[conspire] return to a state of grace

Rick Moen rick at linuxmafia.com
Sun Oct 17 19:20:29 PDT 2004

Eh, welcome (more thoroughly) to the Linux world.  I always tell people
that the best way to learn the *ix way of doing things is to dive
straight in and use it -- primarily if not exclusively.  I say this
primarily to those wanting me to help them set up for dual-boot:  I'm
glad to assist them, but suspect they're kidding themselves a bit.
People find themselves spending all their time in one at the expense of
the other.

Quoting Eric De Mund (ead-cabal at ixian.com):

> My main system at home is now running Slackware Linux 10.0.

Slackware's neat.  One thing that's either an advantage or not depending
on your perspective is that it tends towards a sparse system with the
system framework exposed to view, and not a lot of "desktop" clutter
between you and that framework to deliberately conceal the latter.  (At
least, that's how Slackware seemed when I last used it.)

> q1:  Is there a (non-commercial) graphical web browser under Linux that
>      remembers all my open tabs from the last session? 

Meaning no criticism, I'll echo Don's mild cavil, on this point:  The
axis of commercial vs. non-commercial relates to commerce.  The axis of
free / open-source vs. proprietary relates to licensing.

Opera happens to be a particularly fine proprietary Web browser -- but I
doubt your concern about it relates to acquisition cost, but rather to
whether its development model gives third parties the right to do with
it whatever they wish.  If I bought a copy of Opera and then gave it to
you, then your acquisition would be non-commercial but the code would
remain proprietary.  If, by contrast, I sold you a copy of Mozilla
Firefox, your acquisition would be commercial but the code would be open

Anyhow, I know the sort of thing you're talking about:  Opera crashes or
your laptop battery runs out.  Later, you reopen it and Opera offers to 
re-fetch all the tabs you had open when it terminated.

Anyhow, check your installed copy of Firefox to see if it has that
capability built in, yet.  It might.  Or look and see if any of the
couple of hundred Firefox extensions linked from
http://update.mozilla.org/ adds that. 

Or try Galeon, which does exactly what you speak of.  {sigh}  OK, I'd
better provide a bit of browser history.[1]

The first Web browser was one called, oddly enough, WorldWideWeb,
written in 1990 at CERN for NeXTStep -- because that's where Tim
Berners-Lee was.  A bunch of others followed, including most notably
Mosaic, created in 1993 by students at University of Illinois's National
Center for Supercomputing Applications (NCSA).  Per standard university
copyright policy, which was not as enlightened as that of University of
California (which gave us BSD licensing), U. of I / NCSA made source
code available to the public for non-commercial use only.  If you wanted
commercial-usage rights, you had to negotiate a separate deal for big
bucks.  So, NCSA Mosaic was well known and ported to every graphical
environment -- X11, Win16, MacOS, Win32 -- but was never actually open

Several people who worked for a while at NCSA (including Marc Andreesen, 
primary author of Mosaic), after getting their degrees formed in 1994 a
company called Mozilla Development Corporation, with the specific aim of
out-doing and crushing Mosaic: Thus the name Mozilla.  They quickly
developed and released as beta code a really good browser under that
name (but binary-only).  

The university evinced no sense of humour about this, and threatened
tiny Mosaic Development Corporation with a trademark-violation lawsuit,
claiming that the name "Mozilla" was confusingly similar to their
"brand" name, and was likely to mislead customers.  Mosaic dealt with
this threat by renaming itself Netscape Development Corporation, with
a similar change to its product, which become Netscape Navigator, and 
eventually part of a suite of related apps called Netscape Communicator.
(The README for the browser, however, always concluded with something
like "And remember, its name may be Netscape but it's pronounced

Meanwhile, several firms made deals with NCSA for commercial development
of its proprietary NCSA Mosaic browser.  The best known and latest of
those was Spyglass, Inc., which sold a variant form of Mosaic for
MS-Windows as part of a suite of Spyglass Internet applications.

In 1995, Microsoft Corporation woke up to discover that the
open-standards Internet, whose importance it had dismissed while trying
to push MSN as a private, Microsoft-controlled ghetto instead, was in
danger of overwhelming its control of the desktop.  In particular,
Netscape Communications, Inc. was threatening to further develop its
popular suite, Communicator, as a full-blown desktop application
platform.  In a panic, Microsoft looked around for browsers to acquire
for money (Viola, Cello, Amaya, Chimera, etc.), and found hapless
Spyglass, Inc.   

Microsoft told that firm that it wished to sub-license its Mosaic rights
from NCSA to create a browser to be offered by Microsoft to Windows users,
and offered a deal where Spyglass would be paid a modest quarterly fee
plus a percentage of Microsoft's revenues for the browser.  Spyglass
signed, and then found that they'd been tricked when Microsoft started
giving away its MS-branded version of Mosaic, dubbed "MS Internet
Explorer", starting later in 1995.

In March, 1998, Netscape made the then-astonishing announcement that it
intended release all possible portions of its flagship product, the
Netscape Communicator suite, under open-source licensing terms -- citing
specifically the reasoning about product development Eric S. Raymond
wrote about in his essay "The Cathedral and the Bazaar".  Because of
third-party encumbrances on the code, the initial code release on March
31, 1998 would barely compile and wasn't particularly stable.  (I know,
because I compiled and ran it.)  Netscape used the occasion to
officially revive the name Mozilla for this new effort.

Worse than the big gaps in the code caused by removal of third-party
modules was the fact that the Netscape 4.x architecture had become, over
the years, embarrassingly badly tangled.  The Mozilla Organisation's
first year was spent largely arguing over the problem and summoning up
the will to make a painful decision:  Almost all of the old codebase had
to be scrapped.  Work started in late 1998 on a from-scratch central
rendering engine.  At first, it was called Raptor -- but that name
turned out (again!) to have trademark problems.  So, Mozilla started
referring to it as NGLayout (Next Generation Layout).  Netscape, which
was bought out by AOL in late 1998 but still makes heavy use of Mozilla
code, simultaneously started referring to it as "NGT", and more
informally as "Gecko" -- the name by which it's best known even though
it's technically named NGLayout.

Gecko turned out to be astonishingly good -- vindicating the Mozilla
Project leaders' decision in the face of widespread condemnation over
the delay.  But the portable runtime front-end for the browser, XPFE 
(the cross-platform front-end), remained controversial.  People on
numerous platforms were doubtful of the utility of all that extra code:
They didn't need something that was cross-platform programmable (in a
scripted language called XUL, by the way) and could (potentially)
support an entire office suite; they just wanted a browser.

So, a number of groups made platform-specific substitutes for XPFE.
One of them, on Mac OSX, was Chimera, which was soon renamed Camino when
its author remembered there's already an existing X11/Unix browser named
Chimera (plus an entire hypertext system of that name at UC Irvine --
again, trademark issues).  Camino is a browser using its own front-end
engine, and thus user interface, using the OSX/NeXTStep "Cocoa"
libraries, and Gecko for rendering.

Similarly, back over in the X11 world, a similar project built the
Galeon browser using the GTK+ rendering libraries and its own front-end
engine, using Gecko for rendering.

Through its v. 1.2 series, the Galeon project brought to the Mozilla
world a number of innovative features.  Some were implemented following
the example of the talented Norwegian coders at Opera Software ASA --
such as the "remember and reopen my tabs after recovering from a browser
crash" feature you speak of.  Disappointingly, Galeon 1.2.x wasn't
significantly smaller in RAM than its Mozilla 1.0.x rivals, but it _was_
definitely faster and took up a lot less disk space, along with having a
number of advanced features.  

Over time, though, the Mozilla Project picked up and reimplemented the
bulk of Galeon's best features.  Also, the third-party community of
people implementing optional modules ("extensions") in XUL started
taking off.  Best of all, the Mozilla Project announced that it would be
establishing fully modularised versions of its applications in the
future, so you could have just a browser without also getting a
mail/news client in the same big code hairball, and vice-versa (not to
mention a built-in IRC client and HTML editor).

The completely-separate Web browser was at first dubbed Firebird -- but
then was renamed Firefox after the developers of the relational database
Firebird (forked from Borland Interbase, which Borland first
open-sourced and then took proprietary again) complained bitterly.  The
completely-separate mail/news client is called Thunderbird.  

At the same time, you can still download the combined browser / mail /
news / IRC / HTML-editor thing.  It's called Mozilla Communicator, now
at version 1.7.3, and its development project is codenamed "SeaMonkey".  
The fate of SeaMonkey seems to be not-qyite-decided, but plainly
developers are leaning towards the fully modular codebases.

Anyhow, that's partly background for talking about the disappointment
that is the Galeon 1.3.x series.  Maybe you can see this coming:  The
browser started being infected by GNOME disease.  

As he started building a new Galeon version around GTK 2.0 starting in
March 2002, the lead developer, Marco Gritti, started stripping out of
Galeon numerous configuration and other features deemed potentially
"confusing" for the "Everyman" who was considered its primary intended
audience.  Additionally, numerous changes were made to dialogue design
and similar elements without regard to the logic of their content and
situations but just to impose a "consistent look and feel" and
comformity to the GNOME Human Interface Guidelines.

I was one of the numerous longtime Galeon users who were extremely
annoyed by those changes.  The lead developer -- in exactly the same
fashion as mainline GNOME developers -- took to dismissing the critics'
concerns completely, saying we were simply not the intended browser
users (and thus, please go away).  However, debate (or flamewar, if you
prefer) continued through 2002 because Gritti had _also_ alienated all
the other Galeon developers through his attitude and priorities.
Finally, in November 2002, Gritti left the Galeon project and started a
new one, Epiphany, which will become a very simple, HIG-compliant
browser, while Galeon would be freed to become a power-user browser. 

In consequence, Galeon is slowly recovering from GNOME disease -- but
was somewhat damaged from the infection.  Meanwhile, Firefox (which is
nearly at v. 1.0) continues to look really, really good -- especially if
you count the myriad "extensions" available that have been written in
XUL.  (Of course, many of those extensions are lame and pointless.
Sturgeon's Law applies.)

Anyhow, you want to keep an eye on both projects -- Galeon and Firefox
-- to see if they meet your needs.  Right now, it looks like Galeon has
been eclipsed by Firefox slowly gaining all the early Galeon virtues at
the same time that Galeon was dealt a severe setback -- but it's worth
watching them both.

> q2:  What application rips audio CDs, using error-correction? 

I defer to other people's answers, having had no experience with it,
myself.  The classic tool for that purpose, though, is cdparanoia, 

> q3:  Is CUPS now the preferred way to go, for printing?

Without a doubt.

>      I need to figure out printing. I have a NEC non-PostScript laser
>      printer (NEC Super-Script 870) directly connected to my system.


I personally got tired of not having a single place to go for Linux
driver questions, so I created one, which you're welcome to use as a
starting point for such questions:  "Help Resources" on
http://linuxmafia.com/kb/Hardware/  It's nothing special; just a link
farm, but I try to keep it current as a central starting place for Linux
hardware support questions.  

> q4:  What is an easy CD/DVD burning solution? Consulting my notes data-
>     base, I see that it was K3B that you were discussing and praising,
>     recently. Is that the way to go?

Seems like.  My stuff on the subject is at "DVD" on

> q5:  Is XMMS still king, for playing audio?

As usual with AV questions, I'm better off deferring to others, but
certainly XMMS is nice if a little funky-looking.

> q6:  What do people use for playing movies (e.g. avi's, mpg's)?

Old-school:  xine and mplayer.
New-school:  VLC Media Player (formerly VideoLAN Client)

> q7:  What do people use for scanning? Is SANE still the way to go? I
>     have an old SCSI scanner (ach! I forgot about this, and didn't se-
>     lect a SCSI-capable kernel at OS install-time), a Umax Astra 610S,
>     that as a Domex (DTC Technology) DTC-3181LE SCSI controller card in
>     the system.

(Again, you might want to start at my Help Resources page.)

SANE, first of all, is a software _framework_.  It's a modular design:
People write back-end pieces for diverse types of scanner hardware,
which basically presents a _generic_, universal programmable interface
that other (front-end) software can then talk to.  You can then use your
choice (preference) from various, diverse types of front-end (user
interface, control) software.

Anyhow, not at all surprisingly, the "umax" back-end drivers perfectly
support your Umax Astra 610S.  See:
http://www.sane-project.org/sane-backends.html#S-UMAX  That's not
surprising because, in general, SCSI-type scanners are a slam-dunk.
It's the poor schlubs with parallel-port scanners and other funky types
of connection who tend to have problems.

In any event, it's not the scanner that will be the problem, but rather
that cheap-ass ISA SCSI card of yours, which is based on the DTC-3181x 
chip, apparently requiring use of a Linux driver named "dtc3181x".  
Here's a relevant 1999 thread on the Linux kernel mailing list:

You'll notice that one key step is to _disable_ ISA Plug'n'Play mode via
a jumper on the card.  Of course, then you must ensure that you also 
set jumpers on the card to set hardware resources (IRQ, I/O base
address) to values that (1) are made available in your motherboard BIOS
to "Legacy ISA" devices, and (2) aren't being used by anything else.

And another on a SUSE list, showing how to pass the required
hardware-resource information as options to the booting kernel:

Explanation:  Really cheap SCSI cards such as were traditionally bundled
"for free" with SCSI-type scanners invariably did not include option
ROMs, you see.  One consequence of that is that Linux kernels cannot
autoprobe their existence by looking for the ROM signatures.

Really cheap SCSI cards also tended to be ISA.  ISA Plug'n'Play was an
attempt to make ISA devices self-configuring that never really worked
properly, and never will.  Although there is software available on Linux
to query isapnp controller chips for hardware resource information,
you're therefore better off disabling that mode of addressing on each
such card.  If/when you do, you must then provide the hardware resource 
information at the time the driver loads.

Isn't this fun?  If you'd rather not deal with it, then shop at Action
computer for a cheap PCI SCSI card, instead, e.g., one with a Symbios

[1] See also http://livinginternet.com/w/wi_browse.htm

Cheers,             "< > Kernel support for JAVA binaries (obsolete) (NEW)" 
Rick Moen                            -- Linux kernel v2.2.19 configuration
rick at linuxmafia.com 

More information about the conspire mailing list