> ran frequently in problems I were not able to cope with.
>
>
>>
>> Apart from Mageia (which does everything by itself with no
>> intervention), on Fedora I don't feel comfortable as it is
>
> It can, but usually the user needs to initiate it. After initiation it
> can replace all of the needed packages on its own. Ie, it does not carry
> out updating on its own. You can also specify programs that you do not
> want to have updated.
yes, the gui front-end actually was sufficiently
"synaptic-like" to let me manage stuff (when it worked)
>
>> the only RPM package manager based system I had tried. Maybe
> Mageia is an rpm based system as well.
I know, but its updater is so smart (not much the Gui,
rather vintage, not recognizing the mouse wheel !) that
shields me from any manual intervention.
Mageia gives the impression of tidyness of parts, even if
not so sleek as Fedora.
> It has a user layer, urpmi which
> handles things like dependencies. (they are also using dnf but I do not
> know much about it, but it is still rpm based.)
>> my poor confidence was the problem, with dnf-dragora
>> updater. Dunno.
> ...
>>
> On 2021-05-03, Soviet_Mario <Sovie...@CCCP.MIR> wrote:
> ....
>>
>> Mageia Updater (I have a fresh and very standard system) is
>> again ROCKY, never fails even after a long time down. It
>
> I presume you mean rock solid. Rocky means the opposite-- lots of
> problems.
oh ! Sorry, misunderstanding. Sure, I meant exactly rock
> ....
>>
>> Mageia Updater (I have a fresh and very standard system) is
>> again ROCKY, never fails even after a long time down. It
>
> I presume you mean rock solid. Rocky means the opposite-- lots of
> problems.
solid :D
>
>> patiently downloads a lot of stuff, dependencies, and slowly
>> grinds all and goes through.
>>
>> Fedora : I had used just 28 and 29. Fedora worked well. With
>> 29 I noticed same problems as Manjaro : DNF-DRAGORA updater
>> ran frequently in problems I were not able to cope with.
>
>
>>
>> Apart from Mageia (which does everything by itself with no
>> intervention), on Fedora I don't feel comfortable as it is
>
> It can, but usually the user needs to initiate it. After initiation it
> can replace all of the needed packages on its own. Ie, it does not carry
> out updating on its own. You can also specify programs that you do not
> want to have updated.
"synaptic-like" to let me manage stuff (when it worked)
>
>> the only RPM package manager based system I had tried. Maybe
> Mageia is an rpm based system as well.
rather vintage, not recognizing the mouse wheel !) that
shields me from any manual intervention.
Mageia gives the impression of tidyness of parts, even if
not so sleek as Fedora.
> It has a user layer, urpmi which
> handles things like dependencies. (they are also using dnf but I do not
> know much about it, but it is still rpm based.)
>> my poor confidence was the problem, with dnf-dragora
>> updater. Dunno.
> ...
>>
Carlos E.R.
May 3, 2021, 2:56:07 PM5/3/21
to
On 03/05/2021 18.01, paul wrote:
> Carlos E.R. wrote on 03.05.2021 13:32
>
>> It is impossible to answer, as there are a dozen or two of major
>> distributions and hundreds of other distribution, so that the upgrade
>> methods have a lot of variance. Even inside one distribution there can
>> be several strategies.
>
> I've read the Linux responses to date and I admit it's confusing, especially
> as I own Windows, Android, and iOS so I only know how _they_ update.
> a- I think I know the major component parts of those OSs.
> b- I think I know the frequency of update of those component parts.
> c- And for what period of time those component parts are typically updated.
>
> I'm simply trying to ascertain the answer to those questions - for Linux.
> (Which is the most common consumer desktop Linux distribution?)
That is also open to debate.
> Carlos E.R. wrote on 03.05.2021 13:32
>
>> It is impossible to answer, as there are a dozen or two of major
>> distributions and hundreds of other distribution, so that the upgrade
>> methods have a lot of variance. Even inside one distribution there can
>> be several strategies.
>
> I've read the Linux responses to date and I admit it's confusing, especially
> as I own Windows, Android, and iOS so I only know how _they_ update.
> a- I think I know the major component parts of those OSs.
> b- I think I know the frequency of update of those component parts.
> c- And for what period of time those component parts are typically updated.
>
> I'm simply trying to ascertain the answer to those questions - for Linux.
> (Which is the most common consumer desktop Linux distribution?)
As there are no sales figures, it is impossible to get actual numbers of
Linux usage. If you try to count "downloads", then you should consider
that a single download can be used for several installations, even
thousands.
Popularity also changes depending on the type of user or the
installation they do.
> As an example, when I compared Windows to Android to iOS, I found the
> differences shockingly huge in terms of component parts, frequency of
> update, period of time, who supplies the updates to the consumer, and how
> the consumer obtains those updates (pull or push for example).
>
>>
>> I can only explain how it is done in the distribution I use.
>
> I saw your explanation of how openSUSE updates - which I appreciate.
>
> It's clear each consumer OS (Android, iOS, Windows at least) updates quite
> differently in terms of components, frequencies, and support lifetimes.
>
> But is there any way we can generalize how "most" consumer Linux's update?
> a- What are the component parts and who supplies them to the consumer?
other libraries forming the core, the gnu tools, other tools, the
desktops, and a myriad different "services" and applications. All of
them are supplied to me by the same single source. Single point of
update but from several mirrors.
> b- What is the frequency of update of those component parts?
anywhere in the world, it is up to them when they deliver an update and
its "importance".
> c- For what period of time are those component parts typically updated?
Some may be "long life" distributions, which can last about a decade
tops, I think, others are continuous updating (rolling), and in a month
everything has changed.
> d- How does the consumer obtain those component parts (pull or push)?
course, if you have a support contract and do not update, the contract
may be invalidated if you don't update in time.
Say, a hack is found, a patch is published, you do not update, and then
your system is attacked. Well, it is your fault.
I recommend you seek for and read "The Cathedral and the Bazaar".
<https://en.wikipedia.org/wiki/The_Cathedral_and_the_Bazaar>
--
Cheers, Carlos.
Carlos E.R.
May 3, 2021, 4:48:08 PM5/3/21
to
On 03/05/2021 18.51, William Unruh wrote:
> On 2021-05-03, paul <nos...@nospam.invalid> wrote:
...
> On 2021-05-03, paul <nos...@nospam.invalid> wrote:
>> d- How does the consumer obtain those component parts (pull or push)?
>
> That is up to the consumer, but most distributions will have automatic
> notification that component parts have updates available. It is then up
> to the consumer to initiate the updates. So it is announced pull.
system connects to some server and asks somehow if there are updates
available.
The distribution can publish announcements on some web page, or they can
send emails to you (maybe that can be considered "pushing").
With many variations.
--
Cheers, Carlos.
RonB
May 3, 2021, 7:42:49 PM5/3/21
to
On 2021-05-03, paul <nos...@nospam.invalid> wrote:
> In the Windows newsgroup we were discussing how Windows differs from iOS and
> Android in terms of (a) frequency of update (b) length of update (c)
> component separation and (d) mandatory versus voluntary (see below).
>
> Can someone who knows Linux explain how it is most updated by consumers?
> (You can pick the most common Linux consumer OS which I suspect is Ubuntu).
>
> See the context of the question in the discussion included in the sig.
With Linux Mint you just get a notifcation in the panel when an update is
> Android in terms of (a) frequency of update (b) length of update (c)
> component separation and (d) mandatory versus voluntary (see below).
>
> Can someone who knows Linux explain how it is most updated by consumers?
> (You can pick the most common Linux consumer OS which I suspect is Ubuntu).
>
> See the context of the question in the discussion included in the sig.
available. I normally update when I see that notification. Most updates are
small and just take a minute or so. If you opt to update the Kernel (which I
do) it will ask you to reboot after the kernel update (optional, but I
ususally do it) — this happens maybe once every other month (maybe every six
weeks).
With the machines I have running Debian 10.9, I don't have any automatic
setting for update notifications. About every week or so (when I'm using one
of those computers) I just run two commands from the CLI.
sudo apt update
and then...
sudo apt upgrade
In some ways I like that better than being notified for each small update.
--
Saudis bomb & starve Yemeni civilians: (crickets)
Yemenis attack Saudi oil facilities: TERRORISM!
Aragorn
May 3, 2021, 11:15:32 PM5/3/21
to
On 03.05.2021 at 07:20, Melzzzzz scribbled:
> On 2021-05-03, paul <nos...@nospam.invalid> wrote:
>
> > Can someone who knows Linux explain how it is most updated by
> > consumers? (You can pick the most common Linux consumer OS which I
> > suspect is Ubuntu).
>
you're talking of a minor update with only three or four packages.
I'm running the Stable branch of Manjaro, and the latest bundled update
to Stable was on 28.04.2021. The downloading took about 10 minutes, and
the actual update process took about half an hour, not including about
10 minutes for two AUR packages that needed to be rebuilt from sources.
In the past, and with much larger updates, I've needed about an hour and
a half, including the time needed for downloading the packages.
For clarity, I do run a fairly well-packed system here. I have 1982
installed packages — my /usr currently stands at 7.2 GiB,
and that's with zstd compression on a separate btrfs filesystem — and I
did remove some of the junk that comes preinstalled on a new install,
such as gvfs, Steam, AppImage support and FlatPak support — Snap was
never even installed here.
--
With respect,
= Aragorn =
> On 2021-05-03, paul <nos...@nospam.invalid> wrote:
>
> > Can someone who knows Linux explain how it is most updated by
> > consumers? (You can pick the most common Linux consumer OS which I
> > suspect is Ubuntu).
>
> Windows update is horror, mildly speaking. My Manjaro updates in 1-2
> minutes, most time spent is downloading.
1 or 2 minutes? You must be running a very thin system then, or else
> minutes, most time spent is downloading.
you're talking of a minor update with only three or four packages.
I'm running the Stable branch of Manjaro, and the latest bundled update
to Stable was on 28.04.2021. The downloading took about 10 minutes, and
the actual update process took about half an hour, not including about
10 minutes for two AUR packages that needed to be rebuilt from sources.
In the past, and with much larger updates, I've needed about an hour and
a half, including the time needed for downloading the packages.
For clarity, I do run a fairly well-packed system here. I have 1982
installed packages — my /usr currently stands at 7.2 GiB,
and that's with zstd compression on a separate btrfs filesystem — and I
did remove some of the junk that comes preinstalled on a new install,
such as gvfs, Steam, AppImage support and FlatPak support — Snap was
never even installed here.
--
With respect,
= Aragorn =
Soviet_Mario
May 4, 2021, 6:14:26 AM5/4/21
to
Il 04/05/21 05:15, Aragorn ha scritto:
when I tried (for a while) Manjaro "Rolling" (some 1-2 years
ago), not only did an average update take up such a time
(half an hour or so), but that happened really frequently if
not every time, and from time to time it took longer.
I had a much heavier system though, stuffed with any
thinkable goodies from the repo.
The fact imho is : not every single piece of sw needs a so
frantic development cycle.
I can accept web browser, maybe also email client, and every
sw for privacy-security, and maybe even "cloud-oriented"
clients or alike (telegram desktop, owncloud and so).
But MOST of heavy sw (office suites, compilers, graphics,
sound and video) imho are not so urgent to keep up the
frantic pace.
Just my opinion, obv.
ago), not only did an average update take up such a time
(half an hour or so), but that happened really frequently if
not every time, and from time to time it took longer.
I had a much heavier system though, stuffed with any
thinkable goodies from the repo.
The fact imho is : not every single piece of sw needs a so
frantic development cycle.
I can accept web browser, maybe also email client, and every
sw for privacy-security, and maybe even "cloud-oriented"
clients or alike (telegram desktop, owncloud and so).
But MOST of heavy sw (office suites, compilers, graphics,
sound and video) imho are not so urgent to keep up the
frantic pace.
Just my opinion, obv.
Soviet_Mario
May 4, 2021, 6:52:28 AM5/4/21
to
Il 03/05/21 20:55, Carlos E.R. ha scritto:
but also the other way around. I have downloaded much more
distros than I actually use : I just gave them a try.
I have not the palest idea of which phenomenon prevails.
also, I often look in DISTROWHATCH, but, apart from the fact
that I have no clue if this site in particular is more
reliable than many others I may well don't even know, the
mentioned figures are always "per month" or so.
There is no esteem of the cumulative data.
The same difference between "deficit" and "debt".
I see that MX has been standing at 1st rank for a rather
long time. But how much time ? What are the TOTAL numbers ?
They are not reported.
It would be nice to have them.
Also : when a distro changes main number, the total is to be
reset ? maybe not, maybe yes ... after some larger change in
the main number, surely the figures should not be aggregated.
Silly to add the figures of an abandoned distro with its
successor.
It's complicated !
>
> Popularity also changes depending on the type of user or the
> installation they do.
pretty right. Server vs desktop vs mobile devices vs IoT are
hardly comparable.
>
>> As an example, when I compared Windows to Android to iOS,
>> I found the
>> differences shockingly huge in terms of component parts,
>> frequency of
>> update, period of time, who supplies the updates to the
>> consumer, and how
>> the consumer obtains those updates (pull or push for
>> example).
>>
>>>
>>> I can only explain how it is done in the distribution I use.
>>
>> I saw your explanation of how openSUSE updates - which I
>> appreciate.
>>
>> It's clear each consumer OS (Android, iOS, Windows at
>> least) updates quite
>> differently in terms of components, frequencies, and
>> support lifetimes.
>>
>> But is there any way we can generalize how "most" consumer
>> Linux's update?
>
> Nope.
dunno too. I think that a novice will firstly rely upon
built-in updaters, and after a certain time, he will
smoothly switch to more frequent manual update to have more
control.
Also "pinning" is a feature (influencing the update system)
likely not used by newbies, and that only advanced users
exploit.
>
>> a- What are the component parts and who supplies them to
>> the consumer?
>
> The distinction is fuzzy. You could consider the kernel, the
> libc and other libraries forming the core, the gnu tools,
> other tools, the desktops, and a myriad different "services"
> and applications. All of them are supplied to me by the same
> single source. Single point of update but from several mirrors.
>
>> b- What is the frequency of update of those component parts?
>
> Varies a lot. Each single "part" is developed by a different
> group anywhere in the world, it is up to them when they
> deliver an update and its "importance".
>
>> c- For what period of time are those component parts
>> typically updated?
>
> Again, varies a lot.
>
> Some may be "long life" distributions, which can last about
> a decade tops, I think, others are continuous updating
> (rolling), and in a month everything has changed.
>
>> d- How does the consumer obtain those component parts
>> (pull or push)?
>
> Polling and pulling. I'm not aware of anyone doing
> "pushing".
<EVIL>
maybe Snap store ... ;D
</EVIL>
BitBleach (a cleaning tool) looks for its own update at each
run (maybe this can be disabled), even if installed from std
repo.
> But of course, if you have a support contract and
> do not update, the contract may be invalidated if you don't
> update in time.
>
> Say, a hack is found, a patch is published, you do not
> update, and then your system is attacked. Well, it is your
> fault.
>
>
>
>
>
> I recommend you seek for and read "The Cathedral and the
> Bazaar".
>
> <https://en.wikipedia.org/wiki/The_Cathedral_and_the_Bazaar>
>
>
>
--
distros than I actually use : I just gave them a try.
I have not the palest idea of which phenomenon prevails.
also, I often look in DISTROWHATCH, but, apart from the fact
that I have no clue if this site in particular is more
reliable than many others I may well don't even know, the
mentioned figures are always "per month" or so.
There is no esteem of the cumulative data.
The same difference between "deficit" and "debt".
I see that MX has been standing at 1st rank for a rather
long time. But how much time ? What are the TOTAL numbers ?
They are not reported.
It would be nice to have them.
Also : when a distro changes main number, the total is to be
reset ? maybe not, maybe yes ... after some larger change in
the main number, surely the figures should not be aggregated.
Silly to add the figures of an abandoned distro with its
successor.
It's complicated !
>
> Popularity also changes depending on the type of user or the
> installation they do.
hardly comparable.
>
>> As an example, when I compared Windows to Android to iOS,
>> I found the
>> differences shockingly huge in terms of component parts,
>> frequency of
>> update, period of time, who supplies the updates to the
>> consumer, and how
>> the consumer obtains those updates (pull or push for
>> example).
>>
>>>
>>> I can only explain how it is done in the distribution I use.
>>
>> I saw your explanation of how openSUSE updates - which I
>> appreciate.
>>
>> It's clear each consumer OS (Android, iOS, Windows at
>> least) updates quite
>> differently in terms of components, frequencies, and
>> support lifetimes.
>>
>> But is there any way we can generalize how "most" consumer
>> Linux's update?
>
> Nope.
built-in updaters, and after a certain time, he will
smoothly switch to more frequent manual update to have more
control.
Also "pinning" is a feature (influencing the update system)
likely not used by newbies, and that only advanced users
exploit.
>
>> a- What are the component parts and who supplies them to
>> the consumer?
>
> The distinction is fuzzy. You could consider the kernel, the
> libc and other libraries forming the core, the gnu tools,
> other tools, the desktops, and a myriad different "services"
> and applications. All of them are supplied to me by the same
> single source. Single point of update but from several mirrors.
>
>> b- What is the frequency of update of those component parts?
>
> Varies a lot. Each single "part" is developed by a different
> group anywhere in the world, it is up to them when they
> deliver an update and its "importance".
>
>> c- For what period of time are those component parts
>> typically updated?
>
> Again, varies a lot.
>
> Some may be "long life" distributions, which can last about
> a decade tops, I think, others are continuous updating
> (rolling), and in a month everything has changed.
>
>> d- How does the consumer obtain those component parts
>> (pull or push)?
>
> Polling and pulling. I'm not aware of anyone doing
> "pushing".
maybe Snap store ... ;D
</EVIL>
BitBleach (a cleaning tool) looks for its own update at each
run (maybe this can be disabled), even if installed from std
repo.
> But of course, if you have a support contract and
> do not update, the contract may be invalidated if you don't
> update in time.
>
> Say, a hack is found, a patch is published, you do not
> update, and then your system is attacked. Well, it is your
> fault.
>
>
>
>
>
> I recommend you seek for and read "The Cathedral and the
> Bazaar".
>
> <https://en.wikipedia.org/wiki/The_Cathedral_and_the_Bazaar>
>
>
>
--
Aragorn
May 4, 2021, 6:58:54 AM5/4/21
to
On 04.05.2021 at 12:14, Soviet_Mario scribbled:
> Il 04/05/21 05:15, Aragorn ha scritto:
>
updates per month. Rarely ever will it be more frequently, and
sometimes it is less frequently, e.g. when a major component of the
system is upgraded to a newer release and as such, more testing is
desirable.
> The fact imho is : not every single piece of sw needs a so
> frantic development cycle.
It's not about a frantic development cycle, but about keeping the
system fully up-to-date — it is after all a rolling release — and not
letting the gap between Arch and Manjaro get too wide.
Arch Stable is on par with Manjaro Unstable, and then the newer
versions of the software trickle down into Manjaro Testing and — if
necessary — Manjaro Stable-Staging, and eventually then into Manjaro
Stable. This puts Manjaro Stable about six weeks behind on Arch, with
the exception of urgent updates for individual packages for stability
or security reasons.
> I can accept web browser, maybe also email client, and every
> sw for privacy-security, and maybe even "cloud-oriented"
> clients or alike (telegram desktop, owncloud and so).
>
> But MOST of heavy sw (office suites, compilers, graphics,
> sound and video) imho are not so urgent to keep up the
> frantic pace.
> Just my opinion, obv.
Not every major update also updates the office suites, or for that
matter, everything else that's installed. But every major update does
by definition include an updated kernel, and when you're running a
desktop environment like KDE Plasma, GNOME or XFCE, then there might be
updates to that, or to one of the subsystems of the environment.
For instance, maybe Plasma itself isn't being updated, but the KDE
Frameworks (formerly kdelibs) might be, or the KDE Applications, which
includes just about all of the utilities in Plasma.
But — again — this is what you sign up for when you're installing a
rolling-release distribution. At least in Manjaro, normal updates are
curated and bundled together, whereas if you run Arch proper, then you
might have (smaller) updates every single day, or possibly multiple
times a day.
> Il 04/05/21 05:15, Aragorn ha scritto:
>
> > I'm running the Stable branch of Manjaro, and the latest bundled
> > update to Stable was on 28.04.2021. The downloading took about 10
> > minutes, and the actual update process took about half an hour, not
> > including about 10 minutes for two AUR packages that needed to be
> > rebuilt from sources. In the past, and with much larger updates,
> > I've needed about an hour and a half, including the time needed for
> > downloading the packages.
>
> > update to Stable was on 28.04.2021. The downloading took about 10
> > minutes, and the actual update process took about half an hour, not
> > including about 10 minutes for two AUR packages that needed to be
> > rebuilt from sources. In the past, and with much larger updates,
> > I've needed about an hour and a half, including the time needed for
> > downloading the packages.
>
> when I tried (for a while) Manjaro "Rolling" (some 1-2 years
> ago), not only did an average update take up such a time
> (half an hour or so), but that happened really frequently if
> not every time, and from time to time it took longer.
On average, the Stable branch of Manjaro sees about two such major
> ago), not only did an average update take up such a time
> (half an hour or so), but that happened really frequently if
> not every time, and from time to time it took longer.
updates per month. Rarely ever will it be more frequently, and
sometimes it is less frequently, e.g. when a major component of the
system is upgraded to a newer release and as such, more testing is
desirable.
> The fact imho is : not every single piece of sw needs a so
> frantic development cycle.
system fully up-to-date — it is after all a rolling release — and not
letting the gap between Arch and Manjaro get too wide.
Arch Stable is on par with Manjaro Unstable, and then the newer
versions of the software trickle down into Manjaro Testing and — if
necessary — Manjaro Stable-Staging, and eventually then into Manjaro
Stable. This puts Manjaro Stable about six weeks behind on Arch, with
the exception of urgent updates for individual packages for stability
or security reasons.
> I can accept web browser, maybe also email client, and every
> sw for privacy-security, and maybe even "cloud-oriented"
> clients or alike (telegram desktop, owncloud and so).
>
> But MOST of heavy sw (office suites, compilers, graphics,
> sound and video) imho are not so urgent to keep up the
> frantic pace.
> Just my opinion, obv.
matter, everything else that's installed. But every major update does
by definition include an updated kernel, and when you're running a
desktop environment like KDE Plasma, GNOME or XFCE, then there might be
updates to that, or to one of the subsystems of the environment.
For instance, maybe Plasma itself isn't being updated, but the KDE
Frameworks (formerly kdelibs) might be, or the KDE Applications, which
includes just about all of the utilities in Plasma.
But — again — this is what you sign up for when you're installing a
rolling-release distribution. At least in Manjaro, normal updates are
curated and bundled together, whereas if you run Arch proper, then you
might have (smaller) updates every single day, or possibly multiple
times a day.
Carlos E.R.
May 4, 2021, 8:56:08 AM5/4/21
to
On 04/05/2021 12.52, Soviet_Mario wrote:
> Il 03/05/21 20:55, Carlos E.R. ha scritto:
>> On 03/05/2021 18.01, paul wrote:
>>> Carlos E.R. wrote on 03.05.2021 13:32
> Il 03/05/21 20:55, Carlos E.R. ha scritto:
>> On 03/05/2021 18.01, paul wrote:
>>> Carlos E.R. wrote on 03.05.2021 13:32
>>> I'm simply trying to ascertain the answer to those questions - for
>>> Linux.
>>> (Which is the most common consumer desktop Linux distribution?)
>>
>> That is also open to debate.
>>
>> As there are no sales figures, it is impossible to get actual numbers
>> of Linux usage. If you try to count "downloads", then you should
>> consider that a single download can be used for several installations,
>> even thousands.
>
>
> but also the other way around. I have downloaded much more distros than
> I actually use : I just gave them a try.
> I have not the palest idea of which phenomenon prevails.
Ah, yes, that too.
>>> Linux.
>>> (Which is the most common consumer desktop Linux distribution?)
>>
>> That is also open to debate.
>>
>> As there are no sales figures, it is impossible to get actual numbers
>> of Linux usage. If you try to count "downloads", then you should
>> consider that a single download can be used for several installations,
>> even thousands.
>
>
> but also the other way around. I have downloaded much more distros than
> I actually use : I just gave them a try.
> I have not the palest idea of which phenomenon prevails.
>
> also, I often look in DISTROWHATCH, but, apart from the fact that I have
> no clue if this site in particular is more reliable than many others I
> may well don't even know, the mentioned figures are always "per month"
> or so.
looking at how browsers to a site identify themselves. But this depends
on the popularity of the site, and at most it will only count computers
used for browsing; a user can have other installs and other computers.
...
>>> d- How does the consumer obtain those component parts (pull or push)?
>>
>> Polling and pulling. I'm not aware of anyone doing "pushing".
>
> <EVIL>
> maybe Snap store ... ;D
> </EVIL>
>
> BitBleach (a cleaning tool) looks for its own update at each run (maybe
> this can be disabled), even if installed from std repo.
--
Cheers, Carlos.
RonB
May 4, 2021, 6:28:57 PM5/4/21
to
On 2021-05-04, Aragorn <thor...@telenet.be> wrote:
> On 03.05.2021 at 07:20, Melzzzzz scribbled:
>
>> On 2021-05-03, paul <nos...@nospam.invalid> wrote:
>>
>> > Can someone who knows Linux explain how it is most updated by
>> > consumers? (You can pick the most common Linux consumer OS which I
>> > suspect is Ubuntu).
>>
>> Windows update is horror, mildly speaking. My Manjaro updates in 1-2
>> minutes, most time spent is downloading.
>
> 1 or 2 minutes? You must be running a very thin system then, or else
> you're talking of a minor update with only three or four packages.
Nope. Full install with all the codecs. I figured it was just because the
> On 03.05.2021 at 07:20, Melzzzzz scribbled:
>
>> On 2021-05-03, paul <nos...@nospam.invalid> wrote:
>>
>> > Can someone who knows Linux explain how it is most updated by
>> > consumers? (You can pick the most common Linux consumer OS which I
>> > suspect is Ubuntu).
>>
>> Windows update is horror, mildly speaking. My Manjaro updates in 1-2
>> minutes, most time spent is downloading.
>
> 1 or 2 minutes? You must be running a very thin system then, or else
> you're talking of a minor update with only three or four packages.
M.2 SSD is that fast. Installs on dual core machines, with a standard hard
drive, take about 20 minutes. So 1.5 minutes on a fast machine with a fast
thumb drive and an M.2 SSD doesn't seem out of the realm of possibility.
> I'm running the Stable branch of Manjaro, and the latest bundled update
> to Stable was on 28.04.2021. The downloading took about 10 minutes, and
> the actual update process took about half an hour, not including about
> 10 minutes for two AUR packages that needed to be rebuilt from sources.
> In the past, and with much larger updates, I've needed about an hour and
> a half, including the time needed for downloading the packages.
I should have made that clearer.) This was the initial install, not the
update that always follows immediately — for me at least.
> For clarity, I do run a fairly well-packed system here. I have 1982
> installed packages — my /usr currently stands at 7.2 GiB,
> and that's with zstd compression on a separate btrfs filesystem — and I
> did remove some of the junk that comes preinstalled on a new install,
> such as gvfs, Steam, AppImage support and FlatPak support — Snap was
> never even installed here.
installation and he said, "Done." I started asking him what went wrong — and
he said, "Nothing, it's installed."
Cörvid
May 7, 2021, 1:13:32 AM5/7/21
to
On 05/03/2021 04:20 AM, Peter Köhlmann wrote:
>
> In short: You are a babbling dimbulb
He's Canadian. We make those people answer a math-skills question before
>
> In short: You are a babbling dimbulb
they're eligible to win a prize.
RonB
May 9, 2021, 2:34:26 PM5/9/21
to
On 2021-05-09, Bud Frede <fr...@mouse-potato.com> wrote:
wife's Windows machines after a failed update. It's always required a
reinstall. In Linux (in 14 years of exclusive use) I've never had an update
fail. But dropping back to the previous kernel is easy if it ever happens.
> I don't think that I fully agree. Microsoft, Apple, and the various
> Linux distros all do things differently, and there are
> advantages and disadvantages for each.
>
> I've seen more issues with OS updates on Windows servers in the past 5
> years or so than I have on Linux or UNIX.
>
> On the desktop, which is kind of a pathological case due to the
> generally unskilled and unknowledgeable users, there are advantages to
> the way Windows and macOS do things. (I think macOS is better than Win
> 10 in this respect, as in most others.)
>
> While I do use Linux as a workstation or desktop OS, and my wife uses it
> for her desktop as well, I don't think that Linux is the desktop OS for
> everyone. It doesn't need to be, and it doesn't need to compete
> in that market.
>
> I also don't think that it's really useful to compare OSes like
> this.
>
> Does the desktop OS run the apps that I want to use? Can it run on
> hardware that I can afford? Does the way it works please me?
True.
> Those are all valid questions, and IMO more important than most of the
> back and forth I've seen in the newsgroups, on the wider net, years ago
> on FIDO and RIME and other networks, at various user groups, etc.
>
> I like Linux and I enjoy using it. It's comfortable to me and it doesn't
> get in my way.
>
> Work provides me with a Mac, so I use that too. It's also pretty
> comfortable and doesn't get in my way.
>
> Does MS have what I use beat on some feature or other? I don't really
> care, because I don't like to use Windows. I would be less happy if I
> were forced to use Windows, and that's that.
>
> If you prefer Windows, that's great. Carry on.
I feel the same. Use Windows if you like it or need it. I (personally) have
zero use for Windows or Windows-only apps. And Linux definitely works better
on my equipment than Windows would.
> Paul <nos...@needed.invalid> writes:
>
>> Melzzzzz wrote:
>>> On 2021-05-03, paul <nos...@nospam.invalid> wrote:
>>>> In the Windows newsgroup we were discussing how Windows differs from iOS and
>>>> Android in terms of (a) frequency of update (b) length of update (c)
>>>> component separation and (d) mandatory versus voluntary (see below).
>>>>
>>>> Can someone who knows Linux explain how it is most updated by consumers?
>>>> (You can pick the most common Linux consumer OS which I suspect is Ubuntu).
>>>>
>>>> See the context of the question in the discussion included in the sig.
>>> Windows update is horror, mildly speaking. My Manjaro updates in 1-2
>>> minutes, most time spent is downloading.
>>
>> It depends.
>>
>>
>
>> Melzzzzz wrote:
>>> On 2021-05-03, paul <nos...@nospam.invalid> wrote:
>>>> In the Windows newsgroup we were discussing how Windows differs from iOS and
>>>> Android in terms of (a) frequency of update (b) length of update (c)
>>>> component separation and (d) mandatory versus voluntary (see below).
>>>>
>>>> Can someone who knows Linux explain how it is most updated by consumers?
>>>> (You can pick the most common Linux consumer OS which I suspect is Ubuntu).
>>>>
>>>> See the context of the question in the discussion included in the sig.
>>> Windows update is horror, mildly speaking. My Manjaro updates in 1-2
>>> minutes, most time spent is downloading.
>>
>> It depends.
>>
>>
>> But on OS Upgrades, Microsoft has you beat. On Windows 10, if
>> the upgrade process goes tits up, the installer can revert to
>> original conditions. Without the user having to make a backup.
>> The process is now so good, it also includes protection
>> against the "reboot into new OS" failing. If the reboot into
>> new OS does not work, it reverts to the old OS. No Timeshift
>> or Clonezilla needed. It may be bloody slow reverting, but,
>> it works. It can take a whole hour to revert, as it unwinds itself.
I've (personally) never had the "reversion" work one, single time on my
>> the upgrade process goes tits up, the installer can revert to
>> original conditions. Without the user having to make a backup.
>> The process is now so good, it also includes protection
>> against the "reboot into new OS" failing. If the reboot into
>> new OS does not work, it reverts to the old OS. No Timeshift
>> or Clonezilla needed. It may be bloody slow reverting, but,
>> it works. It can take a whole hour to revert, as it unwinds itself.
wife's Windows machines after a failed update. It's always required a
reinstall. In Linux (in 14 years of exclusive use) I've never had an update
fail. But dropping back to the previous kernel is easy if it ever happens.
> I don't think that I fully agree. Microsoft, Apple, and the various
> Linux distros all do things differently, and there are
> advantages and disadvantages for each.
>
> I've seen more issues with OS updates on Windows servers in the past 5
> years or so than I have on Linux or UNIX.
>
> On the desktop, which is kind of a pathological case due to the
> generally unskilled and unknowledgeable users, there are advantages to
> the way Windows and macOS do things. (I think macOS is better than Win
> 10 in this respect, as in most others.)
>
> While I do use Linux as a workstation or desktop OS, and my wife uses it
> for her desktop as well, I don't think that Linux is the desktop OS for
> everyone. It doesn't need to be, and it doesn't need to compete
> in that market.
>
> I also don't think that it's really useful to compare OSes like
> this.
>
> Does the desktop OS run the apps that I want to use? Can it run on
> hardware that I can afford? Does the way it works please me?
True.
> Those are all valid questions, and IMO more important than most of the
> back and forth I've seen in the newsgroups, on the wider net, years ago
> on FIDO and RIME and other networks, at various user groups, etc.
>
> I like Linux and I enjoy using it. It's comfortable to me and it doesn't
> get in my way.
>
> Work provides me with a Mac, so I use that too. It's also pretty
> comfortable and doesn't get in my way.
>
> Does MS have what I use beat on some feature or other? I don't really
> care, because I don't like to use Windows. I would be less happy if I
> were forced to use Windows, and that's that.
>
> If you prefer Windows, that's great. Carry on.
I feel the same. Use Windows if you like it or need it. I (personally) have
zero use for Windows or Windows-only apps. And Linux definitely works better
on my equipment than Windows would.
David W. Hodgins
May 9, 2021, 3:46:43 PM5/9/21
to
On Sun, 09 May 2021 14:34:24 -0400, RonB <ronb02...@gmail.com> wrote:
> I've (personally) never had the "reversion" work one, single time on my
> wife's Windows machines after a failed update. It's always required a
> reinstall. In Linux (in 14 years of exclusive use) I've never had an update
> fail. But dropping back to the previous kernel is easy if it ever happens.
Same here. I don't use windows but have relatives that do.
> I've (personally) never had the "reversion" work one, single time on my
> wife's Windows machines after a failed update. It's always required a
> reinstall. In Linux (in 14 years of exclusive use) I've never had an update
> fail. But dropping back to the previous kernel is easy if it ever happens.
I've seen windows systems where it was running really slow. When I investigated
it, the system was repeatedly following the procedure of downloading a very large
update, trying to install it, failing, reverting the update, and going back to the
download step. All of this was going on in the background with the user unaware it
was happening over and over again. No notifications or other info provided to the
user.
As the person was on a metered internet connection it cost them over $50 in
overage usage. That was enough to convince that relative to switch to linux
only. :-)
Regards, Dave Hodgins
--
Change dwho...@nomail.afraid.org to davidw...@teksavvy.com for
email replies.
RonB
May 9, 2021, 5:47:14 PM5/9/21
to
Linux is the superior option. If you want a computer for a gaming machine or
some other specialized use, then that can change things.
Paul
May 9, 2021, 6:10:26 PM5/9/21
to
The one I remember, is a Visual Studio Runtime that kept updating
itself over and over. Rinse and repeat. It took forever to stab
that one with a silver spike and get rid of it. Even if you manually
updated it, it just would not stop. The one in the pipeline wanted
to deliver its load, no matter what.
As a rolling release, that's kinda sorta fixed. A number of the
looping behaviors were truncated at around three or four attempts
or so.
*******
There is a metered Internet setting, to reduce usage on marked
network connections. For example, the laptop might wait until
it's at Starbucks, to do some of the updating activity. There is
also a byte-o-meter, that shows what the "monthly usage" is by
the OS, so you can keep track if you want. There aren't too many
questions about that now, so I can't comment on the health/quality
of that feature.
It's not an OS "fit" for a limited cap networking service.
Using it with a 2GB/mo capped satellite service is a bad idea.
You'd have to drink too much Starbucks coffee to get all
the updating done.
Paul
Aragorn
May 9, 2021, 10:40:04 PM5/9/21
to
On 09.05.2021 at 09:13, Bud Frede scribbled:
> Paul <nos...@needed.invalid> writes:
>
> > Melzzzzz wrote:
>
> >> Windows update is horror, mildly speaking. My Manjaro updates in
> >> 1-2 minutes, most time spent is downloading.
> >
btrfs-enabled version of GRUB doesn't even need any time to revert to a
previous state. You simply boot up in the previous snapshot. ;)
> While I do use Linux as a workstation or desktop OS, and my wife uses
> it for her desktop as well, I don't think that Linux is the desktop
> OS for everyone. It doesn't need to be, and it doesn't need to compete
> in that market.
GNU/Linux was never meant to compete with anything, because it's not a
commercial product — let's not talk now about the commercial entities
like RedHat and Oracle that bundle their distribution with a paid
support contract.
> I also don't think that it's really useful to compare OSes like
> this.
Apple and Microsoft are both commercial businesses, and the people who
buy from either of these two businesses have a commercial (and thus
competitive) mindset. They cannot fathom any other ecosystem than the
commercial/competitive one.
> Does the desktop OS run the apps that I want to use? Can it run on
> hardware that I can afford? Does the way it works please me?
>
> Paul <nos...@needed.invalid> writes:
>
> > Melzzzzz wrote:
>
> >> Windows update is horror, mildly speaking. My Manjaro updates in
> >> 1-2 minutes, most time spent is downloading.
> >
> > But on OS Upgrades, Microsoft has you beat. On Windows 10, if
> > the upgrade process goes tits up, the installer can revert to
> > original conditions. Without the user having to make a backup.
> > The process is now so good, it also includes protection
> > against the "reboot into new OS" failing. If the reboot into
> > new OS does not work, it reverts to the old OS. No Timeshift
> > or Clonezilla needed. It may be bloody slow reverting, but,
> > it works. It can take a whole hour to revert, as it unwinds itself.
>
> > the upgrade process goes tits up, the installer can revert to
> > original conditions. Without the user having to make a backup.
> > The process is now so good, it also includes protection
> > against the "reboot into new OS" failing. If the reboot into
> > new OS does not work, it reverts to the old OS. No Timeshift
> > or Clonezilla needed. It may be bloody slow reverting, but,
> > it works. It can take a whole hour to revert, as it unwinds itself.
>
> I don't think that I fully agree. Microsoft, Apple, and the various
> Linux distros all do things differently, and there are
> advantages and disadvantages for each.
A GNU/Linux installation on a btrfs filesystem and using the
> Linux distros all do things differently, and there are
> advantages and disadvantages for each.
btrfs-enabled version of GRUB doesn't even need any time to revert to a
previous state. You simply boot up in the previous snapshot. ;)
> While I do use Linux as a workstation or desktop OS, and my wife uses
> it for her desktop as well, I don't think that Linux is the desktop
> OS for everyone. It doesn't need to be, and it doesn't need to compete
> in that market.
commercial product — let's not talk now about the commercial entities
like RedHat and Oracle that bundle their distribution with a paid
support contract.
> I also don't think that it's really useful to compare OSes like
> this.
buy from either of these two businesses have a commercial (and thus
competitive) mindset. They cannot fathom any other ecosystem than the
commercial/competitive one.
> Does the desktop OS run the apps that I want to use? Can it run on
> hardware that I can afford? Does the way it works please me?
>
> Those are all valid questions, and IMO more important than most of the
> back and forth I've seen in the newsgroups, on the wider net, years
> ago on FIDO and RIME and other networks, at various user groups, etc.
Exactly.
> back and forth I've seen in the newsgroups, on the wider net, years
> ago on FIDO and RIME and other networks, at various user groups, etc.
J.O. Aho
May 10, 2021, 4:26:26 AM5/10/21
to
On 10/05/2021 04.40, Aragorn wrote:
> Apple and Microsoft are both commercial businesses, and the people who
> buy from either of these two businesses have a commercial (and thus
> competitive) mindset. They cannot fathom any other ecosystem than the
> commercial/competitive one.
There is also the misconception that close source is more secure with
> Apple and Microsoft are both commercial businesses, and the people who
> buy from either of these two businesses have a commercial (and thus
> competitive) mindset. They cannot fathom any other ecosystem than the
> commercial/competitive one.
less vulnerabilities than open source and those the company management
may think it's better go for such even if the tech/devs would prefer
another solution.
--
//Aho
Aragorn
May 10, 2021, 4:49:05 AM5/10/21
to
On 10.05.2021 at 10:26, J.O. Aho scribbled:
And yet the annual Coverity report consistently keeps on proving the
opposite. But unfortunately, the ignorant pointy-haired bosses like
big and corporate sounding names. ;)
opposite. But unfortunately, the ignorant pointy-haired bosses like
big and corporate sounding names. ;)
J.O. Aho
May 10, 2021, 5:12:38 AM5/10/21
to
On 10/05/2021 10.49, Aragorn wrote:
> On 10.05.2021 at 10:26, J.O. Aho scribbled:
>
>> On 10/05/2021 04.40, Aragorn wrote:
>>
>>> Apple and Microsoft are both commercial businesses, and the people
>>> who buy from either of these two businesses have a commercial (and
>>> thus competitive) mindset. They cannot fathom any other ecosystem
>>> than the commercial/competitive one.
>>
>> There is also the misconception that close source is more secure with
>> less vulnerabilities than open source and those the company
>> management may think it's better go for such even if the tech/devs
>> would prefer another solution.
>
> And yet the annual Coverity report consistently keeps on proving the
> opposite.
And that is just what has been discovered outside the company, we never
> On 10.05.2021 at 10:26, J.O. Aho scribbled:
>
>> On 10/05/2021 04.40, Aragorn wrote:
>>
>>> Apple and Microsoft are both commercial businesses, and the people
>>> who buy from either of these two businesses have a commercial (and
>>> thus competitive) mindset. They cannot fathom any other ecosystem
>>> than the commercial/competitive one.
>>
>> There is also the misconception that close source is more secure with
>> less vulnerabilities than open source and those the company
>> management may think it's better go for such even if the tech/devs
>> would prefer another solution.
>
> And yet the annual Coverity report consistently keeps on proving the
> opposite.
know how many they have discovered themselves and how many of the
bugs/vulnerabilities has been fixed.
> But unfortunately, the ignorant pointy-haired bosses like
> big and corporate sounding names. ;)
have pulled off all their hair due of downtime and system breaches due
of unfixed bugs.
--
//Aho
Carlos E.R.
May 10, 2021, 7:24:08 AM5/10/21
to
On 10/05/2021 04.40, Aragorn wrote:
> On 09.05.2021 at 09:13, Bud Frede scribbled:
>
>> Paul <nos...@needed.invalid> writes:
>>
>>> Melzzzzz wrote:
>>
...
>
>> Paul <nos...@needed.invalid> writes:
>>
>>> Melzzzzz wrote:
>>
>> I also don't think that it's really useful to compare OSes like
>> this.
>
> Apple and Microsoft are both commercial businesses, and the people who
> buy from either of these two businesses have a commercial (and thus
> competitive) mindset. They cannot fathom any other ecosystem than the
> commercial/competitive one.
One can buy a commercial OS, and at the same time use Linux.
--
Cheers, Carlos.
Aragorn
May 10, 2021, 7:28:45 AM5/10/21
to
On 10.05.2021 at 13:20, Carlos E.R. scribbled:
Nitpicking... <grin>
Carlos E.R.
May 10, 2021, 8:32:08 AM5/10/21
to
--
Cheers, Carlos.
Steven Carroll - frelwizer
May 13, 2021, 11:02:49 PM5/13/21
to
Poaster. For all the boasting Onion Knight has done on this topic, the 'Social
Media Manager' does not know how to do this. It literally takes a couple
seconds to click and drag across a range and 'PDF convert' it. This forum
is a septic tank. By listening to 'social warriors' like that you get laws
like 'equal rights'. Carried to its (im)moral resolution, the insistence
that it's 'discrimination' for a hetero guy to not wish to marry a transvestite
is established. Just about everything Onion Knight says about what William
Poaster has done is false. Who isn't aware of this by now? HA!
-
This broke the Internet
<https://www.spokeo.com/Dustin-Cook/Tennessee/Kingsport/p40064865906>
https://duckduckgo.com/?q=Steve+Petruzzellis+the+racist+swine
https://duckduckgo.com/?q=dustin+cook+the+functionally+illiterate+fraud
Dustin Cook the functionally illiterate fraud
0 new messages