Slashdot Mirror

GC is an obvious requirement of a "safe" language.

No, it isn't. Memory safety can be achieved without the use of garbage collection, by avoid reallocation of freed memory to a differently-typed object. Access to freed pointers can be caught by page table manipulation. Sure, there's an overhead to these techniques, but then there's a non-trivial overhead to GC as well.

I'm sure that somebody is going to dig up that paging-free GC paper, but pay attention: that is a kernel-level GC.

Which just indicates that there's been a lack of forethought in OS design resulting in the impossibility of implementing this at user level. I'm not sure, but pretty confident it could be done in user mode in a well-designed microkernel (e.g. L4). Not sure how relevant this is at this point, though. We've gone too far down the current line of OS development to just drop everything in favour of a new design just to be able to do a couple of neat tricks like this.

Complaint 3: Swing is ugly/leaks memory
    The first is a matter of opinion.

Yes, but it's a very popular opinion. Swing (and JDK) is also exceptionally slow. Most people don't realise it, but almost all of the slowness traditionally associated with Java can be attributed to those toolkits. Don't believe me? Try writing a program with another toolkit. I haven't tried it myself, but you could give wxjava a go.


Complaint 4: Bad build system
    Java cannot do incremental builds if class files have circular references. In a small project of about ten classes I was working on, the only way to build it was "rm *.class ; javac *.java"

Have you tried jikes? It has a superior incemental build system to sun's compiler. Not sure it'll fix your issue, but anything's possible.

You, sir, are a twit.

http://gdcmac.sourceforge.net/

As an update, I got further along this time, but compilation died when gcc tried to compile libphobos with the oddest autotools bug I've ever seen. Here's what I did:

Download and unpack gcc-4.0.3 into your personal src dir

Download the svn trunk of gdc from http://sourceforge.net/projects/dgcc

Go into gcc-4.0.3/gcc

$ cp -a ../../dgcc/trunk/d .

$ cd ..

$ gcc/d/setup-gcc.sh

$ mkdir work

$ cd work

$ ../configure --prefix=/usr/local/gcc-4.0.3 --enable-languages=c,d,c++ --host=x86_64-pc-linux-gnu --build=x86_64-pc-linux-gnu --target=x86_64-pc-linux-gnu --disable-multilib

edit ../gcc/d/d-lang.cc and change the following on line 268

#ifdef D_CPU_VERSYM
VersionCondition::addPredefinedGlobalIdent(D_CPU_V ERSYM);
#else
#define D_CPU_VERSYM "unknown"
#endif

Run "make"

Compilation dies on libphobos with the following peculiar error:

checking for nan... yes
checking for exp2... yes
checking for log2... yes
checking for fpclassify and signbit... yes
configure: error: Missing fpclassify and signbit
make: *** [configure-target-libphobos] Error 1

Well, I'm getting further along. It would be nice to get D working in gcc, no doubt its more usefull than Objective-C. lol.

Um, we do? Please email us, or have someone email us, if there's an entire country that Slashdot is blocking in some way. I think we provide a contact email address on all our "you've been blocked" messages, but just in case you didn't see one, try pater@slashdot.org. Let us know what the IP is and if there's been abuse, we'll handle it, well, pretty much the way Wikipedia did :)

As for not being able to log in, I think that might be something about your ISP. We block anonymous contributions from an IP number by setting it to 'nopost' or 'nopostanon'. The only place our login.pl code checks that is if you try to mail a password to someone (possibly yourself). We disallow that from nopost-blocked IPs, because it could be used for anonymous harrassment of our users.

So unless an IP number has been completely banned from Slashdot (no access to any pages, whether you're logged-in or not), you should be able to create an account and log in from it, as far as I know.

also wxD for wxWidgets.

http://wxd.sourceforge.net/

Its for situations like these that a few years ago I wrote pagefetch [sourceforge.net], a perl script that would retrieve an html page with its associated images, tar and gzip them all up, and email them back to the original requestor.

Looks like it's called DUI and it works on Linux and Windows.

DUI is D GUI

There is an entire market of novelty toys that play games like poker, sodoku and so forth.

From http://www.theory.physics.ubc.ca/transit.html:

The card requires firmware to be downloaded to the card first for it to work as a soundcard. While it appears on the usb bus with Vendor/ProductID of 0763/2806 this is a very primative usb device that does nothing except wait for firmware.

A firmware loader for M-Audio audio devices is available at http://usb-midi-fw.sourceforge.net/. Interestingly enough, the set-up procedure involves copying the firmware bin file from the Windows driver installation, which is subsequently used by the firmware loader.

Anyways, TFA makes some interesting points:

The firmware, although its being loaded to the device, is still firmware. Do distributions that demand an entirely free sofware environment ship drivers for devices with proprietary firmware? Of course they do. I'm not convinced that this should be any different.

I think that Theo de Raadt, of OpenBSD, has it right:

-azzurro

According to the article, there are three companies that have actually worked with the free software community on drivers. Here is the list:

Ralink Technology

Atmel Corporation

Realtek Linux drivers here

Vote with your money, folks. If you would like to see companies cooperate with the free software community, reward the companies that do so by buying their products.

If you know of a particular piece of WiFi hardware that works particularly well in Linux or BSD, please follow up here so we all know what to buy. (See also this list.)

No wonder there's only been one comment on this entry.

In all seriousness, though...why not just see if you could reach a service like Anonymouse to proxy it, or if you can get to your "American Web Hosting" you could set up your own php-based web proxy. Way easier and more elegant than all that other mumbo jumbo, plus you could reach any other site.

They can, but no one have financed it yet. BBCs Dirac is another project that shows real promisse: http://dirac.sourceforge.net/ and http://www.bbc.co.uk/opensource/projects/dirac/

The technology is there. It needs focus and financing to get real pollished.

From the linked site. It has been relatively easy to get .wmv, .mov, etc. working in Linux for quite some time now. Check out the MPlayer plugin for Firefox. For K/X/Ubuntu or other Debian-based distro users, "apt-get install mozilla-mplayer". I do agree, however, that all government websites should make their content available platform-independent. But then, that would require common-sense, now wouldn't it?

If the password is encoded within Java code, just use JAD for decompiling Java byte code into source code and then look for the password. There is even an Eclipse plugin called JadClipse, which make's password fishing even more fun. :)

Michael

Provide a scalable, manageable platform for collaboration and the development of Web-based business applications with Windows SharePoint Services 3.0, a versatile technology in Windows Server 2003.

And what that supposed to mean to the CIOs??? Or as developer and admin?

Well, we do not develop web services and heck I'm not sure what those are anyway. Last I heard Google doesn't use them - and it is playground of IBM/M$. From all it looks like: another non-technology crap (made up of buzz-words) created to sell something else. I have experienced many of such creations from M$ (and not only). And I'm sure many are yet to come: it is big companies, they need to sell something ... BIG.

As to corporate information sharing my employer suddenly went with Wikimedia's MediaWiki (the same one which powers Wikipedia). Meeting minutes, white papers, tech notes, sales remarks, bug analysis, etc - lands in Wiki were it can be viewed, corrected and extended by others. People especially liked the page audit & version control: that was one of the reasons for the Wiki to win against conventional content management systems. Big plus is of course that it runs on Linux so it can sit on our existing file/print/raid/svn/backup servers. Also, installation was no brainer with basically everything done by already prepared SUSE rpms.

It does actually work, but it lacks some features still...

• isync - Synchronize a local maildir with a remote IMAP4 mailbox
• mailsync - Synchronize IMAP mailboxes
• offlineimap - IMAP/Maildir synchronization and reader support

• isync - Synchronize a local maildir with a remote IMAP4 mailbox
• mailsync - Synchronize IMAP mailboxes
• offlineimap - IMAP/Maildir synchronization and reader support

Fish fillets is worth a try for those looking for a nice puzzle game. It is a well polished game with an interesting humour to it.

http://fillets.sourceforge.net/

http://vegastrike.sourceforge.net/

It's been eating up huge amounts of my time. While it's a bit confusing to learn and there is *a lot* of work to be done to the game balance, it's unbelievably fun still (and quite pretty).

If you like 3d space action/trading games, this one is great!

Google does not allow you to POP all your email.

It does seem as if Google has a time limit as to when you can download/POP mail. I tried to get all my mail from my old Gmail account onto my new hosted domain.

I could not enable POP for all the mail I ever received. I had to resort to libgmail and hacking its demo archive.py script to get all my mail.

The whole story is on my blog

Some Free games are very nice, specially the simple ones (not surprisingly, as they can't compete with the multi-million-dollars production effort of commercial games). The already mentioned Battle of Wesnoth and Freeciv are among my favorites strategy games. For action, Kobo Deluxe and Koules are pretty nice. Recently I found also XMoto, which looks silly, but is very fun to play.

I can play any game previously released (Though some DOS games seem like they are in super turbo mode)

For encoding video to my PSP, I use pspvc http://pspvc.sourceforge.net/ which is ffmpeg based, but is easier to use and "just works" You could also use it to encode for your iPod.

Virtual Dimension might be worth a look: a set of Windows WM enhancements. Open source and everything.

My experience with Ubuntu seesaws back and forth between pleasant surprise and absolute frustration. Installation was relatively painless and only a little complicated because I decided on a dual boot system. But, once I figured out I had to reformat the blank space I set aside for Ubuntu as ext3 the installer did its thing and I ended up on the Ubuntu desktop. Once I got over the turd brown color scheme and general unfamiliarity things were easy. Instead of /Users/Me it's /home/Me, but that's not a big deal. Got me a Desktop folder, so that's the same. Got /etc and all that jazz, with which I am familiar from OS X. Figure out that, as a GUI, Gnome is much more dependent on right click than OS X, so I go around right clicking on everything and find all sorts of useful options. Nautilus works pretty well, and I find a shell and fire up top to see what's going on deep down.

But it's slow. I mean the screen: redraw is slow and there's a lot of tearing, which surprises me, as I had expected the system to fly compared to OS X on six-year old hardware. I boot back into OS X and am surprised that Quartz is much faster on this machine than X/Gnome. Hmmm. . . After some googling I find out about glxgears, which runs somewhere around 40 fps. Not good. Some more Googling and a visit to Ubuntu's forums learns me that 1) the stock ati driver sucks ass and that 2) out-of-the-box video support on this machine is pretty bad, too.

So here's the frustrating part. What follows is about two weeks of googling and digging through Ubuntu's forums to find a series of solutions to make as full use of the machine's built-in card as possible (it's an 8 meg ATi Rage Mobility 128 AGP 2x). Eventually, through a combo of loading the right kernel modules and modifying xorg.conf, I am able to get direct rendering and full AGP support going, which results in a much, much faster desktop and glxgears scores of over 600 fps. But I also discover one of my big gripes with Linux: there is no centralized source for info. I found the answers my video questions in five different places. Now, the cool thing about this is that I learned some stuff, which I like. But I can't think helping that someone less inclined to root around in his machine's internals would've given up and ditched the OS. I know part of this is just a numbers problem: Apple has a very small number of machines to support using paid developers. The Linux community has a huge number of machines to support using, for the most part, people working for free and it's not surprising that some older hardware would be low on the support totem pole. But it is an annoyance.

I decide I want to be able to talk to my G5 running OS X from the Linux machine and vice versa. So, I turn on "Windows File Sharing" on the OS X machine, and share my home folder on my Powerbook. I am pleasantly surprised. Try to mount the Linux machine from the G5, but there's a problem: it's not taking the username and password. More googling reveals I have to 1) edit smb.conf and 2) set the Samba password from the shell. So, another few days of Googling and futzing are required for me to do something which can be done with three clicks in OS X. Frustration, but I eventually get it working. Yay me.

I like the Add/Remove software option, which makes finding programs a breeze. I like Update Notifier, which makes keeping the system up-to-date as easy as anything in OS X. I get comfortable with apt-get, which turns out to be pretty easy. And so I go along, merrily installing software. I find conky, which gives me a nice system monitor and practice installing software and using the shell to edit flat files, but of which are useful. I find that Ubuntu has a smaller memory footprint than OS X, which doesn't surprise me. And I dec

Dosbox is more fun.

given the choice between closed source drivers and none at all (a choice any Linux user makes when using 3D graphics acceleration)

I'm using Radeon 9000 with open drivers. It does everything I need. I'm not a gamer, but supertux runs at about 80 FPS which seems plenty for my needs. All video plays fine. glxgears runs at about 1400 FPS, but I don't use glxgears for anything :)

I know it's not the latest and greatest , but the r300 driver apparently has "works well, no lockups" support for Radeon 9600 and radeon X800. Sure, there are many cards and late model cards not supported, but that doesn't amount to closed source drivers or none for all linux users with 3D graphics acceleration.

and traffic jams at the SUMO site. You can also use their open source simulation software to create your on
traffic scenarios. I have always seen the creation of a traffic jam as a transition from a high density high flow state meta stable state to a high density low flow state. This can be expressed with a lambda shape curve in a density-flow diagram. The cause for exiting the meta stable state can be a
small disturbance, sometimes simulated by a random factor in CA traffic models, e.g. some guy braking without reason. In reality I don't think you can avoid the transition without maybe the help of computer guided cars.

Another useful Linux program that I forgot to mention is "CrossOver Linux." It is not free, but is a more user friendly version of WINE that is optimized to allow certain Windows office software programs to be run under Linux. It doesn't work with all versions of all Windows software, but in the past, I have used it to run Microsoft Office 97 and Office 2000 under Linux. I was using an old version of Red Hat Linux at the time, but it should also work with Ubuntu.

Cedega is another non-free version of WINE. It is designed to allow some Windows games to run under Linux. I am not a gamer and have not tried it, so I don't know much about it. I have heard that both products as well as the normal free version of WINE can all be installed at once without interfering with each other. If you know how to use old DOS commands and have some old DOS games that you want to play, you could download the Linux version of the free DOSBox program and run your old DOS games inside that.

One program that a desktop Linux user won't need is a virus scanner. There have never been any actively circulating Linux viruses, that is a Windows only problem. There are actually a few free and a few commercial virus scanners for Linux. If someone is running a Linux based mail server it might be useful to use one of those to protect your more vulnerable Windows clients from infected email attachments.

According to the article, there are three companies that have actually worked with the free software community on drivers. Here is the list:

Ralink Technology

Atmel Corporation

Realtek Linux drivers here

Vote with your money, folks. If you would like to see companies cooperate with the free software community, reward the companies that do so by buying their products.

If you know of a particular piece of WiFi hardware that works particularly well in Linux or BSD, please follow up here so we all know what to buy. (See also this list.)

steveha

For your publishing activities, you might like to install both Scribus and LyX in addition to the TeX and LaTeX you mention.

While the suggestion to buy a Mac is marked 'Funny', and was, I'm sure, intended to be such, it's actually not such a silly suggestion because Macs do run Linux very well, and if you find you don't like Linux, which while being superbly user friendly, it does tend to be somewhat pickey about the friendships it makes. If you and Linux just do not get on, you still have a very good piece of hardware and software in your possession. Macs will also run the software I have mentioned using the X-11 server from either Apple or Fink. That's in addition to all the proprietary software offered by Apple and their ISVs.

i-installer ftw. Sure, its not "right out of the box," but it might as well be considering its just a simple installer. For working with LaTeX, I've found TeXShop to be one of the best editors out there, and its Mac only.

2) One way to increase logging numbers is by making certain simple helpdesk tasks self-logging. For example, when a client wants their password changed, it's tempting for the helpdesk consultant to just change the password without ever opening a ticket. Why not write the password change utility so that it automatically opens a ticket, provides some minimal level of notes, and then presents this to the consultant?

I'll second that suggestion and add another: make an API that facilitates logging and merge that into your workflow. NOTE: This is all off the top of my head. I expect you'll tailor it to your specific needs. Augment as needed and/or time permits.

I'm thinking along the lines of wrapper functions that implement:

1. StartTicket implement a small program whose sole task is to log the start of a ticket. Record: Date, Time, Caller, Technician, Severity, Short Description

   The date and time can be captured automatically from the system. Ditto for the tech. That leaves gathering who called which could also be captured automatically from the Caller-ID info from your phone system. That leaves the Severity (Urgent! Important. When-you-get-a-chance) and Short Description.

   Another comment suggested carrying around a small voice recorder. With the increasing availability of IVR systems, even these could be captured with a small front-end that the person calling the help desk goes through. If techs were only permitted to work on calls that come through such a system, then everything you need is already there. Log the incoming call's audio as a BLOB in an RDBMS, use some Speech Recognition on that to get a text-formatted problem ticket. Sure, it's not going to be 100% accurate translation, but for now it's good (enough) to go. Get the minimum up and running Right Now. You can enhance, later.

2. LogAnActivity Simply record a text or voice update to the current task. Along with the current date, time, and tech.
3. StopTicket Again, implement a small wrapper program which captures the information you need. Date, Time, resolution, followup required.
4. Write a TASK utility that uses these wrappers.
   1. invokes the StartTicket wrapper,
   2. opens a new shell / window,
   3. tech uses shell to perform requested action; all shell output is logged to a log file whose name is based on the tech's name and the current date time in ISO-8601 format. Take a look at Date (Unix) for details. Don't have a unix date command on your system? Take a look at the GNU utilities for Win32. So, now you can construct an easily sortable date/time stamp;

      export right_now = $(date "+%%Y%%m%%dT%%H%%M%%SZ")
      
      or, under windows:
      
      SET UnxUtils=C:\TOOLS\UNXUTILS\USR\LOCAL\WBIN
      
      COPY CON: right_now.bat
      %UnxUtils%\date "+SET right_now=%%Y%%m%%dT%%H%%M%%SZ" > %TEMP%\right_now2.bat
      CALL %TEMP%\right_now2.bat
      ^Z
      
      right_now
      
      I just ran it on my PC and got:
      
      right_now=20061226T175320Z

   4. tech closes the shell / window
   5. on shell close, invoke the LogAnActivity wrapper
   6. invoke the StopTicket wrapper.
5. Permeate your workflow Leverage these wrappers as a framework and as it becomes clear, write other TASK wrappers as needed.

You are pressed for time right now, so this is going to need to start lean and simple. Just capture enough info to show that you are way too busy. Get some wiggle room from management.

Other Approach Rigorously provide what they request in the way of documentation adn logging!!! If you are short-staffed, then LET THAT FACT BE KNOWN! TANSTAAFL. U

There is already an open source attempt at reimplementing AmigaOS sort of: AROS. It's been around for years.

Thanks for the tip. Found it, but it's actually OpenAVS not CoreAVS at sf.net

http://sourceforge.net/project/showfiles.php?group _id=168676

PyObjC and RubyCocoa are both open source projects, and they will remain open. Cocoa-Java apps still run on 10.4, and will continue to do so, but new features will not be ported to Java.

PyObjC and RubyCocoa are both open source projects, and they will remain open. Cocoa-Java apps still run on 10.4, and will continue to do so, but new features will not be ported to Java.

You can find it right here: http://fsv.sourceforge.net/

Leaving aside the plausibility of a 12-year-old knowing Unix, simply knowing Unix is not enough to immediately use any application running on the system. Yes, she could probably have used vi on the security terminal. But the specialized security system would have required some learning time -- significant learning time if it were built on Unix, which has notoriously inconsistent user interface design and thus makes it harder to transfer skills from one application to the next.

Something to consider is that comparing Linux to consumer operating systems like Windows and Mac OS is like comparing apples to oranges; it's more appropriate to compare Linux to Darwin and Windows' kernel.

Why? One only needs to look at Gaim's download page to understand: http://gaim.sourceforge.net/downloads.php There are 8 different downloads for Linux-based operating systems.

I personally do not consider Linux to be a complete operating system; instead, I consider it an ingredient for use in creating an operating system. Likewise, Linux-based operating systems tend to be similar, thus allowing programs created for Red Hat to be easily ported to Fedora.

What does this mean: "Linux" can not be successful as a consumer OS, but one of the distributions can. The distributions can compete, and have high levels of compatibility, yet the consumer needs to be shielded from the complexity of the "Linux" brand. Personally, I think the most probable model to follow is to have a Linux-based "Dell OS", and a Linux-based "Compaq-OS". This is a similar strategy that Apple took when they brought BSD to the desktop with "Mac OS."

For example, Turbo Tax could run on Windows, Mac, "Dell OS", and "Compaq-OS". Stating that Turbo Tax runs on "Linux" is confusing because each Linux install has a differing set of APIs.

Sure. Check out OSS4LIB for a list of different open-source library systems. On the smaller end of the scale, OpenBiblio, PhpMyLibrary, and Emilda get mentioned a fair bit.

http://magpierss.sourceforge.net/

From reading the patent it seems to me Magpierrs is prior work.

::sigh:: people..

You know your statement is kinda hard to understand because of all the typeo's and gramer mistakes.. but I'll try to make it out.

I think you said, I cant use Wine because someone hasn't sold it as a product yet. ...yeah Its uhh.. painfully ovious that you dont know what your talking about. Allow me to illustrate.

You can just compile wine. You dont need it to be sold by corssover first. PPC is not he *only* hardware I own, its the only Apple hardware I own. I also own ix86, ix86_64, and Sparc. Did you know Mr. Wizard that you actually *can* run it on PPC vis Darwine? Wine on ppc screenshot. http://sourceforge.net/dbimage.php?id=54672

The reason Wine is faster is because it is not emulation, in fact wine is sometimes faster than Windows apps running on Windows XP. Its impossible for a VM to compare to that because VMWare will pass instructions directly to the real hardware underneath it still has the overhead of actually loading its own OS and Drivers in the VM. If you mean however "faster" not by how much of an impact it has on your system but just by user expirence then you may find all of them run Notepad with no signifiant slow downs. When you run more intensive tasks your milage may vary..

Wine has stability problems. (unless "guest" apps are recompiled.)
VMWare (workstation) and Paralles are only arround 80% native speed.
Xen requires modifications to the host. (and was Unix on Unix only till recently)
Boches very neat project can provide ix86 on just about any arch, but its slow.
(and the nameless other stuff I wont get into, QEMU, Virtual PC etc)

In many cases this would be the wrong aproach because if you want to provide applications to users a better aproach may be..
Citrix, X11, Terminal Server, or just write your apps in Java.

Like I said originally, I use VMware to do development and prototyping, so being able to copy snapshots and send them to other people via VMWare Player is a huge benifit and I am excited about it running on Mac's now.

To answer your question seriously, I think if you had a decent understanding of software development, you could use the OpenEEG project's schematics to build two of them, hook them up to a comp via whatever interface (can't recall, but I'm guessing it's serial or USB or such), write your software that takes inputs from the EEG, interprets them (so, you'd want to look for the band mentioned in the article, theta waves, associated with drowsy or relaxed states) and then send a command to an air blower or whatever (whatever method of propulsion you want to hook up to your tube with your ping pong ball or whatnot), which then moves the ball, according to the volume of your type of brain wave your program is looking for.
Here's a link to the OpenEEG project.

We (my family) played this at the MSI in Chicago...it's a great "game"....my years of meditation in college helped me beat everyone, I even went off the chart (at the low end) a few times.

I've been on the look-out for a consumer version to buy, this would be a cool home game, especially with the other EEG applications possible. I know there's an OpenEEG specification, but things still seem a little bit more expensive than I'd like...plus I don't want to electrocute myself or my family (at least, not yet).

The secret is that I reject all but a few hundred of those 11000 spams in SMTP envelope. Correspondents must have some form of id, currently one of:

1. a valid rDNS
2. a valid RFC 2822 HELO that resolves to connect IP
3. an RFC 4408 sender policy (SPF) with a PASS

That gets 3/4 of the garbage. Next, SPF FAIL is rejected, including for HELO. You'd be surprised at how much spam has my own domain for the HELO! For SPF SOFTFAIL, since the sender is requesting debugging info, I send a DSN to the purported sender reporting the SOFTFAIL. For senders with no SPF, I match domains with HELO and rDNS, and look at MX to try to get a match - which is then treated like and SPF pass. For SPF neutral, I do a CBV, and blacklist the sender if it fails.

This reduces the spam from 11000 to several hundred. The content filter is auto trained. A honeypot mailbox provides spam training. Messages from (verified by SPF PASS) senders that users reply to provide ham training. Users have a web interface to the quarantine.

The false positive from content filtering is extrememly low. The biggest problem is VIP correspondents with clueless email admins who are unwilling to educate or fire them. (E.g. one admin insisted I didn't know what I was talking about and "JUPITER" was a valid HELO name...) In these cases, I have extensions to the sendmail access database to provide policy exceptions. I can also provide local SPF records for correspondents to get them a PASS.

One customer had to resort to spamsoap.com because they were getting 2 million spam connection attempts a day, and my python based filter could only process 80000 or so on his 400Mhz server.

I'm not so sure what everyone is complaining about. I'm using SpamSieve as a plug in to Mail.app, and it catches just about everything without much in the way of training. Currently, my statistics as of 2006-11-01 say it's 97.1% accurate (with 71% of my total mail volume being spam, but that includes some legitimate marketing mail that I no longer really want, and I'm too lazy to track down the list maintainers), and that number gets higher every day.

On Windows, I'm using either Mozilla Thunderbird (usually), or SpamBayes as a plug-in to Outlook 2003 (when I have to), and I get similar results.

Of course, what we really need to do is rethink the way that the whole email system is designed, just in terms of MTAs that work separately from MDAs, etc. This kind of filtering really needs to take place at what we currently call the MTA level, with a configurable corpus for each user. The filtering should be done before the mail is permanently accepted, so that the impact on storage resources is as minimal as possible. Granted, it still takes a lot of processing power.

Another thing I need to spend some time thinking about is how RFC822 messages are structured in general. I'm just pulling this out of my ass right now, but the fact is that message envelopes are much to easy to spoof. Why have a separate message envelope to route the mail when the addressing information is already supposed to be contained in the headers? With the way spam is going, the message needs to be processed in its entirety in any case, so perhaps the envelope has outlived its usefulness?

I have noticed this as well and so have my friends and family. In fact, the number of daily spams caught and trashed by my Spam Bayes filter has nearly tripled in the last six months. The probable cause of this increase is a recent surge in the number of zombies now controlled by spam trojans in the bot networks. This was covered here on Slashdot last month in Bot Nets Behind Recent Spam Surge. As for the trusted email addresses, some of us are already doing this with whitelists, but as you say the good guys are losing right now. The one good thing, if you can call it that, that might come out of this whole scenario is that the spammers speed the coming of the day when classic e-mail is retired from general use and something better is put in its place. The greed of the spammers may ultimately prove to be their undoing as they collectively kill the goose that laid the golden eggs.

Mouse - two computers, one mouse? madness. Same for keyboards.