Re: [Pan-users] To explain why I'm unwilling to up date my glib/gtk+/etc libs ATM…

[Duncan]

SciFi posted on Mon, 04 Jul 2011 20:16:07 +0000 as excerpted:

> Essentially, I've been wanting to "jump ship" for a very long time
> (years).

Interesting.  I knew you as the Mac guy who knew code, while I only knew 
some things /about/ code, at the C/C++ level anyway.  And of course I 
knew your very nice xface, especially since that discussion and patch 
sometime back that corrected pan's previous face color-inversion. =:^)

But I didn't know about your wish to jump-ship.  That it is taking you so 
long reminds me of my own journey, back around the turn of the century.  
After running the IE/OE4 betas (really where I discovered USENET, with my 
first real experience on the IE/OE4 beta newsgroups, tho I'd read about 
it before) and having fun with the new desktop extensions they included, 
I was in line at midnight for Windows 98, but that was about the zenith 
of my time with MS, and the afternoon went rather faster than the morning 
had.  Shortly after that (some time in '99) I began looking into Linux, 
but it took me two years to actually be ready to make the switch.

In I think early '99 or so I bought a Mandrake (5.1/Venice, perhaps, 
going from the table in the Mandriva wikipedia entry?) at I think Fry's 
Electronics and loaded it.  It worked, but I didn't know enough of what I 
was doing to be comfortable with it, and I've always hated rebooting, so 
it pretty much just sat there.  IIRC, I upgraded mobo and CPU and the 
version I had quit working, so I eventually upgraded to I believe
6.1/Helios, probably in early to mid 2000.  But still it sat mostly 
unused.

However, by sometime in 1999 I had started confirming Linux drivers for 
all the hardware I bought, and by 2000, was carefully considering 
software purchases as I thought about switching.  Around late 2000, as 
the betas of Office XP developed eXPrivacy anti-features, it was becoming 
very apparent that MS was following a different road than I and I began 
getting more serious, asking questions on the newsgroups, etc.  Sounds 
like you're at about that stage ATM.  With the release of Office eXPrivacy 
and the betas of Windows eXPrivacy developing the same anti-features, in 
early 2001, I began actively planning the switch.

Pretty much as soon as I did, I realized the problem, that I didn't know 
enough about Linux to really /do/ anything with it.  Years earlier, one 
of my best early computer decisions (after I had my own, anyway) had been 
buying the book "Using DOS 6.2", which taught me an ENORMOUS amount about 
computer basics.  But I had bought a number of other books after that, 
both in the Using... series and not, and been VERY disappointed, as they 
really weren't all that useful at all (about as useful as the Windows 
Help system, which I had taken to calling NoHelp, because that's what it 
so often was) and had been a waste of money.

So what I realized I needed to teach me Linux was a book as useful as 
Using DOS 6.2 had been, but I didn't have a lot of money or time (the 
eXPrivacy release was fast approaching) to go experimenting and sorting 
thru all the dreck, first.  Without that, I doubted I'd get very far in 
Linux, as I simply didn't know enough about what I was doing to feel 
comfortable or useful in it.  But MS wasn't leaving me much choice, I was 
running out of time, and by this point I was getting desperate.

So I asked in the newsgroup for my ISP, which at the time had a VERY nice 
sized group of Unix/Linux/BSD folks (the ISP turned bad not long after, 
unfortunately, and people were leaving by the time I did a year and a 
half later), including one with direct commit rights to one of the BSDs 
(IDR which, was pretty fuzzy to me at that point), himself.

Two books were recommended by more than one person replying, Running 
Linux, and Linux in a Nutshell, both from O'Reilly publishing.  Running 
Linux is in a tutorial/textbook style with simple stuff at the front and 
more complex toward the back.  Linux in a Nutshell is, as described in a 
blurb on that edition's back cover, "Like a stack of manpages but simpler 
to read and easier to carry around", or something to that effect.  Both 
of them cover mostly the command-line, tho there were chapters on 
configuring X and on KDE and Gnome in Running.  And they both cover the 
generic Linux tools, save for the section on package managers, not some 
distro-specific stuff.  So it's a good foundation.

They were both in their third edition at the time, and I bought them 
both, reading the Running book nearly cover to cover.  Thinking to ask 
and get those recommendations, then buying those books, was perhaps the 
best computer related purchase I ever made, and I still recommend them, I 
think in sixth edition last I checked, today.  I even bought a newer 
version of Linux in a Nutshell, after using the first until it was quite 
worn out, and continued to use parts of the second (especially the bash 
appendix) until it disappeared some months ago.  I'm sure they saved me 
at LEAST three months of full-time-equivalent work coming up to speed on 
Linux!

You're advanced enough and know FLOSS well enough they won't be the help 
to you that they were to me, but I don't know if you're really 
comfortable on the *ix command line or not.  If not, I'd certainly 
consider buying Linux in a Nutshell, at least, as it should help.  
Running Windows I'd recommend to folks who have no clue about Linux file 
permissions (user/group/world, read/write/execute, setuid/setgid/sticky-
bit), etc, but I expect you're beyond that.

I installed from the Mandrake 8.1 ISOs I had just downloaded and burned 
(no 8.1 disks to buy yet, only the ISOs to download), the same week 
Windows eXPrivacy was publicly released.  Even with those books i took 
three months to really get operational in Linux and quit having to boot 
to Windows for this or that, because I was a Windows poweruser and at 
least VB programmer, and I had certain expectations for my Linux install 
that were either going to be met, or I was going to know the reason why 
they couldn't be.

Among those expectations, I had a triple-monitor setup that worked fine 
in W98, and I expected it to work fine in Linux as well.  Fortunately or 
unfortunately, back when I had been verifying Linux drivers for any 
hardware upgrades, I had purchased an nVidia card, not realizing the 
difference between Linux drivers and freedomware Linux drivers.  I 
learned that pretty fast when I switched!  I ended up learning how to 
hand-edit my xorg.conf (then xf86config) as the automated tools couldn't 
deal with the multi-monitors.  I learned how to download, configure, 
build and install my own kernel, and how to rebuild the nVidia driver to 
match it, because only the nVidia proprietary driver could handle the two 
outputs I was using on the card.  I didn't have the money for a new video 
card at that point, and I suffered with what I had for some time, but you 
can bet that when I DID upgrade video card again, I made **SURE** I chose 
a card with reasonable freedomware drivers, a Radeon as it happened, and 
I'm still buying Radeon's today (tho there were a few years when I didn't 
upgrade video at all, as there simply weren't any decent upgrades with 
good freedomware drivers), for the same reason.

Another expectation was that I could configure the boot manager (LILO, I 
eventually switched to GRUB but LILO was the standard x86 boot loader 
back then) to boot to either Windows, which I had left alone on its own 
drive, or Linux, installed to a brand new, much bigger drive.  That 
required learning how to hand-configure LILO, and not just the simple 
stuff, either, but some more fancy stuff to switch the drives around so 
Windows would think it was on the primary drive.

It took me about two months to get the hardware level all working the way 
I wanted, before I even STARTED looking at the app alternatives I 
wanted.  However by the time I did, KDE was already growing on me, and I 
settled on it, with Konqueror for the browser, and kmail for the mail 
client.  The last major app I needed was the news client.  I eventually 
settled on pan, but it took a bit to find it.  The kde news client, knode, 
just didn't do what I wanted/needed.  By this time, another month had 
passed, so it was about three months before I got situated in Linux, but 
I had quite the learning curve as I had learned to manually configure and 
build the kernel, to manually configure X, and to manually configure 
LILO, in that period, to get the system up and running the way I wanted.  

During that three months, I was rebooting to Windows to do stuff I hadn't 
figure out my Linux alternative for, yet.  The first two months, it was 
pretty much everything software, since I was dealing with hardware 
issues.  Then as I chose browser and mail, I switched them to Linux, but 
I was still having to boot to Windows to do news, and to ask my ISP's 
newsgroup people about questions I had, and there were quite a few.  (I 
had trouble with the kernel config at one point, I recall, because the 
config tool I was using, a Mandrake or KDE took IIRC, wasn't saving the 
config properly or some such.  I ended up using make menuconfig, the same 
thing I still use today.)

But, by the end of the three months, when I'd finished converting to pan 
for news, I was not only as comfortable in Linux as in Windows, the 
earlier "what do I actually do on Linux" thing was already inverting on 
me, and I'd boot to Windows to get something ready to delete or to ask a 
question on the newsgroup, then just sit there after doing that specific 
task, wondering what to actually /do/ on Windows, now that I had booted 
it and finished the specific task I had setout to do!  It was at THAT 
point that I realized my switch was a done deal; there wasn't any turning 
back now!

Ironically, a couple months later, I read some post from somebody calling 
themselves a Red Hat Linux developer, who, it was quite evident, had 
never even built his own kernel!  Bah, humbug, I remember thinking to 
myself! Even *I* knew how to do that!  I'd been on the platform less than 
six months and wouldn't DREAM of calling myself a developer if I hadn't 
ever even compiled a KERNEL!  Oh, well...

> In fact, the people I've conversed with, have all agreed with
> me.  But I thought I would pay the yearly fee to join ADC (Apple
> Development Center, http://developer.apple.com/) to see what I could do
> to help.
> 
> (In case the reader does not know, I am medically disabled / retired,
> with 38+years work on record.  I have "professional" experience in the
> system-level programming for a big huge shop which uses machines made by
> a famous three-letter international company.  I do this open-source
> stuff because I believe in it whole-heartedly.  Plus, in my "twilight
> years", it helps keep my mind sharp.)

I didn't know that much about it, but had known you as a decent coder... 
What had me sort of puzzled was your interest in pan and FLOSS in 
general, being on a Mac and for all I knew satisfied with it in general.  
Now I'm finding out more of the story.

> I only have one single machine here, which also is used for making tons
> of OTA TV recordings.  I don't get much down-time to boot-up any other
> system.

I didn't really have /that/ problem.  I just hated the down-time required 
for rebooting, plus I didn't know what to /do/ in Linux, a problem those 
books and the specific hardware config tasks I set out to do when I first 
switched solved very well for me.  But I understand not really having 
time or wanting to reboot to "the other one", for sure, and it was 
realizing that and figuring out how to overcome it for me, that was the 
key to my own switch.

> I /should/ get another machine.  But the thought of "jumping ship" is
> plaguing my choices:  Do I get another Mac, or do I *really* jump-ship
> and get an AMD-based system, with a *real* open-source system (Linux,
> *BSD, etc.).  If I truly want to jump-ship, I do not want another
> Intel-based system, and I can see how other chipsets are no-longer
> viable (e.g. PowerPC, my fav).

I've never really had the room or desire for a second machine.  Of 
course, if the one was my VCR/DVR, that would change things.

FWIW, I had two good quality 4-head stereo VHS VCRs at one point (to 
watch the previous evening's shows while recording the current evening's 
shows), about 20 years ago now, before I got heavily into computers.  But 
as I got into computers and as my frustration with the idiot commercials 
that with that technology could only be 30-second ff-zapped grew in 
parallel, I spent less and less time with the TV, until in a mobo upgrade 
back before the turn of the century, I realized that the last time the TV 
card that allowed the monitor to be both TV and computer monitor had been 
turned on, was as a functionality check two years before, after the 
PREVIOUS mobo upgrade.  At that point I decided it was safe to remove the 
thing, and I've not had a TV since.

But having had the two VCRs back then, I know how big a deal it can be to 
be able to record the shows so they don't get missed.  And then there's 
/watching/ them.  Today, yes, I'd have a second machine (actually a third 
as I have a netbook, now, running Gentoo, like my big machine, but since 
the big one's 64-bit and the netbook's 32-bit, they can't run the same 
binaries and I build the netbook's stuff in a chroot on the main machine, 
then ssh it over), if the one was the DVR.

> Either way, even if I stick with my current Mac, I am going to be forced
> to recompile every single project, so they'll run under Lion, or so
> they'll run on another machine/system/platform.
> 
> And most if not all of the OSX "packagers" are not ready for Lion -- it
> can take a very long time (at least a year I'd say, based on earlier OSX
> versions experience) to get patches up-stream'd into the projects repos
> everywhere, to be able to compile "out of the box" (src tarballs)
> without further specialized patching.  (I really do detest these
> "packagers" esp'ly for Macs.)
> 
> So, that's where I am at, right now, trying to figure-out what to do.
> Do I jump ship, and if so, to what?

What I did back when I switched was put Linux on a new drive, keeping the 
old one for Windows.  But your situation is different due to the DVR 
thing you have going, and machines are cheaper and more powerful today.  
Thus, I'd say, either leave the existing machine running OSX and doing 
the recordings, and buy a new machine for Linux, or buy a new machine for 
OSX and put Linux on the old one.

Personally, I'd do the new one as Linux, leaving the old as "legacy" both 
in OS and in hardware, much as I did with just the drives in my case.  
Because I knew where my future was and that the legacy was just that for 
me, legacy.  I think making that decision helped me make it so as well, 
as I began thinking of the old one as legacy right away.  If you make the 
new machine a Mac, and put Linux on the old one, you're probably more 
likely to end up with an attitude that matches.  So if you /want/ your 
future to be Linux (or one of the FLOSS BSDs, for that matter), make it 
the new machine and act like you believe it WILL be your future, and 
you'll be more likely to make the switch and have it "stick".

OTOH, if you want to keep OSX as your OS of choice, make that decision, 
make it the new machine, and put Linux on the old one.  But from my 
experience, I can safely say, that's about what you'll then get, too.  
You'll think of Linux as legacy and won't do much with it, and will end 
up emphasizing OSX on your new machine.  Or at least that's the way it 
would happen if it were me and I tried it.

As to what you jump to, that's a far bigger and more loaded question.

Honestly, I've not spent much time learning about the BSDs because I 
really believe more in copy-left licensing.  However, it's quite possible 
that if you DO know the OSX command line reasonably well, you'll be more 
immediately comfortable on FreeBSD or DragonflyBSD, as their kernels and 
command line programs are likely to be closer to OSX than Linux will be.

OTOH, Linux has far more momentum and current development going on, and 
supports a far larger ecosystem.  In a number of critical areas, 
including X with KMS (kernel mode setting), the u* set of utilities (udev/
udisk/upower, etc), and just coming up, systemd, the BSDs have fallen or 
are falling seriously behind and are in danger of becoming irrelevant to 
current state-of-the-art developments, to the point where it's now 
affecting compatibility.  There are certainly advantages to running the 
BSDs, but if you don't want to be in the technological backwaters forever 
fighting to keep up, the fact is, the bsds probably aren't where you 
ought to be looking, because that's a very real challenge they face, and 
it appears to be getting worse, not better.  Maybe that's a challenge 
you'd like to help tackle, tho, in which case...

If you decide to go Linux, then which one?  One of the things to realize 
at this point is that in many ways, each of the BSDs is more accurately 
compared to a distribution, than to Linux in general.  People from the 
BSD camp are often befuddled by all the distributions, all Linux.  But 
it's easier for them if they think of each Linux distribution as like a 
different one of the BSDs, as that's a more practically accurate way to 
treat the distros, than as variants of the same same kernel/BSD, since 
there's really no BSD analogy to Linux distro.

After realizing that, the next basic you want to nail down, is how much 
control you want over your own destiny, and whether building everything 
from source using scripts (much like IIRC it's FreeBSD that has the ports 
collection), to get the extra control it allows, or settling for having 
the distro make most decisions but in turn getting prebuilt binaries, is 
better for you.

Along a line from the whole group of binary distros on the one end, to 
Linux from Scratch (LFS) on the other, Gentoo (what I run) is about in 
the middle between a normal binary distro and LFS, where you build 
everything using make, but following instructions.  Gentoo builds 
everything, letting you choose many of the compile-time options and thus 
run-time dependencies using USE flags, but does so with scripts that 
automate much of the process.  So it takes a long time to build, but once 
you get things setup the way you want, you pretty much let it do its 
thing, and you can do other things while it does.  But Gentoo isn't for 
everyone, and they/we don't pretend that it is.  It does try to provide 
good documentation, but there's less handholding than some of the other 
distros, and they/we expect Gentoo users to be willing to properly manage 
the system, taking care to read warnings when they appear about specific 
upgrades and act on them accordingly, etc.  For those who prefer the 
distro make those sorts of decisions so that it hopefully "just works" 
for the user, Gentoo is the wrong choice, and we're very happy to refer 
you to another distribution where you're likely to be far happier.

Just as Gentoo is about half way between normal binary distro and LFS, 
Arch Linux is about half way between Gentoo and a normal binary distro.  
The core is binary, but many of the non-core packages are built from 
source using user-supplied scripts.  AFAIK it's about half way between 
Gentoo and a normal binary distro in comfigurability as well, more 
configurable than a binary distro, but less configurable, tho that means 
less hassle too, if you don't want to spend the necessary time to 
actually understand all those config options).

Then there's all the binary distros...

Another aspect where a decision up front will help you make the right 
choice is rolling vs. periodic upgrades, and how often you want those 
periodic upgrades to happen, with the size of upgrade and length of time 
between upgrades being strongly related.

Most traditional distributions come out with releases every six months to 
three years.  The community distros tend to have a shorter release cycle 
(6-8 months) and normally only support current plus one back, so you 
pretty much have to upgrade every year (two release cycles) or go into 
unsupported.  LTS, long-term-support release, meanwhile, tend to be what 
the corporate types use, with upgrade cycles of a year and a half or so 
and support out 3 years minimum, out to 7 or even 10 years.  These tend 
to be where the paid support is available, tho there are free (gratis) 
options as well.  Of course, upgrading these LTS systems is a serious 
project, but the timing is far easier for the corporate verification 
systems to handle.

On the other end of the scale is the rolling upgrades I mentioned 
earlier.  These are getting more popular especially with desktop users, 
as they tend to be more upto date, and avoid the "everything changed at 
once and there's a problem, where do I look" issue of periodic updates, 
because individual package updates happen as they are tested and ready.  
A six-month-cycle distro will often have packages released three months 
earlier at ship, due to pre-ship testing and integration, making them 
nine months old when the next release comes out.  A rolling release might 
take two months to test and mark stable, or maybe four if bugs show up in 
testing, but 2-4 months from upstream package release is FAR faster than 
the 3-9 months of even 6-month periodic releases, with the LTS 
stale^h^hble releases being MUCH older than that.  Also as mentioned, 
since individual packages are constantly being updated, when a problem 
does come up, tracing it down and fixing it is generally much easier 
because the search space is only a few packages, not the entire just 
updated release.

Gentoo and Arch are both rolling-update.  There's also a number of 
mainstream binary distros that are rolling-update, tho most are periodic.

As you can probably deduce by now, a few basic decisions at the outset 
begin to narrow your search space for your ideal distro match 
dramatically.  The ones I've emphasized to this point have been toward 
the customized, from-source, rolling-update end, but that's what I'm most 
familiar with.  That end tends to be less mainstream, however, and thus 
offer far more limited choices.

If you decide that a more centralized choices made for you, binary 
approach, is closer to what you want, then there are three other 
decisions that can help narrow down the binary mainstream.

The first of these is default desktop environment (assuming you want X 
and a graphical environment at all, that is, but it sounds like you do).  
With the from-source distros this isn't such a big deal as you make many 
of the compile-time optional support and thus run-time-dependency choices 
yourself, but if you're going with a binary distro, these choices are 
made for you, and they tend to enable optional support in all packages 
for their default desktop, but only selected optional support for other 
desktops, so you probably want a distro that chooses as its default 
desktop the desktop that you will make your own default as well.

The big desktop choices are gnome and kde.  xfce, lxde, and enlightenment 
are a bit lighter on the resources but not as full-featured or as 
popular.  Then there's the real light ones, *box, icewm, etc, and at the 
extreme end are ion and similar tiling window-managers, designed for easy 
keyboard-only use and almost graphical command-line in function.

Then there's Ubuntu's new unity interface on top of Gnome.

I'm a kde user so my views are obviously tilted toward it, but the 
general approach toward configurability differs markedly between it and 
the other two big ones (gnome and unity).  KDE's general attitude is more 
toward the make it customizable end, and they expose a lot of config 
options.  This is great for those that like to tweak things but can cause 
those who just want it to work to get scared at all the available options.

Gnome, OTOH, tends toward a "there's only one right way" approach, 
exposing far less configuration directly, tho a lot is available by 
editing the config files manually or thru a config tool that reminds many 
of the MS registry editor.

Meanwhile, while kde went thru a big and very disruptive upgrade to its 
4.0 about three years ago, the dust from that has for the most part 
settled by now (4.6 current, minor 4.x releases every six months, bugfix 
4.x.y releases monthly), while gnome is fresh into a major upgrade of 
just a few months ago, with first release of the new version still coming 
out on many distros, and it's stirring a LOT of controversy.  kde4 does 
have nice 3D effects, but falls back gracefully to 2D if 3D/composite 
isn't available or the user prefers to do without the flashy stuff.  The 
desktop itself, called plasma, defaults to a widget (called a plasmoid 
since it's a plasma widget =:^) mode (which supports both Apple's widgets 
and Google widgets, BTW, if you have any you'd like to keep that aren't 
Apple-specific-featured).  There are however a number of alternate modes 
including a traditional single desktop folderview, a search&launch view 
that with its big icons and simple layout is reminiscent of mobile touch-
screens, a scrolling newspaper view that puts the widgets in columns like 
a newspaper (again, nice on a smaller screen netbook, tablet, or other 
mobile devuce), and newer grid and grouping layouts that I've not really 
played with.

The previous gnome2 was looking a bit dated, but was very mature and 
stable.  It was originally a 2D interface, tho the compiz 3D effects 
window manager had proven very popular in the last few years.  Also, a 
lot of the original strictness of the "our way is the right way" approach 
had faded and it had gotten rather more customizable over the years.

The new gnome-3 interface is 3D only (there's a fall back similar to the 
old gnome2, but it's severely emasculated when it works, and often won't 
work at all), and again strips much of the additional config 
functionality, with the result being gnome3 as its authors intended, but 
not as many people would prefer it.  It's also very 3D and composite 
effect oriented and thus rather far from the traditional comfort zone of 
many current gnome2 users, many of which are departing for other 
desktops, either kde or xfce/lxde.

Meanwhile, one of the most popular "Linux for ordinary people" distros 
has been Ubuntu, sponsored and founded by a millionaire named Mark 
Shuttleworth.  Originally based on gnome2, Mark evidently decided he 
didn't like gnome3's styling and announced a new gnome front-end called 
unity.  Unity was originally targeted at the smaller screen of netbooks, 
tablets and touchscreen phones, but the vision there is a unified 
interface across all devices, and thus it now appears on the desktop as 
well.  Unity is based on the gnome backend, but is a very different front-
end.  Ubuntu is the only one with it as default, but some other distros 
ship a version of it as an option.  Unlike gnome3, it has a reasonable 2D 
fallback mode, should that be necessary or desired.  Ubuntu/unity went 
its own way with its defaults, which upset a lot of users (remember it 
was gnome based, and gnome isn't exactly known for its configurability, 
this is a crowd that doesn't like to spend time switching from the 
defaults), but for those that are willing to configure them, many of 
these defaults can be changed, so I believe it falls between kde and 
gnome3 on the configurability scale, probably pretty close to gnome2.

Gnome, meanwhile is built on gtk (much like kde is built on qt), which 
also forms the basis for both xfce and lxde, as they're gtk2 based.  As 
mentioned earlier, they're also rather lighter on resources, but 
corresponding less full featured.

You probably well know this already, but just in case, I should mention 
that as long as the distribution you choose includes the packages or if 
you compile from source, you can run programs based on one desktop, in 
the other ones, just fine.

With that background, we can now list the binary distros and their 
default desktops.  Ubuntu as mentioned uses unity.  Fedora and its 
corporate sponsor use gnome.  OpenSuse, which was owned by Novell, who 
just sold itself to Attachemate, is quite well known for its kde desktop 
tho it tries to treat kde and gnome reasonably equally.  There are ubuntu 
based sub-distros for kde (kubuntu) and xfce (xbuntu).  The struggling 
Mandriva is pretty good kde but some of its distro tools are gtk based, 
thus closer to the gnome side, I believe, so it's sort of neutral.  The 
Mandriva fork, Mageia (formed by some of the folks Mandriva laid off) is 
really too new for me to know much about, but it should be much like 
Mandriva at this point as the fork is pretty recent.

There's a new Linux, Bodhi, based on Enlightenment, which I haven't 
discussed much as I don't know much about it, but it's comparable to xfce 
and lxde in weight while being based on different technology.  
Enlightenment has reasonable 3D effects but works in 2D as well.  Bodhi 
is kind of new and small at this point, being the work of one man, 
mostly, but you should be able to use the package repositories of the 
distro it's based on (which I can't recall at this point, Debian, maybe?) 
for other packages.

Debian is a very large and mature community based distribution known for 
its large package repository.  Many child distros including Ubuntu are 
based on it.  I don't believe it favors any one desktop, but has packages 
for all of them I've mentioned (but possibly ion, which isn't fully free 
and has licensing issues), plus more.  Debian is also known for its 
sometimes way longer than intended release development cycles and it's 
sometimes rather heated internal politics, including on just what 
constitutes "free" software, since it doesn't ship non-free in its main 
repository.  (It's also known for its package format, but that's covered 
below.)

Mint is a small but growing in popularity distro, originally based on 
Ubuntu, but now rebased or rebasing (I don't know if it's complete yet) 
on Ubuntu's Debian upstream.  Originally popular primarily because it was 
small enough and its devs apparently out of the mainstream enough that it 
was able to ship some of the unfree packages like flash and mp3 codecs 
that Ubuntu didn't, so it "just worked" out of the box where Ubuntu users 
had to download something else from somewhere else (and both are aimed at 
the "it just works crowd", it's now getting many Ubuntu refugees leaving 
due to the switch to Unity, while Mint is sticking with Gnome2, thus it's 
switch to Debian as its direct upstream.

It's worth mentioning that this Debian derived Mint is one of only a 
handful of rolling-update based binary distros.  I /think/ I saw that 
there's an OpenSuse rolling-update version now as well, tho its main 
version remains periodic release based, 8-month cycle, I believe.

The second factor for primarily binary distro choices that can help you 
decide the perfect distro is packaging format.  The binary Linux distro 
world has two primary formats, rpm, originated by Red Hat (rpm stands for 
redhat package manager), and deb, originated by Debian.  These have more 
or less the same features from a user standpoint now, tho deb led in the 
early years, but the binary distro world remains split between them due 
to there being so many debian based distros on the one hand, while rpm 
has the LSB, Linux Standards Base, going for it on the other, rpm being 
the package format blessed by the LSB.

(LSB, while a useful standard defining many common aspects of a compliant 
distribution, such as a standard filesystem layout thru the fhs, file 
hierarchy standard, is from a user perspective of most used to those 
interested in running non-free binary-only applications, since it's 
generally only these that really depend on the LSB in the first place, as 
source-based packages tend to be flexible enough to allow defining this 
sort of thing at compile time and thus per distro or per individual 
installation.  Other than as a standard to which such binary-only 
packages can be built, the LSB doesn't tend to matter that much to 
ordinary distro users, tho it does matter a bit more to application and 
library devs and even more to distros, since they must keep it in mind 
for their work.)

There's a few others isolated distros using their own binary package 
formats, but they don't tend to be mainline.

So package format isn't really a deciding factor on its own, except that 
the binary distro ecosystem is divided pretty much in half, with the 
debian based and thus deb distros on the one side, and the lsb and thus 
rpm based distros on the other.  So whichever way you go, it's going to 
define where you can shop for alternative packages and your view of the 
FLOSS world thru the package format prism.

The third binary-distro-focused factor is language and localization 
(l10n, l, 10 letters, n).  While many/most distros ship a pretty wide 
l10n assortment, where the distro is based can play a big part in how 
well they can support you in your own language or others in which you may 
be fluent.  Mandriva is popular in France as that's where Mandrake was 
originally based, tho it's pulling out to mostly Brazil (original home of 
Connectiva, which Mandrake merged with to form Mandriva), now, I 
believe.  Mageia is formed of mostly the cut-loose French devs, I 
believe, so it should continue to be popular in French speaking 
countries, while the Connectiva base of Mandriva makes them popular in 
Latin America.  SuSE was originally based in Germany, and is headed back 
to its German roots now that Novell sold to Attachemate which is 
splitting off SuSE into its own nearly independently run department, once 
again, so it's very popular with Germans.  There's a number of Chinese 
and other oriental based distros that I know little about (Red Flag 
Linux, official Mainland Chinese, the Taiwanese Linpus Linux, Fedora 
based, that originally shipped on my netbook, others...), for that side 
of the world.  But most distros tend to be English developed and 
primarily English supported.  That's certainly the case with Debian (tho 
it has a large international developer base, English is the standard) and 
Red Hat and Fedora as they are US based.  On the non-binary side, Gentoo 
at least has contingents from other nations, but like Debian, the primary 
development and communications channels are in English.  Arch seems to be 
English at least primarily, but it may have a secondary that a number of 
its devs speak.

Based on all that, you should be able to narrow down your choices quite a 
bit, first to Linux or one of the BSDs, then if Linux, binary or from-
source and thus how configurable you want, which if you want from source 
gives only a few choices just as the BSDs give you only a few choices.  
If you go binaries, then you can narrow it down by rolling-update vs 
periodic release, with rolling update only giving you a few binary 
choices, and if you go periodic release, you can narrow it down by 
release period and level of support.

If you need more factors, default desktop, choice of package format, and 
choice of language, can help.

By the time you're done, you should have narrowed down your choice 
reasonably effectively and can then choose from only a small handful 
based on more specific factors.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman