On Mon, 6 Dec 1999, Tomasz Wegrzanowski wrote:
> On Sun, Dec 05, 1999 at 09:13:17AM -0700, Richard Stallman wrote:
> > What does it mean to make a new GNU/Linux distro that is more
> > user-friendly and entirely free? How would it differ from Debian?
> > It seems to me it would differ in two ways:
> >
> > 1. It is entirely free. You could achieve this starting with Debian
> > by eliminating the non-free and contrib categories (the definitions
> > may need a little adjustment), getting rid of some links to them,
> > changing some HOWTOs, and a few other things.
> >
> > 2. It has some additional software for user-friendliness.
> >
> > If you think about it, #1 is basically equivalent to what I've been
> > asking people to do since about a year ago: to make *in effect*
> > a version of Debian which has only the free software (main and non-us).
> >
> > But instead of forking the whole distro, I've proposed doing this by
> > making some structural and configuration changes to make it possible
> > for these two versions of GNU/Linux to share almost everything.
> > That seems both more efficient and more friendly.
> >
> > As for #2, that software needs to be written, but Debian would
> > gladly accept it I am sure. So there is no need to fork, no need
> > to make a separate project--not for reason #1 and not for reason #2.
>
> I see three other reasons to eventually fork :
>
> 1. I cant recommond Debian to GNU/Linux newbies. So I have to
> reccomend them some not-100% free distro. The whole Debian is too
> difficult to change. It *may* be easier to fork.
Yes. See my comments below.
>
> 2. Debian's packages quality is very inequal. We cant force
> Debian to make Quality Standards, and if maintainer of some
> package dont think this is a bug, we can do nothing.
>
> 3. Debian is going out very rarely. Potato and half is
> very good idea (is anyone sure woody will be in 2000 ? Im not)
>
> Our possible distro have to be .deb compatible.
> And will probably share also most of base and
> many other packages.
>
I'd just like to add my two cents (they're rather big...perhaps
they're two dollars?) to this little discussion.
There's a problem with the GNU/Linux world. Namely-- I can grok it. You
guys can grok it. Most people with a brain and enough time on their
hands-- and the desire to a learn a better way of computing-- can grok it.
However, the vast majority of people out there either CAN'T (for lack of
intelligence) or WON'T (for lack of desire to expend the requisite energy,
-which is significant-) grok it. Many people simply feel paralyzed when
they're not in front of a GUI. Not even just a windowing system like X--
but a _GUI_. They _need_ a desktop metaphor. They _need_ drag 'n' drop.
They _need_, in short, a perfect little Mac-ish WIMP interface-- or they
can't use the computer. It's not "user-friendly". It's "too hard". It's
"scary". It's "ugly". And they don't touch it.
What we have in the GNU/Linux world is a phenomenal set of tools,
available to be freely used and modified. But, as of just yet, there is
really no way to convince Joe Average to use the stuff. Most people can't
simply go from a Windows/MacOS world to the Unix world without extreme,
trying efforts. There's just no way to wean themselves slowly off of the
GUIish monolithic thought patterns (as opposed to modular thought, as
things are in Unix-- whereas Windows people expect OSes to be made up of a
few HUGE monolithic apps that each perform a billion functions, and if you
can't find the function you want on the menus, well, it's just not there,
Unix people expect OSes to be like pails of Legos out of which you can
build anything-- the shell script/Perl or even C metaphor, in other words)
and onto Unix's more console-oriented, modular thought. Perhaps even more
importantly, while free software has created programs -better- than the
equivalent proprietary software in many areas of the hacker-oriented world of
programming languages, compilers, OS kernels and the like, there is no
free software GUI system that's better than the drek they put out in the
Windoze or MacOS worlds. It's a sad fact, and one that doesn't bode well
for the future of the free software philosophy.
Like it or not-- and I know I don't-- most people are not only selfish,
but lazy. They will only switch to-- and advocate-- free software if they
are shown that it's not just something that a bunch of hackers (which many
of them think means "computer criminal") can benefit from, but them-- the Joe
WinIdiots of the world-- as well, and (and this part is critical!)
without a massive learning curve involved. User-friendliness is NOT
something that can be merely grafted onto an existing system through the
addition of a few programs. It's something that must be worked into every
package in the system. Why should only geeks be able to even FATHOM
something like grep? Where is the _Grep for Smarties_-- the guide that
explains things like grep in no-nonsense language to anyone with a brain,
but little to no knowledge of computer terminology like "string" or
"case-sensitive"-- perhaps explaining these terms along the way? Why is it
that even my friends who -are- highly intelligent, but have little to no
Unix experience, take so long to learn even this critical and relatively
simple tool?
I fully agree that the important thing about the free software movement is
not making the software popular, but making it high-quality. But let me
put it this way. The free software movement is DYING. I am sorry to make
this bold assertion, but it's really true. While the number of true
free-software hackers may still be going up, (at least for the
moment...), this is solely due to the exploding numbers of people in
the GNU/Linux world, and the numbers of "open source" people-- that is,
people who, in the past, might have been a part of the free software
world, but now place marketing and money above freedom (most important of
all) and quality (also important, but not as important) on their priority
lists, are going up even faster. The "open source" crowd, with very
different goals than the hardcore "free software" advocates, is edging out
the "free software" crowd. Nearly every time someone finds out I'm into
Linux and I dislike proprietary systems, they make the assumption that I'm
into-- and I quote them directly-- "open source". If I had a nickel for
every time I'd been assumed to be an "open source" person, a follower of
ESR, a Red Hat-lover, or the like, I would be quite wealthy. And while all
of this is happening, Windows's "market share" is ever rising. Whole
generations of children are growing up on Win32 systems, not knowing that
anything else exists-- and if they do, it's an obscure little system that
they simply call "the Mac", and most ridicule merely since they don't
understand it. In the minds of the masses, the concept of "free software"
simply does not exist. In the minds of the coders, it's a quaint little
idealistic notion led by a wild-haired freak named Richard Stallman, who,
while most of the smarter ones will concede is directly responsible for
huge chunks of what they know and love today about GNU/Linux, or even
Unix, is "an extremist" and "too militant about his views." I'm not
expressing my own ideas here. I'm expressing what I'm seeing from my
fellow coders. Free software is dying, and I'll tell you why.
With every passing year, the stakes in the computing world grow higher, as
more and more people make computing a significant portion of their lives,
and those who already use computers significantly begin using them -more-
significantly. Even your average idiot nowadays has an e-mail address and
a network connection, be it permanent or sporadic. Think about that from a
historical perspective. Had somebody told me ten years ago that everyone
would be connected in a global computer network, with e-mail and graphics
and sound and the ability to transfer dozens of files in mere minutes,
within ten years, I would have laughed. I would have said, "Of course
not.. that will never happen!" But it did. Each year, computers grow more
and more important. And in the Western world, when something gets
important, it gets proprietarized.
This process began years ago, and its effects were what prompted RMS to
start the Free Software Foundation. For a while, the things that RMS
created-- the framework of the GNU system, which the addition of the Linux
kernel made into a viable OS-- were popular among geeks, hackers, computer
lovers. People spoke of "the revolution"; people spoke of the dream of
free software finally coming to fruition.
But more recent events, I believe, are making that dream become, bit by
bit, more hopeless, more distant, more unlikely than ever before.
The reasons are quite simple. The use of computers is ever-rising, as are
the profits of proprietary-software companies. Now, there are two primary
groups of people joining the GNU/Linux world: proprietary software people,
such as Corel, who obviously have no love for the free software ethos, and
"open source" people, like those behind Red Hat, who care more about money
than about freedom. (Witness the fact that Red Hat's first CEO was
admittedly not a hacker, but a businessman. This alone is extremely
telling. IMHO, any tech company that really gives a crap about the tech
they deal with should be headed either by a geek-and-suit duo such as the
old Wozniak/Jobs combo at Apple, or by one of those rare beasts who
understands -both- tech and business.) The original, freedom-loving "free
software" people are being crowded out on their own turf by people who
have no commitment whatsoever to promoting and increasing the freedom of
those who use software. Their priorities lie, quite simply, in their own
pockets.
And as these people run amok on old "free software" territory like the GNU
project and, more recently, Linux, they drown out the voices of the
freedom fighters (for that is what, in their own little way, free software
hackers are). As their momentum increases, they will convert more and more
people-- and as the size of the computer-using community increases year
after year, the ability of individuals motivated by freedom and software
quality (and not money) to stop the tide will shrink, wither and,
ultimately, all but disappear. If the chances that free software would
overpower proprietary software were one in ten in 1990, they will be one
in twenty in 2000, and one in a hundred by 2010. In 1990, the free
software community-- which was the last great hope of the hacker ethic--
was at least pure, and committed, and full of youthful energy. Now, as we
approach 2000, the very free software world is being polluted. Companies
like Corel, Caldera and Red Hat quite clearly exist for the primary
purpose of making money. Their actions have demonstrated equally clearly
that they are -not- in the same category as RMS. Some of them aren't even
in the same category as ESR! They are not part of the "free software"
crowd, and many are not even part of the "open source" crowd. Were I to
draw a spectrum where the left side represents those who care about
mostly freedom, and the right side represents those who care about mostly
money, RMS and the "free software"/"GNU" crowd would be on the far left,
with ESR and the "open source" people somewhere around the center and
moving ever rightward, and some of the companies now joining the "Linux"
bandwagon (as they call it.. many of us would call it the "GNU/Linux"
bandwagon) would be even further to the right. Corel, for instance, has
spent over a decade producing proprietary software. Come on, people, do
you -really- think they're going to change? And Red Hat, the darling of
every financially-inclined Unix hacker's eye-- do you really think that
they care about "open source", as they say? Please consider carefully that
these are the people who started out with one decent distribution with no
major proprietary software components, and now-- in addition to selling
proprietary software like Red Hat Motif, sell various "enhanced"
distributions of software with all sorts of fancy encryption "products"
included-- which, I would wager, are as non-free as Windows itself. The
Red Hat concept is to make money off of Linux, and if they have to bend or
break the social rules of the "free software" or even the "open source"
philosophy to make that money, they will do so.
What has happened here is simple. As the ranks of computer users grow, the
percentage of computer users who are hard-core geeks, hackers or whatever
you choose to call them, shrinks. The importance of Joe WinIdiot has been
growing, and it will continue to grow. The people who created the modern
"open-source" movement-- which is in fact just a corrupted/modified
offshoot of the "free software" movement in much the same way that modern
Christianity is a corrupted/modified offshoot of Judaism-- recognized
this. By way of cashing in on this fact, they are (not due to any vast
conspiracy-- I am not so paranoid-- but merely as the natural conclusion
of their dreams of ever-higher profits) slowly turning their own
corrupted spinoff of RMS's original dream into something strongly
resembling Bill Gates's dream of profits built on proprietarianism.
Eventually, the two will merge. In time, the "open source" movement will
be so similar to the traditional Microsoft/Lotus/Apple/Corel/Oracle
mindset that these two branches of computerdom will become one-- in fact if
not in name. The focus is already shifting away from freedom and towards
profits-- and as a result, we are seeing Linux slowly turning into
Windows. The "open source" people are backing KDE-- whose parent library
only was made freer because GNOME made it virtually impossible to remain
so non-free-- which is, basically, Yet Another a clunky (like Windows) and
slow (like Windows) GUI. They are backing companies like Red Hat, Corel
and Caldera, who routinely mix proprietary software with their dists,
advertise it on their sites, and sell proprietary software packages
outright. It is abundantly clear that none of these companies places any
significant weight on freedom. They do, however, put significant weight on
profit. It was this same sort of thought that created the proprietary
software world that RMS got so upset at in the early '80s-- and that right
now is corrupting the world of Linux-based systems into something that,
fairly soon, won't be able to be called "GNU/Linux". It's only a matter of
time until the GNU components of most GNU/Linux systems (Debian, possibly
and hopefully, excluded) are removed.
Back to my earlier point about the Western process of proprietarization.
As something becomes more important to the general public, in the Western
mind, the "need" to propietarize it grows and grows. Or, to put it more
bluntly, when the potential user base of something grows beyond a certain
point-- when the technology becomes mature enough that it becomes feasible
to adopt it to mass use-- everyone stops doing it for the fun of it, or
for the love of it, and starts doing it for money. There exists--
as much as I dislike it, as much as so many of you may dislike it-- a
perception that once a particular NON-HACKER-ORIENTED technology matures
past a certain point, it should be commercialized-- and, in today's world,
that means proprietarized.
GNU/Linux systems, today, have reached that point. For a while-- that is,
until they came "under" corporate America's "radar", as they might say--
they escaped serious notice. But then, just as companies did back in the
early '80s when proprietary software finally became a major thing, people
started realizing-- "Oh my God, we can MAKE MONEY OFF OF THIS STUFF!" And
when they did that-- just like in the early '80s-- the priority on
anything !money went way down, and the priority on money shot up like a
Saturn V. Now, the world that RMS helped to create-- the world of viable
non-proprietary software-- is becoming corrupted, and it will only be a
matter of time before it will once more be impossible to create a full,
viable OS out of 100% free software. Sooner or later, you will find that
everyone is using proprietary libraries made or endorsed by Red Hat,
Corel or Caldera, or some other company yet to join the GNU/Linux world,
and that you can't run anything they can run, while they -can- run what
you can. You will receive files from someone running Red Hat 10.2 or
Corel Linux 5.0 and you will realize that they're only readable by a
proprietary program being licensed from Adobe or Caldera. And then one
will be forced to wonder-- when did the GNU/Linux world become so much
like the old proprietary software world of the mid-'90s?
If it can be of use to Joe WinIdiot-- if it's something that more than
just hackers can grok-- development and user choice (e.g. the number and
diversity of programs available) in the free-software arena will
lag behind that in the proprietary-software arena by years. Consider the
following examples:
* Adobe Photoshop, Corel Photo-Paint and many other high-quality,
-proprietary- paint programs were out for years before coding even BEGAN
on The GIMP.
* WordPerfect, Microsoft Word, AmiPro and many other high-quality,
-proprietary- word processors were out for years before coding even
BEGAN on any of today's free-software word processors-- which hardly
anyone has even heard of, and which are all, from what I can tell, still
in the process of creation-- none are nearly as stable or mature as the
proprietary offerings. And that scares me!
* Windows NT and Mac OS were out for years before GNOME was a glimmer in
its creators' eyes. To this day, GNOME (and KDE for that matter, though
KDE's not really part of the "free software" world, now is it?) doesn't
have the functionality that EXPLORER.EXE or the Finder do, while it may
be a bit stabler. It also gobs just as much RAM as any of the
proprietary "desktop interfaces"-- if not more. In fact, it chunks,
speed- and efficiency-wise.
The reason for all of this is that "free software" is, was and-- unless we
do something--will forever be thought of as merely something geeks do to
help themselves and their fellow geeks. As I have pointed out, in this
society, when something reaches the point where the masses can benefit
from it, it becomes proprietarized. Furthermore, in the process of being
proprietarized, it becomes corrupted. "Brute-force" logic becomes the way
to go-- rather than seeking to make everything as efficient as possible,
companies rely on the ever-advancing forces of technology alone to improve
their "products", and focus on getting things out the door-- and "to
market"-- as quickly as possible.
To summarize, and bring it all home-- the focus on the end-user is rising,
and yet end-user apps are still, in the coding community at large,
perceived as something you just -don't do- unless you're getting paid, and
the results of your work will be proprietary. What this means is that
unless we making a free-software OS that's truly user-friendly-- and no,
Debian + GNOME doesn't cut it (since Debian is -not- user-friendly, and
GNOME, frankly, sucks)-- free software will die. Simple as that.
What I am proposing is that a system be created that borrows the best
concepts and programs that Debian has to offer, and arranges them in a way
that benefits hardcore hackers like myself (e.g. people who don't use
Windows AT ALL unless they have to, and avoid MacOS, BeOS, etc.) and the
WinIdiots as well. A system where one can start out in a pretty GUI that
rivals even the Mac OS for ease of use -and- has the stability one should
expect from a Unix system-- then, bit by bit (not suddenly, like people
who come out of the Windows/MacOS world to the Unix world are now expected
to adapt), lose their dependence on silly GUI tools, and begin to make
clever use of the shell, clever use of scripting, of Perl, and the like.
What this would necessitate is several things.
1: *EVERY PACKAGE* should be documented. And WELL-documented. Perhaps the
best way to do this would be to write a book, _(NameOfDist) for Smarties_.
I've started a project to create high-quality, GPLd documentation-- the
free documentation that RMS has been asking for for so long-- and it's up
at www.forsmarties.net. If there could be a book that would eliminate the
need for man pages-- and perhaps the contents of it could be integrated
into a parallel man system (I'm sorry, but even -I- find 'info' to be
a pain to use... it'd have to behave more like man, IMHO)-- (perhaps call
it 'newman', for "Newbie Manual")-- giving info in a more newbie-friendly
format, then so many more people would be able to take the plunge into the
Unix world.
2: Error messages in Unix are of -very- inconsistent quality. Sometimes,
their meaning is obvious. Sometimes, their meaning is obvious-- IF you are
a hardcore kernel hacker (which even I am not). Sometimes, their meaning
isn't even present. (E.g. "Whoops! Badness." or the like). The error
messages in this system would have to be modified, even down
through the kernel-- e.g. a simple kernel patch, which perhaps
could be integrated as an option in future releases of the kernel,
would change all the error messages to something more readable,
while still containing the same information. All that this proposed change
to the system would requires is a series of simple patches-- or,
alternatively, what the errors REALLY mean would have to be documented in
newman, in the book, or in both. (I think it would be best if such an OS
was documented both in a man-like system and a book, which would contain
the same data). This path would leave the existing code untouched and
unpatched, but would be a bit of a kluge.
3: There needs to be a good GUI. A GOOD GUI. I'm sorry-- GNOME and KDE
don't cut it. In addition to the fact that the KDE people don't seem to
give a hairy rat's ass about free software, -both- GNOME and KDE are slow,
klunky and -still- not up to par with even Windoze's GUI, let alone Mac
OS's. I tried them both. They -both- suck. They're huge, they require a
bajillion megs of libraries (you KNOW something's wrong when the GUI
starts to edge out EMACS for RAM use) and they can't even get such simple
things as storing your settings (KDE-- which sometimes seems to completely
ignore your setting changes) or having the "desktop" automatically update
upon file additions. (both KDE and GNOME... okay, perhaps this sort of
thing would have to be done via a small kernel patch, or more conveniently
a module, as I'd wager there's no event to be caught by an app (KDE or
GNOME) whenever a new file is created-- but still, something needs to be
done about this very real problem)
Perhaps something nice and lightweight like XFce (www.xfce.org) would make
a good starting point? XFce is also based on GTK, a la GNOME, and comes
with a "GNOME compliance module" (whatever that is), so it might make it
feasible to run GNOME programs in it, and perhaps even vice versa?. I know
little about this "desktop environment" (as they call it), but perhaps it
would make a good base for a GUI for the system I'm proposing. It's
GPLd, and it's lightweight-- which I think is the primary obstacle to
making a GUI worthy of being associated with the name "Unix". All major
existing GUIs for Unix are either non-free, big 'n' clunky, or both. Why
should we, the people who care about efficiency and quality, borrow such
horrid traits from the Windows world?
4: There needs to be a good Web browser, as it has been pointed out.
5: There needs to be a good office suite.
6: Much as I hate this part-- there needs to be a good graphical coding
environment (ugh!) to appease the VC++ types.
Any thoughts on what I have said would be much appreciated. Please read
and consider this message carefully... I fear that if people do not heed
messages like this, eventually we'll find the free software ethic dead as
a doornail.
--Caspian
--
= Jon "Caspian" Blank, right-brained computer programmer at large =
.--------------------------------------------------------------------.
| Freelance coder and Unix geek / Founder, The Web Union (twu.net) |
| Information wants to be free! Visit www.gnu.org. |
| WANTED: Writers who share the GNU philosophy. www.forsmarties.net. |
| - - - - - - - - - - - - - - - - - - - - - - - |
| E-mail: [EMAIL PROTECTED] Web: http://caspian.twu.net |
| Send a short message to my cell phone: [EMAIL PROTECTED] |
`--------------------------------------------------------------------'
