[ale] Kind words for Windows? - was The latest from Gigabyte

Michael B. Trausch mike at trausch.us
Tue Feb 8 12:16:48 EST 2011


On Tue, 2011-02-08 at 01:25 -0500, Ron Frazier wrote:
> Hi all,
> 
> The line I'm quoting from Chris Fowler (below) inspired this note. If 
> you're offended, blame me, not him. But, I hope no one will be
> offended, just spurred to thought. But, lots of it is just some random
> things I've been thinking about, and maybe ranting for frustration
> about some problems I've been trying to solve. I'm not trying to start
> a flame war, but, I've been using Windows ever since it was invented.

While I haven't used Windows 1.0 or 2.0 outside of checking it out years
and years after the fact, I go all the way back to MS-DOS 2.11.  My
first encounter with Windows was Windows 3.0, followed by 3.1, 3.11, WFW
3.11, Windows NT 3.1, Windows NT 3.5, Windows 95, and Windows NT 4.0.  I
have, since that point, run every version of Windows for short periods
of time for no other reason than to study the way the system works (and
the changes from the last version) in order to stay up on how it works.
While I may not use Windows for my everyday computing (in fact, I could
not imagine doing so at all; when I have to, I find so much base
functionality missing from the system that I'm unable to effectively
work.  Adding that missing functionality via Cygwin is pretty much
unbearable, given the high cost of process spawning in Windows and the
Unix philosophy that everything lives in multiple processes.).

> Modern versions of Windows are very slick, refined, functional, and,
> most of the time, reliable. In some ways, particularly in terms of
> user interface, it's more advanced than the Ubuntu I'm typing this on.
> In some cases, I can do maintenance procedures on Windows MUCH easier
> than I can in Linux, as I've detailed to some extent in other posts.

Define "more advanced".  For any definition I can think of, this simply
does not make sense.  The only thing I can think of that Windows
supports in its UI that GTK does not is MDI, and there is honestly no
need for an application to use MDI when tabs tend to work better
(especially in a proper implementation where you can pull tabs out of
the window they are in and move them to another window or create a new
window out of them---that's something that I have wanted in many
applications for a long time, but only Pidgin and Chrome support that
model of working as far as I am aware).

> I'm here, basically trying to use Linux 98% of the time, for 4
> reasons. They are, in no particular order:
> 
> 1) Intellectual challenge. I love to learn new things, most of the
> time, and maybe, there might be some usable job skills I'll gain.

That's good; I'd guess then that you spend much of your time learning.
However, it is possible to learn the ins and outs of everything from the
kernel up and then no longer is it an "intellectual challenge" to keep
up.  Once you truly understand a system---any system, and I mean far
more than just the surface or the UI it exposes---you will be able to do
things with it without thinking.

I aim to keep my knowledge as up-to-date as I can on a set of operating
systems, including the entire BSD family, various Linux distributions
(and more importantly the components that go into it), Android (can't be
lumped with normal Linux systems; it's that different) and the Windows
family itself.  I tend to find myself in situations that call for far
more intimate details of all of these operating systems than most people
care to learn, for whatever reason.

Windows is probably perhaps the best example.  The Windows operating
system is an insanely complex suite of interconnected DLLs and
executables.  There are hooks present at multiple levels to override and
redirect functionality.  Even the boot process, starting with Windows
Vista, is no longer straightforward like it once was.  I've used the
knowledge that I have gained from studying Windows to fix systems that
most people would give up and say "reinstall".  When reinstallation
isn't an option, people come to me.  I'm not cheap for those things, but
when you have your data on only a single system, and you need it back,
and you don't have a means for recovery...

And believe me when I say:  It is far easier to find and fix problems in
a Linux distribution than it is to find and fix them in Windows.  Having
the ability to access the source code makes a _huge_ difference.  I've
only actually had to refer to sources a couple of times to answer
questions that I've had about how things worked, but both times it has
been absolutely valuable.  If I had the money, I'd pay for a Windows
source license.  Why?  Because I'd be able to learn a lot more
effectively if I knew the true, underlying structure of things.  Even
with the restriction that the source could not be modified, just the
ability to read it would be, IMHO, quite valuable.  I'm not going to say
that I have learned all I can from a suite of black boxes, because there
are always methods to learn more (including disassembly).  But I don't
have the time to learn by disassembly.  That takes a lot longer.

(Though I will say this: reverse engineering, difficult as it is, can be
_quite_ fun.  But it is a very time-intensive task.  And, for me, it
requires the use of my dot-matrix printer...)

> 2) Windows cost, for a new PC, this is only about $ 30 - $ 50. If you 
> want to buy a copy for a build your own pc or do an upgrade, it's more
> expensive.

Indeed.  One of my biggest problems is the cost (both monetarily and in
terms of freedom).  Like anything, if the price of something is too high
it is going to be stolen more frequently.  There are some people that
will steal anyway, just to steal; I don't know what could be done to
alleviate that problem other than shooting kleptomaniacs in the head or
something.

But for the normal, every-day person, Windows is prohibitive in cost.
Those who don't know that there are other operating systems will be
driven to steal it.  Who can afford to drop $200 for a pretty dual-layer
DVD with a hologram on the front?  Most people on this list can, I
think.  But I know lots and lots of people that really cannot afford
that.  Add to that the fact that there are arbitrary restrictions on all
but the most expensive version of Windows, and you have a setup where I
think theft is encouraged---either you replace your computer every time
you need to upgrade Windows, and you get a Home edition and then burn
additional money to get an "anytime upgrade" to the edition that you
need (should the home edition not provide the functionality required),
or you buy/steal the next version and have to upgrade (though usually
it's easier and less a waste of time to reinstall cleanly).

For example, all of the features that can be used to really bolt down a
Windows system are not available on consumer editions.  That makes me so
angry I could scream.  There is literally no logical reason for this;
sure they want to screw people out of their hard-earned paychecks, but
requiring someone to give up a whole week's worth of pay just to have a
system that can be joined to a DC is nuts.  Irrational.  Driven by
greed, nothing more.

> 3) Maintenance issues. Tomorrow is the 2nd Tuesday, Microsoft patch
> day. So, I need to update the Windows side of my dual boot computers
> as well as family members' computers. Ubuntu is similar in the number
> of patches. I'm getting flagged by the updater every few days, and 
> sometimes there are quite a few patches. Linux is able to update
> without a reboot more often than Windows but it certainly is not
> always the case. Probably 1 out of 3 or 4 patch cycles asks me for a
> reboot.

It seems that Windows wants a reboot when nearly _anything_ changes.
There is, I think, a very good reason for this: Windows is so complex
that almost the only way to be guaranteed that the state of the system
is sane is to restart the system and let it come into a known state.
There are so many API calls at so many levels of abstraction that I find
it hard to believe that a group of any size could truly fathom all of
them.  Even updating a few DLLs, or installing an application, can
"require" a restart.

The reason that you don't find that on most other operating systems is
that you can update software while it's running.  The kernel on a
running Linux system requires restarted after an update (though there
are efforts to make that go away, as well).  If a security update for a
core library such as glibc is installed, the system should likely be
rebooted as well---this of course depends on the exact vulnerability
that was patched and whether or not a full restart is required.
Typically just bouncing whatever servers are running on a system is good
enough, because they will load the new glibc.

You can save a bit of memory by rebooting in the core-library-updated
case; applications that are using the old glibc and holding that image
open will continue to share it.  Newly started applications will start
using the updated glibc, which will have to be loaded in memory.  It
will be shared among all new processes.  A reboot can make that
increased memory usage go away because all processes will once again be
using the same glibc image.

This works because when glibc (or anything) is updated, the binary files
are unlinked and new files are put in their place.  On Linux and most
other operating systems (all of those that aim for POSIX compatibility
at the level of filesystem semantics, anyway) when a file is unlinked
(deleted), if it is the last link then it will be removed unless
processes are holding it open.  In that case, no new processes will be
allowed to open the file, and it will appear to disappear, but all
processes that have a open file descriptor to that file can continue to
use it.  IOW, after updating glibc, a reboot will allow all software to
use the same library and reclaim the disk space used up by the previous
version of glibc.

OTOH, restarts are required in Windows in part because of its (IMHO,
pretty stupid) rules.  You cannot delete an open file.  WTF.

An aside: The semantics of file unlinking in POSIX systems is often
_extremely_ useful.  It permits one to do something like open a
temporary file and then immediately delete it.  This makes it more
difficult (but not impossible!) for other processes to get at the
contents of that file.  This trick is often-used in the programming of
applications on Unix-family systems.  I'm not aware of an analog in
Windows.  One could possibly consider using ADSes in a file for a
similar effect, but I think that is asking for trouble.

> The package manager in Ubuntu is what I think is a great advantage
> over Windows. In Windows, I have to update (separately) the system,
> non MS AV system, Java, Firefox, Firefox addons, Adobe Reader, Adobe
> Flash, Itunes, and basically anything else that connects to the
> Internet. It's great to know that everything I've installed using the
> package manager in Ubuntu will be updated.

That is a very good reason to not go behind the package manager's back
unless there is a very good reason to do so.  Even then, instead of
going behind the package manager's back, you can build both source and
binary packages to install software, and you can keep a local repository
such that you can update your local repo (and all of your local systems
will then be able to update efficiently).  There are some gotchas and
you have to know how to use a lot of the functionality of the package
manager (especially when you build binary packages without source code)
but it can be done.  As with anything, once you learn it, it is not
difficult.

> However, this doesn't apply to things I've installed separately. So,
> after I'm done updating Windows, I'll do a few housekeeping things on
> Ubuntu (on 3 computers) including: force a system update via the
> Updater tool, update the Firefox addons, check the Firefox options
> configuration in case something got changed, check Flash rev level and
> configuration, and check the Java rev level. The versions of these in
> the Ubuntu repositories sometimes lag behind the manufacturers'
> versions, but I don't want to override the install manually since
> Synaptic may no longer handle the updates. Usually, they're not too
> far behind though. There are a few apps which I couldn't install from
> the repositories, and I may have to update them manually.

I am looking forward to more people doing what Google has done with
Chrome: make it available such that updates can come via the package
manager without any extra effort at all.  When you install Chrome on
Ubuntu, they enable the APT repository for Chrome and thus you get
updates transparently, just as you do from Ubuntu for all of the
software that is included.  (I don't actually use Chrome on Ubuntu,
though; I use Chromium because they have more frequent updates for
testing and I like being on the bleeding/slightly scabby edge of
things).

> Whether Windows or Linux, I try to do periodic disk integrity checks
> and backups. With 3 dual boot PC's of my own, plus my Son's and Dad's 
> Windows machines, this all adds up to a lot of work.

Backups are absolutely the #1 thing anyone should do, of course.

However, hard disks do not require any active management these days,
really.  I mean you have to periodically defrag NTFS for performance
reasons, but that has nothing to do with the blocks on the drive itself.
Even doing what you're doing with checking integrity, what does that
really accomplish?  Not much---in fact, you're not even likely to notice
the types of problems that you're looking for when you do that.

If you are concerned about integrity, you should be using zfs or btrfs,
both of which can detect what is known as "silent corruption".  If you
are using the redundancy features built into either filesystem, that
"silent corruption" will be repaired and the block on the disk that was
corrupt will be marked as bad.  This is a very effective means of
handling the problem.  The way that *I* would handle it is to push all
of the important data to a server running a filesystem that can handle
silent corruption and mount that filesystem over the network.  Don't put
anything on your workstations -- especially laptops -- that you're not
prepared to lose.

A side note, though: I wouldn't use btrfs.  Not just yet.  Something
they did do it somewhat recently made it slow-slow-slow-slow.  I'm going
to pull a newer kernel at some point and see if that's all been fixed,
but at least as of the kernel in Ubuntu 10.10, btrfs is almost less than
useless.  It won't eat your data, but it'll sure make you wait for it.
At least presently.

> 4) Security! This is the big one. This is one of the main reasons I'm 
> now mostly a Linux user. There are just too many security risks with 
> Windows.

There is a very logical reason for this:  Microsoft Windows is a
multiple gigabyte collection of binary software, all managed by a single
central authority.  I don't know how many gigabytes of source code there
are, but the binary installation for Windows 7 is something like 12 GB.
Probably half of that is data files.  The rest would thus be executable
code at one layer or another in the system.

Windows also attempts to do everything for the user, and in doing so it
makes security sacrifices for the sake of convenience.  This is a real
problem.  Even with Windows 7 systems, I _still_ see applications that
want administrator privileges to do simple things.  Applications that
require this "elevation", as Microsoft calls it, are still far too many.
I don't quite understand it; I very rarely need to actually have my
system perform things as the superuser.  I do almost everything as my
own individual user.  This is perhaps the biggest reason I cannot use
Windows for my daily work: security, and the lack of ability to
successfully audit it.

While on the topic of security, I have this to say.  There is an
often-quoted expression: "Knowledge is power."  Indeed it is, but
knowledge is also security.  A system could be the most secure system in
the known universe, but its status as such is absolutely worthless if it
cannot be known.  I certainly am not going to take a vendor's claims of
security on faith.  I've seen too much proprietary source code to ever
trust anyone on faith when they much such claims.

Probably the largest reason that Windows is still as insecure as it is
is backwards-compatibility.  With a massive change in its architecture
(which can only come at the cost of backwards-compatibility) it would be
possible to make the system significantly more secure than it is.
However, they have a very real economic disincentive to do that:
breaking backwards-compatibility costs time and money.  However, many of
the decisions that are still causing security problems today originated
back in the days of Windows NT.

> Now, in fairness, if Linux had a billion users, and if the 
> hackers had a financial incentive to hack it, you'd see a lot more
> Linux viruses.

First point: hackers are the people who write code; e.g., "Linux
hackers" are people that work on the Linux kernel.  I believe you
intended to refer to crackers --- those are the people who crack
software with the goal to obtain access to it in some way which is
disallowed.

Yes, there are means by which people can exploit software on Linux
systems.  No system is 100% secure; I don't believe that there will ever
be such a system.  Nevermind the fact that the idea of "absolute
security" is unattainable.  I take that back: absolute security is
perfectly attainable.  Eject a computer box out into space such that it
floats away and nobody can touch it, give it no radio, and it will be
perfectly secure.  However, it will also be perfectly useless.

I see this "if Linux had more users" argument all the time.  I'd love
for someone to actually submit some sort of proof; people have attempted
to disseminate Linux-based viruses.  They do not thrive in the wild.
Viruses have to replicate, and in the normal use case (no matter if
we're talking desktop or server here) with a Linux box that is properly
configured (and truth be told, most of that "proper configuration" is
"don't run as root, you idiot") it is next to impossible for a virus to
replicate.  Where's it going to go then?

Also, take note of the mechanisms that viruses use to function in
Windows.  Most of them simply do not exist in Linux (or in any other
operating system save for perhaps VMS and a few other operating systems
that I've not been introduced to).  There are hooks everywhere;
literally, there are something like 10 different ways to install hooks,
and the list of hooks that are available seem endless.  Such hooks do
not exist on Linux systems.  There is LD_PRELOAD, but that is a
significantly more difficult system to abuse, again owing to the
security model that is enforced by Unix-family systems.  (And some
Unix-family systems do not even have LD_PRELOAD functionality made
available by their dynamic linkers---no surprise, for security reasons.)

> Yes, I know we can debate endlessly about which system is more secure,
> and Linux certainly has it's merits. A well patched and protected
> Windows system is one of the most secure platforms, akin (I have
> read), to a condo in the city with bars on the windows (no pun
> intended).

I think if you were truly familiar with the architecture of Windows at
all, you would laugh pretty hard at that statement.  However, there is
no way that anyone can convey that knowledge to you---not Steve Gibson,
not myself, not anyone on this or any other mailing list really.  Spend
a few years learning how the components of Windows work and interact,
and you will start to get an idea.

> Mac is like a normal house in a suburb, and Linux is like the
> farmhouse 20 miles in the woods. Linux is safer, to some extent,
> because it's not as big a target.

The design of the security model an operating system uses is what
more-or-less decides how secure it is.

You've heard about the Android virus that has been going around?  If you
take a look at the security model of Android, it's somewhere between
what Windows is and what the normal GNU/Linux distribution is.
(Remember that Google rewrote a significant part of the stack; Android
is its own operating system and is distinct from GNU/Linux systems.)
Android does not leverage the kernel as much as it could, from my
understanding of it, to enforce security.  Dalvik handles a large part
of that.

> However, no OS is immune to attack, particularly if the user clicks on
> a malicious binary file as a result of a phishing attack. A poorly
> patched or poorly configured Linux system can certainly be vulnerable.

A universal truth: any system that is not maintained becomes a
liability.  The difference is the amount of maintenance required and the
turnaround time on the issues.  I've never seen a critical security
issue go 18 months after discovery without a patch in a free software
operating system unless the project itself was dead.  Even then, not 18
months.  While 18 months does not appear to be usual for Microsoft, it's
not unheard of---there was an IGMP vulnerability that discovered and
went 18 months before Microsoft released a patch for it.  And that was a
privilege escalation exploit, if I recall correctly (and something that
neither NAT nor a stateful firewall could protect against in many
circumstances).

That brings back the point that workstation, server security are
absolutely critical to the security of any network.  But, I digress.

> But, nevertheless, the lack of security problems is one big reason I'm
> running Linux. At the moment, I'm not running anti virus on Linux, but
> I am running the firewall using Firestarter.

Why?  Are you running services on the system?  If you have no services
listening on your ethernet interface's IP addresses, you have no need
for a firewall to be running at all and are simply wasting CPU cycles
and memory.  Not a significant amount, mind, but still.

Properly applying security requires, to a degree, a knowledge of when,
what, why and how.  Blind application might not do any harm, but it's
certainly wasteful.

> At some time in the future, I may find it necessary to run anti virus.

Again, why?  The only Linux boxes I run AV on are those that handle mail
for the networks they're attached to, or file servers.  Those are the
only systems that need it.

> Whether on Linux or Windows, I run the NoScript addon for Firefox,
> which disallows all scripting except for sites I explicitly trust. I
> think that's only prudent. Combine that with Xmarks, and you can save
> the NoScript configuration in a bookmark and make it common across all
> your computers, Windows and Linux. That works really well.

I'd just steer clear from sites that I don't trust, but that's me.  Call
me crazy.  (Hell, maybe I am; I believe that if you want to avoid a
situation, you make it provably impossible to occur.)

> -1) Here's NOT a reason I'm using Linux - ease of use.

And I respectfully submit that you don't know what "ease of use" is.
Like most people you are confusing "ease of use" with "ease of
learning".  Guess what?  The combination of an accelerator and brake
pedal in the usual arrangement, accompanied by a clutch and a shifter,
an ignition switch, and a steering wheel seem perfectly easy to use for
us.  Why?  Because we're familiar with it and we already learned how to
use it.

But nobody will teach you how to use your operating system like they
will teach you the subtleties of operating a motor vehicle.

Study after study shows that a user's first operating environment
defines what the user calls "easy to use", because what they really mean
to say is "like the system I come from".  But, if you are equally
well-versed on both systems, both are equally easy to use.  If you're
familiar with the UNIX command line, you'd define most systems in the
UNIX family as "easy to use", while people would scoff at you.

It really pisses me off when people claim "ease of use", though; it
conveys the idea that somehow Ubuntu or other operating systems should
mimick Windows.  If that's what you want, check out ReactOS.  Otherwise,
the "fix" for this problem is to study and learn the system.  No
operating system is intuitive out of the box---everyone requires some
sort of education or training to get started working with one,
regardless of which one it is.

>  (This echos the sentiment in a recent letter to the editor in Linux
> Journal.) Frankly, most of the time it's similar to Windows, sometimes
> it's worse. I wouldn't recommend it to my Dad, unless I was still
> around to do setup, maintenance, and troubleshooting. Ubuntu 10.04,
> which I'm running, is the first Linux I've tried which just worked on
> the hardware I have.

My parents have two computers.  One runs Ubuntu, and the other one runs
Windows.  They use both.  They even manage to keep both systems
up-to-date, most of the time.  However, the only one that constantly
requires any work when I visit is the Windows one.  Go figure.

No, they're not "computer literate" (I hate that phrase, too; it's
completely nonsensical).

> (Tried previous Ubuntu. Haven't tried other distros.) Even then, I had
> to load some proprietary drivers, but I'm OK with that. On a day to
> day basis, Gnome is fine, and it does what I need. However, there have
> been a number of times, witnessed by my calls for help on this forum,
> when Ubuntu has driven me up the wall and crazy trying to do some
> maintenance procedure or configuration.

As I have said before and will say again: learning the system at the
appropriate level of abstraction will help make a lot of that go away.

Sometimes, software authors do unexpected things.  But in the normal
case (with most installed-by-default software, or preferred packages,
anyway) there is the basic philosophy that everything should do one
thing, and do it well.  It simplifies things greatly by
compartmentalization.

Not everything can be found or fixed or solved by that method; but you'd
be surprised how many things can.  A little bit of knowledge can be a
dangerous thing, but the more you go and the more you learn about how
all the expected components of a system work, the more you understand
how it all works, the more you can go in and find what's up, disable it,
work around it, or fix it.  As is usually the case, the operating
system's core tools (and some really nifty other ones, like htop) are
your friend, and help to identify problems.  The /proc filesystem can be
useful for that, as well.

> At those times, I have to go scrounging through Google, forums, man
> pages, config files, and command lines. Now, I can edit config files
> and scrounge for info with the best if I have to, and I've done plenty
> of that in Windows over the last 26 years. But, I believe that should
> be a rare occurrence, not a frequent one.

Personal opinion, of course.

And one that I (to a certain extent) share.  Fortunately, things like
configuring printers or mounting remote filesystems over SSH no longer
require working at the command line.  However, complex things do.  Many
of those complex things have no analog in Windows, or are things that
can be solved by 50 characters in a configuration file in UNIX or 150
clicks in Windows.  There is an advantage to not leaving the keyboard:
speed.

I _do_ think that there would be benefit in a distribution where the
"core" was defined to be a bit bigger than most distributions define it,
where tight integration of components (and central configuration of them
without losing anything) was the rule of the land.  However, the reason
that such a thing does not truly exist yet is because it's an extremely
difficult problem to solve.  There are tasks that I know how to do in
both the average Linux distribution and in Windows, and believe me, I do
them in Linux quicker, because I usually don't have to leave the
keyboard.

There's a fundamental difference between the two operating systems, and
unless you truly get both methods of operation, you'll be lost.  UNIX
keeps everything at your fingertips, ready for use, all the time, and in
an infinite combination of possibilities.  Windows, OTOH, is mostly
designed to make frequent tasks "easy" or quick, and in nearly all cases
you can only do with the system that which was explicitly designed to be
done with it (in terms of configuration and various other things).

The way UNIX works, where the GUI programs are front-ends for the
ever-more-powerful command line backend, works really well.  If that
would be raised up to the next level of abstraction, it'd be even
better.  I have spent a significant amount of time thinking about such a
system (nothing better to do when I'm having trouble going to sleep at
night), but as is normally the case for any of my ideas, I haven't the
time to do anything about it.  It's something I'd love to see, and if I
ever get the time, love to do.

> Here are some examples of comparative experiences with Windows and 
> Linux. In general, I want to keep my Windows installs and Linux
> installs functionally equivalent. In every case, the experience was
> harder with Linux, and in some cases, I have no solution. Now, my
> intent is not to rag on Linux, but a lot of people like to rag on
> Windows. Both are tools, they have pros and cons. These are just
> examples, and I'm sure I'll ask for help in other posts, so that's not
> my purpose here.

"Ragging on" any system is completely unproductive (and frankly,
irrational).  I have a laundry list of things that I strongly dislike
about Windows, of course, including its horrible rigidity at its various
command shells (and PowerShell---ugh, they took Unix commands and
completely neutered them).  Now that I think of it, they made Unix
commands eunuchs on Windows, lol...

Sorry, couldn't resist.  :)

> * Plugging external monitor into laptop screws up screen resolution
> and scrambles my icons. Windows - does just fine Linux - problem
> exists, no solution yet - reads monitor capabilities wrong

Intel chipset?

> * Desire to have fine grained control over shutdown sequence on
> battery power when battery is low.
> Windows - built into OS, set up parameters in minutes via GUI
> Linux - have to install UPS control daemon - options thus far seem
> very non obvious - few GUI's to set them up

Not sure I understand the problem.  My laptop does precisely the right
thing come critical power-down time.

> * Desired to run all my email on the same system. Now using Eudora
> OSE.  Windows - setup was easy and quick - double click the EXE -
> found my old email database (from Eudora Classic) and imported - (That
> WASN'T quick - 12 hours.)
>
> Linux - setup more difficult - required running a script to install - 
> not available in repository - didn't find old database and gave no 
> choice for manual selection - program must be run from a terminal
> window (no Double click) - I had to figure out how to put a launcher
> icon on the desktop (should be automatic) - terminal remains open all
> the time for whatever reason

Sounds like a problem at the application level.  What does this have to
do with GNU/Linux or even just the Linux kernel in particular?  I'm not
sure I get your gripe here.

Then again, IMHO, all MUAs suck.  Some just suck less than others.

> * Desire to be able to clone my hard drive as a backup, then just 
> install it if the main HDD dies.
> Windows - I have a nice, sophisticated, GUI based backup tool (not
> free) - Acronis TrueImage - I have no problem paying for good quality
> software if I have to. People deserve to be compensated for their
> work. Of course, if people choose to put forth a good quality free
> tool, like this email program, I'll use it.
> Linux - haven't found a comparable Linux tool yet - used TrueImage to 
> clone the drive and install it for testing - Windows boots fine -
> Linux fails to find it's swap partition - I resort to a swap file -
> GRUB freaks out the next time it updates and demands that I tell it
> where to put itself (I'll tell it - @#*&%*! Bang! Zoom!)

O.o.

dd.  Built-in.

If you require the ability to better compress the filesystem, fill it
with zeros first (also using dd) then pipe the fs image to gzip for
storage.

As I've mentioned in the thread where you were discussing all of this:
filesystem utilities do not make any guarantees about what they do
outside of filesystem preservation.  And no FS utilities I know of for
Windows will do anything with a Linux swap partition.  You probably have
to mkswap it.

That means that filesystem utilities don't guarantee that any operating
system will boot.  You still have to make sure that the bootloader(s)
involved are happy.  That's a pretty complex problem.  And I guarantee
you that Acronis only handles it for the Windows bootloader(s),
depending on which ones it supports (there are a few now; there didn't
used to be).

However, if you know what you're doing at that level, reinstalling
bootloaders is pretty easy (especially since the configuration is
already on the boot fs).

> * Desire to have my HotSpotVPN account active, so my public browsing
> at hotspots is encrypted.
> Windows - Relatively painless. Required running two install programs. 
> Boom. Up and running.
> Linux - haven't succeeded or really even tried yet - read the
> procedure once - gave me a headache - not automatic - many steps -
> have to run from terminal

Don't use hotspots?

Or just use SSH to provide SOCKS or a quick-n-dirty tunnel?

The further we go, it seems that you operate like a member of management
in a company: dictate extremely precise requirements to the exclusion of
considering easy and simple solutions that are within an arm's reach.

> * Possibly want to TrueCrypt the drive to make it look like gibberish
> to thieves and nosy govt agencies.
> Windows - I hear it's pretty easy.
> Linux - don't even want to think about it, but I don't know what's
> involved.

Or just use LUKS.  It's pretty simple and supported at install time, and
can be used to encrypt the whole disk.  Windows can't even do that
without having a very new Windows and a TPM chip.  We're getting to the
point of "you very obviously didn't even bother to search google".

> Oh, here are a few more. Some of these I may be able to run under
> Wine, some not.
> 
> * Audible books - Linux cannot do it

DRM?  Pester the companies that have cock-blocked Linux.  Otherwise,
nothing can be done; if you want to wear chains, you have to accept the
price of them.  No different than car vs. plane:  If you want to fly
somewhere, then accept the cost and let the digits up your butt.
Otherwise, do what I do: get in a flipping car and go.

> * Sansa Fuze+ MP3 Player - updater and media converter - Linux cannot
> do it

"Linux" doesn't do MP3 players.  Have you tried Banshee?  Do you know
the first thing about which of the eleventyone standards this particular
player uses to communicate with the computer?

> * TomTom GPS - updater - Linux cannot do it

I'm sure I've seen people update TomTom devices without Windows.  I
don't have one, but I am almost positive I've seen it done.  In any
case, you're right.  Linux cannot do it, because an _application_ has to
do it.  I'd hope that Linux doesn't do that in the kernel.  That'd be
awful and useless bloat and it would piss me off.

> * Evernote - online notes system - Linux cannot do it - OLD one works
> in Wine

Tomboy.  Even syncs with Ubuntu One.  zim desktop wiki, another nifty
solution for that sort of thing.

> * MetaTrader 4 - currency trading software - Linux cannot do it -
> works in Wine

PayPal.  They support this nifty standard you might have heard of that
works on all platforms: HTTPS.

> You get the idea.

Indeed.

> Now, I intend to stick with Linux, and I intend to get these problems
> solved. Since I'm a geeky kind of guy, and I'm tenacious, AND since I
> have lots of GREAT friends on this forum to run to for help, I can
> probably get it done. However, the average Joe on the street won't be
> so able or willing to work through these problems. He'll just use what
> he knows works, and what he knows he can call several people for help,
> and what's easy and available - Windows or Mac.

Do you file bug reports?  Do you follow up on them?  Because honestly
that is what more people need to do.  I know that I have overlooked bugs
before and forgot to report them, or reported them and not followed up
on them.

For example, right now.  I'm having trouble using my computer.  It seems
that composing very long emails in Evolution causes it to chew up
memory.  Thankfully I have some swap; I'm 274 MB deep into it at the
moment.  Evolution is using 3.7 GB of RAM.  I'm going to have to file
that bug like as soon as I restart Evo.

> So, I hope this hasn't offended anyone. I also hope that Windows and
> Mac will draw on the strengths that Linux has AND I hope that Linux
> will draw on the strengths that Windows and Mac have, and that all the
> systems will improve. Of course, I'd like to see more people using
> open source software.

The only thing I see is that you attribute things to "Linux" that should
be attributed to the applications.

I'm certainly not going to blame "Linux" for my problem with Evolution.
But I can't write anymore... so I'm sending this and then killing Evo
before it brtem to a complete halt.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part
Url : http://mail.ale.org/pipermail/ale/attachments/20110208/469b23ae/attachment-0001.bin 


More information about the Ale mailing list