Slashdot Mirror

Shortly after I emailed the link to d/l M5 to some friends, one replied with a link to JWZ's resignation. I guess he thought that JWZ leaving meant the death of Mozilla. Tell me if I'm off-base, but here's how I replied:
Losing Zawinski was a great punch in the stomach for Mozilla, but I think by no means is it dead. I think after Mozilla gets over this rocky stage, it will flourish. Here's why:
1. The code is just now becoming something presentable. Too much code done internally made it hard to follow in early releases.
2. People won't contribute until there is something there. I think that the nearer and nearer that Mozilla gets to their target, the more and more people will jump on board. The organization will probably explode when they release the final product, and the general public sees how good it is.
3. Milestone 5 was reached and released yesterday, without Zawinski. Life will go on.

I just thought I would also share my insight with all of you. Let me know your opinion, and thank you for your time. BTW, I am running M5 from my office on Windows NT, and it is pretty cool.
-NG



+--
Given infinite time, 100 monkeys could type out the complete works of Shakespeare.

Keep in mind a good point that JWZ mentioned in http://www.jwz.org/gruntle/aol.html :

Bear in mind that, for a publicly-traded company, if a CEO makes a decision because it's the right thing rather than because it's the most profitable thing for the shareholders, he will lose his job, and possibly be sued into oblivion. That's the way the rules work.

Just something to think about...

Try www.jwz.org/xscreensaver. It's a much better collection of screen savers than you would get with Windows, and (of course) it's free software. The 3D ones require Mesa, however. libXpm is also required or recommended for a few others.

It's this kind of story that really enforces one's belief that working for yourself can only be a good thing.

OTOH, M$ working methods would have to be drastically altered. It all seems like a very long shot for these reasons alone. Would they continue with daily builds? Would these be posted? Near the end when stabilization is the preeminent concern, would they really want outsiders opinions about the hack vs. fix trade-off decisions? How would they handle the product readiness and release to manufacturing decisions? Openly? That would be rich. The outside tinkerers and perfectionists would certainly be an irritation at the endgame. That's one phenomena the Mozilla crew has yet to face.

The question is, what kind of graphic visualization of the search should be done? What does their Windows client look like, anything?

Perhaps the display should contain multiple kinds of data, maybe a map of the sky in one part of the screen, and a graph of area searched over time, or hits found over time, or something like that?

Having it be an OpenGL hack might be interesting too: what if the "sky display" was a rolling sphere with a star-map drawn on it, with grid-cells that lit up for the area being searched, or the areas that had already been covered by others? (But maybe the SETI client doesn't work that way?)

Another thing nobody's mentioned is the ridiculous amounts of memory it officially takes to build the thing. We starving programmers don't usually have 128M of RAM lying around in our home boxen. Until recently, in fact, I did all my non-work development on a 486 with 16M. I still only have a K6/233 with 96M of RAM. I'm sure there are lots of very good programmers in the same situation.

For an explanation of this phenomenom, try reading:


http://www.jwz.org/worse-is-better.html

On the other hand, there are challenges ahead. But that doesn't diminish Netscape's past accomplishments.

I think this is a bad sign for Netscape/Mozilla. With Jamie Zawinskileaving and the recent layoffs at both Netscape and AOL, I'm beginning to worry about the future of the Mozilla project.

At the same time I think they can rid themselves of all doubt simply by releasing a product as small as iCab, has many more innovative features, and runs faster. There are two things lacking in iCab. First, it crashes on my computer. I expect this to be fixed soon. Second, it is slow in rendering images. I think Netscape/Mozilla can out do this easily, which would restore faith in the project, and enhance Mozilla's standing in the browser wars.

Good luck guys.

-Ben

It seems to me that a large part of the motivation for a free software contributor is the license under which the software is released. While Netscape was the first big-name corporation to release the source code to an important product, they were also the first to come up with a non-free open source license. There are still some ties to the corporation that birthed it, so contributors are basically unpaid developers for said corporation. If Netscape used the GPL instead of NPL, they may have attracted more outside developers. Many people enjoy contributing to the public good for free, but few enjoy lining the pockets of the stockholders for free.

I believe this issue will also haunt all of the other 'open-source' but not quite free software that is being made available on the net. These include Apple's APSL, Sun's CSL, TrollTech's new QT license, etc. Any license that even mentions the corporate donor/supporter will be shunned by free software developers everywhere.

I agree with many of Z's other conclusions, especially the fact that Netscape did not release the source code to a working, usable product. As a programmer, one definitely likes to see tangible results, not just X more lines of source in version control.

Finally, while the difficulties of Mozilla.org should not deter other companies from taking similar directions, others should definitely learn from their mistakes. Learn from the pioneers how to avoid the arrows in the back.

Cheers.

Of course, Mozilla isn't his only project. The xscreensaver open-source project has always been a favorite of mine, and various other UNIX types in the company. (Who can't help but to love those sproingies?)

Of course, Mozilla isn't his only project. The xscreensaver open-source project has always been a favorite of mine, and various other UNIX types in the company. (Who can't help but to love those sproingies?)

JWZ resigns from mozilla

Reading this paper was a very interesting experience. Its a wake-up call with some obvious clues.

Firstly, open source projects don't happen unless people contribute.

Secondly, no one really wants to spend their free time digging through cruft in order to figure out how to contribute. Most developers appear to prefera clean slate to cruft. JWZ seems to think that this was a mitigating factor to getting more outside help on Mozilla.

I can't say any of this astounds me - I program for a living, and its only fun some of the time. Other times its a pain in the rump - its work. Most people don't want to dig through mountains of cruft pro bono, and I can't say I blame them.

What a journey: twentieth employee at Mosaic, then part of the fastest growing company ever, then an employee of the AOL juggernaut.

Read jwz's resignation if you haven't since it's a fascinating read. I found his explanation of why mozilla has failed to ship in a years time facinating. And I especially liked how he ended the paper by saying that mozilla's failures haven't been because of it's going open source, but mostly in spite of it (except he uses the word pixie).

It's easy to understand why he finally quit. Anybody who is there at the beginning, who works best when the pressures are highest and the excitement is greatest, is unlikely to fit in later on when a company grows large and boring.

I just hope he continues to write great code and the occasional gruntle prose and occasionally liven up the mozilla newsgroups.

Jamie's got some stuff on his site. See the following urls:

Resignation and Postmortem

Netscape and Aol

Jamie's got some stuff on his site. See the following urls:

Resignation and Postmortem

Netscape and Aol

--

look -- it's YANNI !!!

and he's wearing black nail polish!

I just did a quick inventory of my Linux workstation, and an rpm -qa|sort|less revealed that around 15% of the packages installed on my system are of FSF origin. Surely originating 15% of the code does not give RMS the right to rename a project that's not even his! GNU/Linux indeed. RMS has HURD. Let him have his GNU/HURD. Without the work done by Linus, Alan Cox, David Miller, and scores of others, Stallman would have only ever been known (thus far) as the guy who heads up that project to write some free third party Unix utilities, and a compiler. Conversely so, Linus, et al, benefited greatly from Stallman's efforts. Does this mean that the Linux developers and distribution creators have any obligation to RMS to rename their materials/software/whatever? Nope. Take for example, WindowMaker, a window manager for the X Window System. WindowMaker is distributed under the GPL. Do we see RMS attempting to force the maintainers of that project to rename it to GNUWindowMaker, or somesuch? No.

Take for example, Richard's recent rant about why library developers should use the GPL rather than the LGPL. The gist of his argument was that linking against a GPL'd library caused the whole application to become a "derivative work" of the library, thus causing the application to be GPL'd as well. Rather, the LGPL, permits linking without necessarily GPL'ing the resulting program.

I'm very pro-Open Source, but RMS is hurting the cause by constantly acting out, causing the media to portray him as the "angry lone nut".

Would we be better off without RMS? Probably not. Could we do without RMS behaving like the "angry lone nut"? Yep.

Ok, so the current record industry sucks, and musicians hardly get paid, and CDs are overpriced. Granted.

And it's easy to see how the widespread use of MP3 could eventually do some serious harm to the music industry, by removing their revenue stream.

My question is, if you take down the music industry, what do you replace it with? How do you feed the artists who are recording music? How do you cause the music to continue to get produced? What's the business model for this new world?

• --
• jwz

[...the Broadway licensing debacle...] XFree86 has become more important than any of the proprietary X servers, and its implementation would have become the defacto standard.

I wish I thought this was true. I think you're wrong though. People in the Linux and (to a lesser extent) the remainder of the `Free Software' movement seem to have a rather blinkered view on many X related things. Here are some examples:

• Motif is the standard X toolkit (toolkits, I suppose). There's no getting away from it. GTK may look nice and be nicer to code, but Motif offers a whole lot more. JWZ posted a rant here some time about it (back in the days when it was possible to read all of slashdot...), and I think that he knows what he's talking about in this arena.
• CDE is the standard desktop. Personally I think GNOME and/or KDE stands a fighting chance of taking this crown (ever tried adding an icon to the control panel on CDE? Yuck.)
• XFree86 is important, but it's not that important. Even amongst Linux/*BSD users AcceleratedX and MetroX are very popular products. Move outside that, and you've got a huge number of different implementations. Neither Sun, Digital^WCompaq nor anyone else is going to be shipping anything `XFree86 compatible' any time soon: they're going to implement whatever the Open Group says is standard.

Don't get me wrong: I'm not doing down the importance of free software here. I'm trying to inject a bit of realism. Motif is better than you think it is (resources, anyone?). Thinking about it, XFree86 might just have a wider distribution than any other single X server, and I bet it'd be more likely to get upgraded than the rest of the X servers combined: they'll probably only get upgraded in sync with OS upgrades if at all. Nevertheless there is a huge inertia behind the installed base of non-Free X machines out there, and it wouldn't do to get carried away.

I seem to be rambling now, so I'll stop. Does anything I've said make any sense?

If so, how quickly?

Well, you've already got Control, Shift and Alt on the keyboard, so the obvious use for the pedals is as Meta, Myper and Super modifiers.

M-x 1000 praise-emacs

Of course, if you've got one of those dodgy Windows keyboards you can always use those wacky extra keys instead -- remember, xkeycaps is your friend.

It's amazing to me how even labelling of charsets turns into a Microsoft flame fest here every time! Don't you people ever get tired of gazing into your own navels?

For the record, this other character set that Microsoft uses (that is a superset of Latin1) is, in fact, an ISO-registered character set. It is properly known as ISO-8859-1-Windows-3.1-Latin-1.

In this particular case, Apple's poor decisions have been more of a problem than Microsoft's, because the character set that Macintosh tools use (MacRoman, which is not ISO-registered) is not even a superset of Latin1 (ISO-8859/1.)

If you're interested in facts, you can see an article I posted to comp.infosystems.* about this way back in 1997.

If you were really interested in doing something productive, you could contribute to Mozilla so that future web browsers could (as the old saying goes) be lenient in what they accept, and in a cross-platform way.

Or, you could just continue your endless whining here, feeling superior without actually doing anything, and wait for someone else to fix it all for you.

-- jwz

[...] one day they might even decide to implement the BSOD [...]

Hmmm. I take it you haven't seen a recent xscreensaver release? Amongst other things, it comes with bsod, which is a rather fantastic emulation of the fabled um, BSOD. It's got Windows, Amiga, Mac modes, and so on -- groovy.

The program you want is called xkeycaps. It lets you select a keyboard from a database of lots of different ones, and then writes out a file that is readable by xmodmap (the program that loads modifications to the keyboard map into X).

Get it at: http://www.jwz.org/xkeycaps/

So, just use xkeycaps to set up the keys to do whatever you want, save to a file (say, ~/.xmodmap-localhost) and put in your .xsession:
xmodmap ~/.xmodmap-localhost

You should read the xkeycaps man page after you install it, it sheds a lot of light about keyboards and how X handles keys. Good luck.

See (hear) the following:

Why cooperation with RMS is impossible