I've got a '486DX4-75 (NEC Versa 2000C); my wife has a P-120 (NEC Versa 4080H). The two of them handle primary and secondary DNS for several domains (yes, we have other secondaries that are geographically and topologically distant), as well as WWW and SMTP for those domains.
Both are a minimum of 5 years old now. Both were bought secondhand. When my Windows-using friends say things like "Well, I've got this old machine that's only a PentiumII-300," I just laugh.
Oy. 2 hit WTC, yes. 1 hit Pentagon. Pennsylvania, Pittsburgh and near Camp David were all the 4th plane, which had been rumored to be shot down, but US says they haven't shot down anybody. And nothing crashed at DCA - the one that hit the Pentagon had taken off from there.
Although I haven't done much in the way of free software, I've been involved in web development for years, and built things that a few thousand people a month seem to find informative or
useful or whatever. I am not at all offended by money, particularly since I've got to pay for this metered DSL connection...
Of course, in the free/open source software arena, popularity doesn't impose too great a burden on the developer (other than spikes in download traffic, and a higher volume of bug reports, fix/enhancement requests, etc.) Web stuff unfortunately costs more to run as it becomes more popular.:)
After having me (and my classmates) on-campus for a semester, the first university I attended decided that having an "Ethics in Technology" course, required for graduation, was a splendid idea.
I can't imagine where they got that idea... I mean, we were just traversing the 'net by telnetting from one Cisco to another (this was in the late '80s).
Anyway, I wound up not coming back the next semester (my grades weren't good... I flunked Freshman Orientation, which should tell you something) and started playing with computers for money instead. I guess maybe I've picked up some ethics along the way.
Actually, things like Magistr@ and this one, which send random files, can have some pretty nasty results. I get hundreds of copies of every trojan that comes along, since I manage a large mailing list with an average subscriber IQ of right around 100. Many, many times I've seen trojans pull random content out of random files to use as a "message body" - and wind up with stuff about bank accounts and whatnot.
Of course, this isn't as bad as plain ol' human stupidity, like the folks who mail me M$ sex-sells spreadshits showing all their employees' personal info including SSNs...
And my cow-orkers wonder why I'm so cynical about humans.
What other registrar can you think of that, a full month or more after you've transferred a domain elsewhere, would send you _postal mail_ to let you know that it's past due and you should send them money... and then a month after that, send you mail _again_, this time letting you know that this is your FINAL chance to send them money for it?
I've got a little '486 running Exim to handle mail for me and my wife. It checks the various MAPS services. The number of connections it refuses varies from month to month, but it averages out to about 60... each connection represents a spam that would have been sent to at least one address for at least one of us (and often multiple addresses).
So... yeah, it has an effect. No, it doesn't stop all the spam. But what's left is easier to deal with. --
This is quite right. I've had some dealings with a company that was contracted to develop a web-based transactional system for another company about four or five years back. Okay, they went out and got an SGI Challenge box. It ran okay for a while, then melted under an increasing transactional load. So they got a slightly bigger SGI box. Around that time, they picked up an application platform that happened to run on IRIX. Okay, cool. They wrote a massive application on that platform and ran it on whatever that box was.
Fast-forward a few years. The application platform went through a couple revisions, and the latest versions are no longer available for IRIX. Oops, they're now stuck with something that's at least 1 major revision old. And oh, the transaction load went up. A lot. And gee, the SGI product line is... how do I put this... sparse at the mid-to-upper range? (Speaking as a former IRIX admin:)
So they've now got this app running on a whole farm of Origin 2000's, which are distinctly not cheap, using an outdated version of an application platform. They pretty much locked themselves into a hardware solution *and* a software solution.
They could, of course, if they really wanted, ditch the software and replace it with, say, AOLserver (the solution they currently use is also Tcl-based, so that would probably be the easiest port). And if that step was successful, they could then port it over to commodity hardware.
Amusingly, I'm currently involved in building the system that's going to REPLACE the system I just described. It's using a newer version of the same software solution, and Sun hardware. Still kinda "locked in," but at least we're not paying extra premium prices for the privilege of having the letters "SGI" on the front of the boxes.:)
Oh, and the most experienced technical people on the project would all absolutely love to be doing it with BSD or GNU/Linux, AOLserver, PHP4, Zope, or stuff like that. But this is the corporate world, so that will have to wait. At least we get to code in Tcl instead of ASP or whatever.:)
--
A cluebie on a BBS recently asked me why anyone would pay to register Opera, when there are other browsers available for free. (Uh... maybe 'cos it's smaller, faster, supports more standards and is more stable?). I asked him why he paid for the OS his "free" browser (IE) runs on, when there are other OSes available for free.
If product A is free, and product B is not, going out and spending money on product B instead of using product A for free pretty much amounts to a donation. You're giving someone money that you don't have to give them.:)
I definitely agree that resting one's wrists on the part of the laptop case below the keyboard, or for that matter on just about anything, while typing, is a Bad Thing.
I find that (using a Dell Latitude CSx) my wrists are usually above or beyond the edge of that area while typing, though - the heel of my hand, or the pad at the base of my thumb, rests in that area. So I'm apparently not causing myself too much in the way of grievous harm.
This thread did make me realize that the squishy gel rest I've been using since before $COMPANY issued me the laptop is worse than useless in front of anything but a desktop keyboard.;)
Indeed, Jamie, it is important that large, heavily-visited sites remain accessible. I fully agree, and suspect most others would as well.
I am a little bit uncertain, though, about your positions on the following statements:
ISP's should not be allowed to block spam; it should be done individually by the user.
Even if small ISP's serving end-users are allowed to block spam, large ISP's that serve other ISP's should not be allowed to.
Large ISP's and backbones are common carriers that cannot block anything.
Only the specific IP address(es) from which spam originates should ever be blackholed.
Operators of large, popular or heavily-visited sites should be given the benefit of the doubt before being blackholed, more than smaller, less-known or less-used sites.
It seems that opposition to MAPS occurs primarily when they block people's personal favorite sites, and never any other time. No one RBL'ed has tried to mail me for months, but this past winter there were attempts from dot.net.in, glptt.gx.cn, and insight.com. I don't recall hearing you, Bennett, or anyone else coming to the defense of *those* domains; in fact, everyone seemed to be pretty happy about not getting spam from them.
I am unclear why Macromedia deserves to be treated differently if they spam than Insight was, or the others.
Maybe you could explain your thoughts on that, and on the position statements I listed.
Thanks!
To the best of my knowledge (as someone who's read a lot that Bennett has had to say about this, and a lot that other people have had to say as well), the chronology was like this:
Peacefire got hosted by Media3 in netblock A.
Media3 netblock B hosted a bunch of spammers.
MAPS received complaints about netblock B.
Media3 *moved* Peacefire from A to B.
MAPS blocked netblock B due to the spammers.
Step 4 above is roughly akin to surrounding your biological-weapons plant with women and children so you can claim it was a school, hospital or residential neighborhood when it gets blown up.
There are some things I do not know, namely:
Whether MAPS and Media3 were in communication before Peacefire was moved. (In other cases I'm aware of, MAPS has seemed to respond to spam complaints with almost glacial speed, so there may have been some bureaucratic back-and-forth in the weeks or months preceding the RBL listing.)
If so, whether Media3 made Peacefire aware that the netblock into which they were moved was likely to be RBLed.
I find it hard to believe that Media3 wouldn't have known they were likely to be RBLed for hosting so many spammers in that block. From what I've heard over the years, an RBL nomination is unlikely to be accepted if the person submitting it hasn't already communicated with the target without success.
If Media3 did know (which seems likely), and failed to tell Peacefire, Bennett has every right to be pissed - at Media3. If someone pushes you in front of a bus, is the bus company liable? No, the person who pushed you is.:)
If Media3 did know and told Peacefire, and Peacefire willingly consented to be placed in a netblock that was likely to be RBL'ed, it becomes a manner of "why the heck would you do that? *slap slap slap*"
Come to think of it, that's pretty much the reaction Bennett got on comp.dcom.telecom...
Oh, and Ethereal, I should note that I agree wholeheartedly that purely technical measures such as the RBL are not going to solve a largely social problem like spam.
The RBL is so infrequently applied that it really doesn't stop much spam for me.
The RSS (Relay Spam Stopper) stops _27 times_ more spam than the RBL for me, and the DUL (Dial-Up List) stops _31 times_ more spam than the RBL.
Given how *few* messages the RBL blocks, I don't know whether I'd describe it as "out of hand." I'd be inclined to agree that it's certainly the biggest stick MAPS has - but in terms of raw numbers _for a given recipient_ it doesn't seem to be the most effective.
(Mathematically, folks who wind up on the RBL are probably causing grief to a lot more recipients than the average two-bit relay-raping or direct-from-dialup spammer, of course...;)
Given the low amount of mail the RBL stops, I I may tell Exim to use it in "warn" mode, like I do with ORBS, rather than "reject" mode. That is to say, don't block it, but let me know that it's suspect.
The HTTP stuff that got disrupted was in the same/24 (256-address block) as the mailserver(s) the spam came from.
There are... relatively few people, MAPS or otherwise, who still think that "oh, we'll just block that IP" works.
It's nearly trivial for the spammer to change IP's within a single block of addresses. If the block is blocked, it's harder.
I believe someone already pointed out that Macromedia had their spam server, their web server *and* their DNSes in the same network block.
Do we need to mention Microsoft's habit of having all the DNSes for one of its domains in the same netblock, and the result that had a few months back, again?
"Single Point of Failure" comes to mind. "All your eggs in one basket" comes to mind. Heck, I run my domain off two obsolete laptops at the end of a DSL line, and even *I* have geographically and topologically distant backup DNS.:)
Anyway, MAPS tends to block/24's (thus Peacefire's problem) from which spam originates, _if_ the folks who run them refuse to do anything about it. And they don't tend to act on complaints unless there's a clear indication that the person making the complaint has exhausted all other channels. That's all old news.
Peacefire's offers of *free* hosting elsewhere were not coming from the same people who were blacklisting them. They were coming from other people who said "gee, that really sucks, it's a pity you got blacklisted, and it's a real shame it happened 'cos your own provider is so spam-friendly. Here, we'll give you a free connection on a network that isn't full of spammers." Had Peacefire moved, it would have been a *good* thing, both in terms of making their site more accessible, and giving the spam-hosting folks one less thing to point to and say "but look, we host this GOOD PUBLIC SERVICE site!" Everyone involved in the Usenet threads fully understands this.
Being fairly familiar with various parties mentioned, and having seen far more of Peacefire's attempt at a smear campaign against MAPS on Usenet (and the rebuffing it got from all manner of administrative sorts not associated with MAPS) I can say with very little doubt that Inquisitor has the story at least 99.999% right, and probably 100%.
Peacefire's whining over their provider getting RBL'ed for being a pit of spammers, and more importantly their refusal to change providers, even though they were offered free hosting other places, made it quite clear that they're not interested in actually getting along with folks, and merely want to be seen as some sort of martyrs.
Merely mentioning Peacefire made it clear from the outset that this wasn't going to be an objective story.
Censorship is bad. Spam is bad. Censorship is content-driven. The RBL and things like that are not content-driven. They are activity- and method-driven. Censorship is oppressive. Is the RBL oppressive? Sure. Is spam oppressive? Every bit as much as the RBL is.
Anarchy is not about getting rid of rules - it is about getting rid of oppression.
After NT 4.0 committed suicide by knifing its own bootloader a few weeks back, I decided that trusting Microsoft products with *physical* partitions was a Bad Thing, and snagged VMWare. Here, NT, go play inside a file under Linux!
Things are much better now. No worrying about OSes trampling the MBR. Hopefully VMWare (or Bochs, *plex, other similar free/open source projects) will continue getting updated to support new versions of Windows.
It's easy for NASA to get all the funding they need. All they have to do is tell the Dick and the Bush that the latest information from their probes indicates that there is oil - enough of it to last earth thousands of years - on Mars.
Wham-O, there's your funding for manned missions to Mars, and so on and so forth.
And for our oil-industry nerds, please note that I'm not badmouthing you. Just two particular people who have erroneously been put in control of lots of big red buttons and stuff.:)
Hmmm... maybe I meant "Nasa Funding From Dummies." --
Bring aboard some folks who're not only used to writing good maintainable code, but used to writing good maintainable code that compiles on a whole bunch of architectures, and you might be able to give "the community" more than just a couple paychecks.
If your codebase becomes more portable, the expense and difficulty of ports to new platforms is greatly reduced. How many games does Midway offer for Linux these days? If new games could be offered for Linux without having to move mountains, would they be?
Of course, I'm saying this out of personal bias, since I loved the original "Gauntlet" and would gladly pay good money if "Gauntlet: Dark Legacy" could be had for Linux...
If you check various *.binaries.* groups - and I don't mean just alt.binaries.* either! - you'll see "This group is no longer archived."
I can understand this - binaries are high-bandwidth. Though since Google strips out the encoded binary, like Deja did, not archiving the remaining text has got to yield some seriously diminishing returns.
More interesting (and baffling/troubling, IMO) is the rather... selective approach Google appears to have taken with regard to other alt groups, particularly in certain hierarchies. Want to read about bondage? Okay, they archive alt.sex.bondage... but not alt.sex.stories.bondage. Into animals? alt.sex.bestiality.hamster.duct-tape is yours for the browsing, but not alt.sex.hedgehog.ouch.ouch.ouch. alt.alien.vampire.flonk.flonk.flonk is okay, but alt.alien.visitors is no longer archived. alt.rock-n-roll.symphonic is mysteriously no longer archived, while other groups in that sub-hierarchy are.
I wasn't able to find anything in the Google Groups help explaining what their criteria are for deciding that a group should no longer be archived.
Better suited to "internal" applications than external ones. I would love to use this to serve up charts of web server log analysis, instead of using WebTrends (*spit*). I'm sure the bean-counters would adore the ability to "fly through" the charts.;)
Mentioning Tcl and CURL in the same posting is likely to cause nasty twitches in those of us who're programming in Vignette V/5 (which is Tcl-based and has a CURL command) at the moment.:)
This is the sort of the flip side of a question I was asked elsewhere a few months back - what forms of information will the Internet obsolete and ultimately destroy?
The folks that raised that question were actually asking about newspapers in particular, but it applies to a few other forms as well.
I'll preface this by saying that I've worked in reference at a library, that I've written for a newspaper, that a 'zine I edit on-line turns 10 this year, and that I have every single back-issue available.
The category of information I have the most trouble finding on the 'net, period - not a week goes by that I fail to find something in this category - is "old media stuff." Old here basically means anything prior to 1990.
There are some newspapers that have been on-line for a few years, and have done a good job of archiving their material in that time. There may even be a few (nytimes.com perhaps?) that have managed to put older material up in electronic form. But it's probably pay content, and it's definitely not something I can find in a search engine.
Magazines are no better, and probably worse. Newspapers at least make some effort to put all their content on-line; magazines are obsessed with the idea of teasing you into buying hardcopy. Even OLD articles, no go. A friend referenced an article in a year-old Utne Reader, and I'll have to go to the library for that one. I wonder if they're selling enough back-issues to make that sort of ploy viable. I doubt it.
And those are the exceptions. Most media sites on the Internet aren't designed with serious long-term archive availability in mind. That's just not the way they're thinking. Archives are for dead-tree formats, microfilm or microfiche or maybe CD-ROM if you're lucky. Not on-line.
Needless to say, my answer to the folks who were concerned about the 'net killing newspapers had a lot to do with it not happening until the 'net folks start taking archival more seriously. On the flip side, the 'net will have a whole lot of informational holes in it until that happens.
NEW technologies? Doesn't have to be NEW not to work. Folks have had plenty of time to implement decent support for standards like HTML (2.0, 3.2, 4.0), CSS (1,2), Posix, etc. Plenty of broken cruft floating around there - and as much as I loathe M$, they by no means have a monopoly on broken cruft in _those_ departments. (Just in ones like, let's see, Java support, Kerberos... I'm sure I missed a few.;)
The company I work for has a mailing list its customers can subscribe to. Guess who's the lucky guy who watches over it? Yup, you got it...
Anyway, the marketeers like to be able to track ROI and whatnot. So a little bit back, we started sending multipart MIME messages, and including the company logo in the HTML version. Result: the ability to tell them "Okay, in the first 7 days after the mailing, N people opened the message with HTML-capable mailreaders while online." Obviously, the actual number of people reading it is greater, but these days, probably not by much.
Around that same time, I modified the Perl stuff that sends the mailings to stick a query string on those images, i.e. "/hdr1.gif?7kdtP-SeV" or whatever, populating it with an encoded version of a string containing stuff like the date it was sent, the filename of the message that was sent, and the registered userid (on our site) the address corresponded to.
On the back end, more Perl looks at various and sundry logs, and goes through the process of "Hey! CMDRTACO read the e-mail. Hmmm. CMDRTACO clicked through to the site from the e-mail. Hmmm. CMDRTACO logged into the site. Hey cool, CMDRTACO bought something, cha-ching!" and so on.
I'm actually doing some finessing today to automate things a bit. Perl hacking, fun fun fun.
--
If Linux is going to compete against MS on the desktop market, it needs to be easier to install/configure. Once Linux is installed and configured, it tends to be at least as easy to use as anything else out there, if not easier, simply because it doesn't go belly-up and demand to be reinstalled, used in safe mode, or whatever. A machine with Linux properly pre-loaded by the OEM is no doubt a joy to behold. --
Slashdot Mirror
Delete Windows, install *Linux or *BSD.
I've got a '486DX4-75 (NEC Versa 2000C); my wife has a P-120 (NEC Versa 4080H). The two of them handle primary and secondary DNS for several domains (yes, we have other secondaries that are geographically and topologically distant), as well as WWW and SMTP for those domains.
Both are a minimum of 5 years old now. Both were bought secondhand. When my Windows-using friends say things like "Well, I've got this old machine that's only a PentiumII-300," I just laugh.
Oy. 2 hit WTC, yes. 1 hit Pentagon. Pennsylvania, Pittsburgh and near Camp David were all the 4th plane, which had been rumored to be shot down, but US says they haven't shot down anybody. And nothing crashed at DCA - the one that hit the Pentagon had taken off from there.
Of course, in the free/open source software arena, popularity doesn't impose too great a burden on the developer (other than spikes in download traffic, and a higher volume of bug reports, fix/enhancement requests, etc.) Web stuff unfortunately costs more to run as it becomes more popular. :)
No can do... We already named her after an old brand of supercomputers. Maybe the second-born, if there's one...
I can't imagine where they got that idea... I mean, we were just traversing the 'net by telnetting from one Cisco to another (this was in the late '80s).
Anyway, I wound up not coming back the next semester (my grades weren't good... I flunked Freshman Orientation, which should tell you something) and started playing with computers for money instead. I guess maybe I've picked up some ethics along the way.
Maybe.
--
Of course, this isn't as bad as plain ol' human stupidity, like the folks who mail me M$ sex-sells spreadshits showing all their employees' personal info including SSNs...
And my cow-orkers wonder why I'm so cynical about humans.
--
What other registrar can you think of that, a full month or more after you've transferred a domain elsewhere, would send you _postal mail_ to let you know that it's past due and you should send them money... and then a month after that, send you mail _again_, this time letting you know that this is your FINAL chance to send them money for it?
Greedy fraudulent bastards.
(2 domains now with OpenSRS, 1 with Qroute)
--
I've got a little '486 running Exim to handle mail for me and my wife. It checks the various MAPS services. The number of connections it refuses varies from month to month, but it averages out to about 60... each connection represents a spam that would have been sent to at least one address for at least one of us (and often multiple addresses).
So... yeah, it has an effect. No, it doesn't stop all the spam. But what's left is easier to deal with.
--
Fast-forward a few years. The application platform went through a couple revisions, and the latest versions are no longer available for IRIX. Oops, they're now stuck with something that's at least 1 major revision old. And oh, the transaction load went up. A lot. And gee, the SGI product line is... how do I put this... sparse at the mid-to-upper range? (Speaking as a former IRIX admin :)
So they've now got this app running on a whole farm of Origin 2000's, which are distinctly not cheap, using an outdated version of an application platform. They pretty much locked themselves into a hardware solution *and* a software solution.
They could, of course, if they really wanted, ditch the software and replace it with, say, AOLserver (the solution they currently use is also Tcl-based, so that would probably be the easiest port). And if that step was successful, they could then port it over to commodity hardware.
Amusingly, I'm currently involved in building the system that's going to REPLACE the system I just described. It's using a newer version of the same software solution, and Sun hardware. Still kinda "locked in," but at least we're not paying extra premium prices for the privilege of having the letters "SGI" on the front of the boxes. :)
Oh, and the most experienced technical people on the project would all absolutely love to be doing it with BSD or GNU/Linux, AOLserver, PHP4, Zope, or stuff like that. But this is the corporate world, so that will have to wait. At least we get to code in Tcl instead of ASP or whatever. :)
--
If product A is free, and product B is not, going out and spending money on product B instead of using product A for free pretty much amounts to a donation. You're giving someone money that you don't have to give them. :)
--
I find that (using a Dell Latitude CSx) my wrists are usually above or beyond the edge of that area while typing, though - the heel of my hand, or the pad at the base of my thumb, rests in that area. So I'm apparently not causing myself too much in the way of grievous harm.
This thread did make me realize that the squishy gel rest I've been using since before $COMPANY issued me the laptop is worse than useless in front of anything but a desktop keyboard. ;)
--
I am a little bit uncertain, though, about your positions on the following statements:
- ISP's should not be allowed to block spam; it should be done individually by the user.
- Even if small ISP's serving end-users are allowed to block spam, large ISP's that serve other ISP's should not be allowed to.
- Large ISP's and backbones are common carriers that cannot block anything.
- Only the specific IP address(es) from which spam originates should ever be blackholed.
- Operators of large, popular or heavily-visited sites should be given the benefit of the doubt before being blackholed, more than smaller, less-known or less-used sites.
It seems that opposition to MAPS occurs primarily when they block people's personal favorite sites, and never any other time. No one RBL'ed has tried to mail me for months, but this past winter there were attempts from dot.net.in, glptt.gx.cn, and insight.com. I don't recall hearing you, Bennett, or anyone else coming to the defense of *those* domains; in fact, everyone seemed to be pretty happy about not getting spam from them.I am unclear why Macromedia deserves to be treated differently if they spam than Insight was, or the others.
Maybe you could explain your thoughts on that, and on the position statements I listed. Thanks!
--
- Peacefire got hosted by Media3 in netblock A.
- Media3 netblock B hosted a bunch of spammers.
- MAPS received complaints about netblock B.
- Media3 *moved* Peacefire from A to B.
- MAPS blocked netblock B due to the spammers.
Step 4 above is roughly akin to surrounding your biological-weapons plant with women and children so you can claim it was a school, hospital or residential neighborhood when it gets blown up.There are some things I do not know, namely:
- Whether MAPS and Media3 were in communication before Peacefire was moved. (In other cases I'm aware of, MAPS has seemed to respond to spam complaints with almost glacial speed, so there may have been some bureaucratic back-and-forth in the weeks or months preceding the RBL listing.)
- If so, whether Media3 made Peacefire aware that the netblock into which they were moved was likely to be RBLed.
I find it hard to believe that Media3 wouldn't have known they were likely to be RBLed for hosting so many spammers in that block. From what I've heard over the years, an RBL nomination is unlikely to be accepted if the person submitting it hasn't already communicated with the target without success.If Media3 did know (which seems likely), and failed to tell Peacefire, Bennett has every right to be pissed - at Media3. If someone pushes you in front of a bus, is the bus company liable? No, the person who pushed you is. :)
If Media3 did know and told Peacefire, and Peacefire willingly consented to be placed in a netblock that was likely to be RBL'ed, it becomes a manner of "why the heck would you do that? *slap slap slap*"
Come to think of it, that's pretty much the reaction Bennett got on comp.dcom.telecom...
--
The RBL is so infrequently applied that it really doesn't stop much spam for me.
The RSS (Relay Spam Stopper) stops _27 times_ more spam than the RBL for me, and the DUL (Dial-Up List) stops _31 times_ more spam than the RBL.
Given how *few* messages the RBL blocks, I don't know whether I'd describe it as "out of hand." I'd be inclined to agree that it's certainly the biggest stick MAPS has - but in terms of raw numbers _for a given recipient_ it doesn't seem to be the most effective.
(Mathematically, folks who wind up on the RBL are probably causing grief to a lot more recipients than the average two-bit relay-raping or direct-from-dialup spammer, of course... ;)
Given the low amount of mail the RBL stops, I I may tell Exim to use it in "warn" mode, like I do with ORBS, rather than "reject" mode. That is to say, don't block it, but let me know that it's suspect.
--
There are... relatively few people, MAPS or otherwise, who still think that "oh, we'll just block that IP" works.
It's nearly trivial for the spammer to change IP's within a single block of addresses. If the block is blocked, it's harder.
I believe someone already pointed out that Macromedia had their spam server, their web server *and* their DNSes in the same network block.
Do we need to mention Microsoft's habit of having all the DNSes for one of its domains in the same netblock, and the result that had a few months back, again?
"Single Point of Failure" comes to mind. "All your eggs in one basket" comes to mind. Heck, I run my domain off two obsolete laptops at the end of a DSL line, and even *I* have geographically and topologically distant backup DNS. :)
Anyway, MAPS tends to block /24's (thus Peacefire's problem) from which spam originates, _if_ the folks who run them refuse to do anything about it. And they don't tend to act on complaints unless there's a clear indication that the person making the complaint has exhausted all other channels. That's all old news.
Peacefire's offers of *free* hosting elsewhere were not coming from the same people who were blacklisting them. They were coming from other people who said "gee, that really sucks, it's a pity you got blacklisted, and it's a real shame it happened 'cos your own provider is so spam-friendly. Here, we'll give you a free connection on a network that isn't full of spammers." Had Peacefire moved, it would have been a *good* thing, both in terms of making their site more accessible, and giving the spam-hosting folks one less thing to point to and say "but look, we host this GOOD PUBLIC SERVICE site!" Everyone involved in the Usenet threads fully understands this.
--
Peacefire's whining over their provider getting RBL'ed for being a pit of spammers, and more importantly their refusal to change providers, even though they were offered free hosting other places, made it quite clear that they're not interested in actually getting along with folks, and merely want to be seen as some sort of martyrs.
Merely mentioning Peacefire made it clear from the outset that this wasn't going to be an objective story.
Censorship is bad. Spam is bad. Censorship is content-driven. The RBL and things like that are not content-driven. They are activity- and method-driven. Censorship is oppressive. Is the RBL oppressive? Sure. Is spam oppressive? Every bit as much as the RBL is.
Anarchy is not about getting rid of rules - it is about getting rid of oppression.
--
Things are much better now. No worrying about OSes trampling the MBR. Hopefully VMWare (or Bochs, *plex, other similar free/open source projects) will continue getting updated to support new versions of Windows.
--
Wham-O, there's your funding for manned missions to Mars, and so on and so forth.
And for our oil-industry nerds, please note that I'm not badmouthing you. Just two particular people who have erroneously been put in control of lots of big red buttons and stuff. :)
Hmmm... maybe I meant "Nasa Funding From Dummies."
--
If your codebase becomes more portable, the expense and difficulty of ports to new platforms is greatly reduced. How many games does Midway offer for Linux these days? If new games could be offered for Linux without having to move mountains, would they be?
Of course, I'm saying this out of personal bias, since I loved the original "Gauntlet" and would gladly pay good money if "Gauntlet: Dark Legacy" could be had for Linux...
--
I can understand this - binaries are high-bandwidth. Though since Google strips out the encoded binary, like Deja did, not archiving the remaining text has got to yield some seriously diminishing returns.
More interesting (and baffling/troubling, IMO) is the rather... selective approach Google appears to have taken with regard to other alt groups, particularly in certain hierarchies. Want to read about bondage? Okay, they archive alt.sex.bondage... but not alt.sex.stories.bondage. Into animals? alt.sex.bestiality.hamster.duct-tape is yours for the browsing, but not alt.sex.hedgehog.ouch.ouch.ouch. alt.alien.vampire.flonk.flonk.flonk is okay, but alt.alien.visitors is no longer archived. alt.rock-n-roll.symphonic is mysteriously no longer archived, while other groups in that sub-hierarchy are.
I wasn't able to find anything in the Google Groups help explaining what their criteria are for deciding that a group should no longer be archived.
--
--
The folks that raised that question were actually asking about newspapers in particular, but it applies to a few other forms as well.
I'll preface this by saying that I've worked in reference at a library, that I've written for a newspaper, that a 'zine I edit on-line turns 10 this year, and that I have every single back-issue available.
The category of information I have the most trouble finding on the 'net, period - not a week goes by that I fail to find something in this category - is "old media stuff." Old here basically means anything prior to 1990.
There are some newspapers that have been on-line for a few years, and have done a good job of archiving their material in that time. There may even be a few (nytimes.com perhaps?) that have managed to put older material up in electronic form. But it's probably pay content, and it's definitely not something I can find in a search engine.
Magazines are no better, and probably worse. Newspapers at least make some effort to put all their content on-line; magazines are obsessed with the idea of teasing you into buying hardcopy. Even OLD articles, no go. A friend referenced an article in a year-old Utne Reader, and I'll have to go to the library for that one. I wonder if they're selling enough back-issues to make that sort of ploy viable. I doubt it.
And those are the exceptions. Most media sites on the Internet aren't designed with serious long-term archive availability in mind. That's just not the way they're thinking. Archives are for dead-tree formats, microfilm or microfiche or maybe CD-ROM if you're lucky. Not on-line.
Needless to say, my answer to the folks who were concerned about the 'net killing newspapers had a lot to do with it not happening until the 'net folks start taking archival more seriously. On the flip side, the 'net will have a whole lot of informational holes in it until that happens.
-Dan
--
NEW technologies? Doesn't have to be NEW not to work. Folks have had plenty of time to implement decent support for standards like HTML (2.0, 3.2, 4.0), CSS (1,2), Posix, etc. Plenty of broken cruft floating around there - and as much as I loathe M$, they by no means have a monopoly on broken cruft in _those_ departments. (Just in ones like, let's see, Java support, Kerberos... I'm sure I missed a few. ;)
--
Anyway, the marketeers like to be able to track ROI and whatnot. So a little bit back, we started sending multipart MIME messages, and including the company logo in the HTML version. Result: the ability to tell them "Okay, in the first 7 days after the mailing, N people opened the message with HTML-capable mailreaders while online." Obviously, the actual number of people reading it is greater, but these days, probably not by much.
Around that same time, I modified the Perl stuff that sends the mailings to stick a query string on those images, i.e. "/hdr1.gif?7kdtP-SeV" or whatever, populating it with an encoded version of a string containing stuff like the date it was sent, the filename of the message that was sent, and the registered userid (on our site) the address corresponded to.
On the back end, more Perl looks at various and sundry logs, and goes through the process of "Hey! CMDRTACO read the e-mail. Hmmm. CMDRTACO clicked through to the site from the e-mail. Hmmm. CMDRTACO logged into the site. Hey cool, CMDRTACO bought something, cha-ching!" and so on.
I'm actually doing some finessing today to automate things a bit. Perl hacking, fun fun fun.
--
If Linux is going to compete against MS on the desktop market, it needs to be easier to install/configure. Once Linux is installed and configured, it tends to be at least as easy to use as anything else out there, if not easier, simply because it doesn't go belly-up and demand to be reinstalled, used in safe mode, or whatever. A machine with Linux properly pre-loaded by the OEM is no doubt a joy to behold.
--