Linux-Advocacy Digest #63, Volume #31 Tue, 26 Dec 00 12:13:03 EST
Contents:
Re: Red hat becoming illegal? (Chris Ahlstrom)
Re: linux does NOT suck (oh yes it does) (Chris Ahlstrom)
Re: Question with Security on Linux/Unix versus Windows NT/2000 ("sandrews")
Re: Conclusion ("Adam Ruth")
Re: Question with Security on Linux/Unix versus Windows NT/2000 ("sandrews")
Re: Conclusion (The Ghost In The Machine)
Re: Conclusion ("Adam Ruth")
Re: Uptimes (R.E.Ballard ( Rex Ballard ))
Re: linux does NOT suck (oh yes it does) ("Ayende Rahien")
Re: New to Linux, and I am not satisfied. (Bob Hauck)
----------------------------------------------------------------------------
From: Chris Ahlstrom <[EMAIL PROTECTED]>
Crossposted-To:
alt.destroy.microsoft,comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy,alt.rush-limbaugh
Subject: Re: Red hat becoming illegal?
Date: Tue, 26 Dec 2000 15:12:15 GMT
DishDude wrote:
> > >
> > > > I find Al Franken far more irritating than Rush. Mostly because he is an
> > > > painfully non-funny comedian. He, David Brenner, and Garry Shandling set
> > > > my teeth on edge.
>
> You can add Dennis Miller to that list as well, AFAIAC. :)
Aw, man, Dennis was in rare form on MNF last night. "Tackle by Claude Rains",
that was really a good one. Trouble is, people don't get Dennis's jokes
because the references are older than N'Sync.
> > Some of Al Franken's early stuff on SNL was pretty good, i'll admit that.
> > You'll never convince me that Brenner or Shandling have any talent, though.
>
> Agreed. Typically whiny humor that only New Yawkas/Yiddish really
> appreciate. Doesn't translate well to the rest of the country though.
Get literate, buddy. Read some books and get some learning.
You'll find life more enjoyable.
Chris
------------------------------
From: Chris Ahlstrom <[EMAIL PROTECTED]>
Crossposted-To: alt.linux.sux
Subject: Re: linux does NOT suck (oh yes it does)
Date: Tue, 26 Dec 2000 15:15:51 GMT
MH wrote:
>
> >
> > But the desktop is all you're concerned about, right?
> >
> > Well, have patience - things are developing, but these
> > things take time. Give it say 5 years, and then we'll see.
>
> God, where have we heard this before?
Microsoft???
>
> As for your rant, does the word "Mindcraft" have any meaning to you?
> Reputable benchmarking, look into it.
That stuff is ollllllllllllllld and now obsolete.
------------------------------
From: "sandrews" <[EMAIL PROTECTED]>
Subject: Re: Question with Security on Linux/Unix versus Windows NT/2000
Date: Tue, 26 Dec 2000 10:27:07 +0500
In article <OjZ16.177$[EMAIL PROTECTED]>, "Erik Funkenbusch"
<[EMAIL PROTECTED]> wrote:
> "mlw" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
>> Erik Funkenbusch wrote:
>> >
>> > "mlw" <[EMAIL PROTECTED]> wrote in message
>> > news:[EMAIL PROTECTED]...
>> > > Erik Funkenbusch wrote:
>> > > >
>> > > > "Philip Neves" <[EMAIL PROTECTED]> wrote in message
> news:Xu_06.862967
>> > > > > As of right now There are no viruses that I know of for
>> > > > > linux.
>> > > >
>> > > > Guess you've never heard of Bliss:
>> > > >
>> > > > http://math-www.uni-paderborn.de/~axel/bliss/
>> > > >
>> > > > It's been around for a while. Not to mention that there have
>> > > > been virus-like things, such as the morris internet worm.
>> > > >
>> > > > > I've been using linux for five years and I have never heard
>> > > > > of
> one.
>> > > >
>> > > > I guess ignorance, is Bliss.
>> > >
>> > > The thing you are missing is that most NT users must operate with
>> > > OS privileges which would allow a virus to spread
>> >
>> > That's not true. NT is perfectly capapble of being used in locked
>> > down
> way.
>>
>> This is clearly not 100% true.
>
> It is 100% true.
>
>> As I have said before:
>>
>> Some Windows programs can not install for administrator and also put
>> their settings into a user's account. (You must know this if you use
>> NT.) They don't know how. So, these programs must be installed as the
>> user who will be using them. To do this, you must have the privileges
>> which would allow a virus to spread.
>
> There are any number of solutions to this problem. 1) give the user
> account temporary admin privs, install the program, remove those
> privs. Problem solved. 2) Using Win2k, simply install the program
> with admin privs as the dialog box that pops up allows. 3) Monitor
> the keys created via any number of tools, then recreate those same
> keys in each user hive.
>
> There are other solutions, but those are just off the top of my head.
>
>> Some Windows programs assume system wide access to operate. To
>> operate these programs you must also have privileges which would
>> allow a virus to spread.
>
> Name one. The only one I know of that requires this out of the box is
> Office 95, and that can be worked around pretty easily.
>
Borlands Delphi versions 3,4 and 5.
>> The answer is to not use Windows NT for these programs OR use NT in
>> an insecure way. Most users will choose to use NT in an insecure way.
>
> No, the answer is to use it the way it's intended to be used.
>
>
>
>
------------------------------
From: "Adam Ruth" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: Conclusion
Date: Mon, 25 Dec 2000 09:35:51 -0700
You're making the same mistake you've been making all along. It's getting
NO uptime from you so you can't use yourself as an example. If what happens
in your situation happens "in many cases", then we'd see a lot of sites with
no uptime reported. What site does
it pull the OS from the webserver and the uptime from the firewall? You
claim this happens but you haven't given an example. In your case Netcraft
knows that the firewall isn't providing the data correctly or isn't
providing the data at all (we're not privy to what they're getting, only
their interpretation).
Show a concrete example where this happens, not the theoretical possibility.
Adam Ruth
"Erik Funkenbusch" <[EMAIL PROTECTED]> wrote in message
news:DFE16.142$[EMAIL PROTECTED]...
> "Jim Richardson" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > If the OS is detected correctly, and the uptime returned for that system
> is
> > accurate. Then what diff does it make whether it is listed as a
webserver
> or
> > firewall if what you are after is uptime ?
>
> You're making the same mistake you've been making all along. You're
> assuming that Netcraft will identify a web server with a firewall as the
> firewall, but that's not what happens in many cases (including my own).
>
> Netcraft reports the server and OS as Linux, but it's getting it's uptime
> data from my firewall, which is neither Linux or Unix based (actually it's
> getting no uptime at all because my firewall doesn't give out that data).
>
>
>
>
------------------------------
From: "sandrews" <[EMAIL PROTECTED]>
Subject: Re: Question with Security on Linux/Unix versus Windows NT/2000
Date: Tue, 26 Dec 2000 10:30:25 +0500
In article <[EMAIL PROTECTED]>, "mlw"
<[EMAIL PROTECTED]> wrote:
> Erik Funkenbusch wrote:
>>
>> "mlw" <[EMAIL PROTECTED]> wrote in message
>> news:[EMAIL PROTECTED]...
>> > Erik Funkenbusch wrote:
>> > >
>> > > "mlw" <[EMAIL PROTECTED]> wrote in message
>> > > news:[EMAIL PROTECTED]...
>> > > > Erik Funkenbusch wrote:
>> > > > >
>> > > > > "Philip Neves" <[EMAIL PROTECTED]> wrote in message
>> news:Xu_06.862967
>> > > > > > As of right now There are no viruses that I know of for
>> > > > > > linux.
>> > > > >
>> > > > > Guess you've never heard of Bliss:
>> > > > >
>> > > > > http://math-www.uni-paderborn.de/~axel/bliss/
>> > > > >
>> > > > > It's been around for a while. Not to mention that there have
>> > > > > been virus-like things, such as the morris internet worm.
>> > > > >
>> > > > > > I've been using linux for five years and I have never
>> > > > > > heard of
>> one.
>> > > > >
>> > > > > I guess ignorance, is Bliss.
>> > > >
>> > > > The thing you are missing is that most NT users must operate
>> > > > with OS privileges which would allow a virus to spread
>> > >
>> > > That's not true. NT is perfectly capapble of being used in
>> > > locked down
>> way.
>> >
>> > This is clearly not 100% true.
>>
>> It is 100% true.
>>
>> > As I have said before:
>> >
>> > Some Windows programs can not install for administrator and also
>> > put their settings into a user's account. (You must know this if
>> > you use NT.) They don't know how. So, these programs must be
>> > installed as the user who will be using them. To do this, you must
>> > have the privileges which would allow a virus to spread.
>>
>> There are any number of solutions to this problem. 1) give the user
>> account temporary admin privs, install the program, remove those
>> privs. Problem solved. 2) Using Win2k, simply install the program
>> with admin privs as the dialog box that pops up allows. 3) Monitor
>> the keys created via any number of tools, then recreate those same
>> keys in each user hive.
>>
>> There are other solutions, but those are just off the top of my head.
>>
>> > Some Windows programs assume system wide access to operate. To
>> > operate these programs you must also have privileges which would
>> > allow a virus to spread.
>>
>> Name one. The only one I know of that requires this out of the box
>> is Office 95, and that can be worked around pretty easily.
>>
>> > The answer is to not use Windows NT for these programs OR use NT in
>> > an insecure way. Most users will choose to use NT in an insecure
>> > way.
>>
>> No, the answer is to use it the way it's intended to be used.
>
> That may be all well an good, but you can't argue, on one side, that
> Linux is hard to use because one sometimes needs to manually
> manipulate settings, and then argue that NT isn't for the same reason.
>
> Yes, there are solutions to these of course, but does the average user
> know them? Do you argue that they have to read a FAQ or HOWTO? If so,
> you are being a hypocrite.
>
> The user will user a system with the least amount of work. Just
> because there are work arounds for this limitation, and it is a
> LIMITATION, does not mean the average user will use the more complex
> solution. Most of the NT users I have seen just give themselves admin
> privileges.
>
yes you are correct, as this is what I have done when using nt4. I
just gave myself admin privileges as it was too much bs to try to use
nt on a daily basis without admin privileges.
------------------------------
From: [EMAIL PROTECTED] (The Ghost In The Machine)
Crossposted-To: comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: Conclusion
Date: Tue, 26 Dec 2000 16:04:01 GMT
In comp.os.linux.advocacy, Tom Wilson
<[EMAIL PROTECTED]>
wrote
on Sun, 24 Dec 2000 22:51:58 GMT
<i6v16.311$[EMAIL PROTECTED]>:
>
>"The Ghost In The Machine" <[EMAIL PROTECTED]> wrote in
>message news:[EMAIL PROTECTED]...
>> In comp.os.linux.advocacy, Tom Wilson
>> <[EMAIL PROTECTED]>
>> wrote
>> on Sun, 24 Dec 2000 13:58:32 GMT
>> <cin16.175$[EMAIL PROTECTED]>:
>> >
>> >"Chad C. Mulligan" <[EMAIL PROTECTED]> wrote in message
>> >news:SG816.23891$[EMAIL PROTECTED]...
>> >>
>> >> "sfcybear" <[EMAIL PROTECTED]> wrote in message
>> >> news:922u16$hpp$[EMAIL PROTECTED]...
>
><snip>
>
>> >>
>> >> Repeating the big lie, Joe Goebbells would be proud of you.
>> >
>> >Looking for the 'H' word, I take it? ;)
>>
>> Hell?
>>
>> Harpies?
>>
>> Hatred?
>>
>> Hat rack?
>>
>> Hoopla?
>>
>> Hacked-up numbers?
>>
>> Hacked systems?
>>
>> Hurts?
>>
>> Hertz?
>>
>> Hambone?
>>
>> (um...can you give us a hint, here? :-) )
>>
>> [.sigsnip]
>>
>> --
>> [EMAIL PROTECTED] -- insert random word beginning with 'H' here
>> up 88 days, 9:23, running Linux.
>
>
>Last hint: Heil!
Oh yeah, that German (actually, Austrian) guy.
That was my next guess. Really. :-)
>
>PS: I don't believe invocation works without someone calling sombody else a
>Nazi first. You may call me one if you wish as i'm fairly thick-skinned and
>not likely to be offended.
Personally, I'm not one to offend anyone if I can help it; it
hadn't occurred to me that Godwinizing was occurring in this thread. :-)
--
[EMAIL PROTECTED] -- insert random oopsie here
up 89 days, 16:26, running Linux.
------------------------------
From: "Adam Ruth" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: Conclusion
Date: Tue, 26 Dec 2000 08:56:45 -0700
I'm not suggesting that it's impossible, I'm suggesting that it is unlikely.
Your example is a great example of why it's unlikely, Netcraft was able to
determine in your case that the uptime numbers are inaccurate with your type
of "firewall", and therefore correctly ignored them. Here are the likely
scenarios that could have happened in your case:
1) They couldn't get the uptime nor the OS from your router, so they went
with the OS of the web server because it's available in the header strings.
2) They couldn't get the OS but they could get the uptime of your router.
They chose to ignore the uptime and go with the OS of the webserver (an
uptime without an OS or the wrong OS attached is pointless you know).
3) They couldn't get the uptime but could get the OS of the router. They
chose to ignore the OS and go with the web server because this is a "web
server survey".
Of course there is the possibility of what you suggested, which is that if
they could have gotten the uptime from the router they would have just
reported that. Well, it's all conjecture you know, without any evidence.
And since very few sites put their webservers behind NAT routers (they
sometimes use proxys, but mostly it's filters), your situation is rare
anyway.
But of course the burden of proof is on you and your claim, because of the
impossibility of proving a negative proposition (which is that there are no
sites with the problem you suggest). Since you claim to have much
experience with complex web server setups, perhaps you could tell us about
some so we could see if Netcraft's numbers are right.
Adam Ruth
"Erik Funkenbusch" <[EMAIL PROTECTED]> wrote in message
news:4_N16.151$[EMAIL PROTECTED]...
> That's exactly the point. Without inside knowledge of the site, you don't
> know if the statistics are correct or not. My point is that if a firewall
> can interfere with providing uptimes, then it could also give inaccurate
> ones. Or are you suggesting that this is impossible?
>
> I can't give an example, because I don't have inside information on all
the
> sites that netcraft is reporting, only my own.
>
> "Adam Ruth" <[EMAIL PROTECTED]> wrote in message
> news:927t9a$8eo$[EMAIL PROTECTED]...
> > You're making the same mistake you've been making all along. It's
getting
> > NO uptime from you so you can't use yourself as an example. If what
> happens
> > in your situation happens "in many cases", then we'd see a lot of sites
> with
> > no uptime reported. What site does
> > it pull the OS from the webserver and the uptime from the firewall? You
> > claim this happens but you haven't given an example. In your case
> Netcraft
> > knows that the firewall isn't providing the data correctly or isn't
> > providing the data at all (we're not privy to what they're getting, only
> > their interpretation).
> >
> > Show a concrete example where this happens, not the theoretical
> possibility.
> >
> > Adam Ruth
> >
> >
> > "Erik Funkenbusch" <[EMAIL PROTECTED]> wrote in message
> > news:DFE16.142$[EMAIL PROTECTED]...
> > > "Jim Richardson" <[EMAIL PROTECTED]> wrote in message
> > > news:[EMAIL PROTECTED]...
> > > > If the OS is detected correctly, and the uptime returned for that
> system
> > > is
> > > > accurate. Then what diff does it make whether it is listed as a
> > webserver
> > > or
> > > > firewall if what you are after is uptime ?
> > >
> > > You're making the same mistake you've been making all along. You're
> > > assuming that Netcraft will identify a web server with a firewall as
the
> > > firewall, but that's not what happens in many cases (including my
own).
> > >
> > > Netcraft reports the server and OS as Linux, but it's getting it's
> uptime
> > > data from my firewall, which is neither Linux or Unix based (actually
> it's
> > > getting no uptime at all because my firewall doesn't give out that
> data).
> > >
> > >
> > >
> > >
> >
> >
> >
> >
> >
> >
>
>
------------------------------
From: R.E.Ballard ( Rex Ballard ) <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: Uptimes
Date: Tue, 26 Dec 2000 15:58:44 GMT
In article <9dx%5.11043$[EMAIL PROTECTED]>,
"Erik Funkenbusch" <[EMAIL PROTECTED]> wrote:
> "R.E.Ballard ( Rex Ballard )" <[EMAIL PROTECTED]> wrote in message
> news:91lsdc$9fm$[EMAIL PROTECTED]...
Bottom line: The EULA is ambigouous and unclear, subject to
interpretation by lawyers and judges.
Reccomendation: Get a very clear contract with Microsoft that
states exactly what will be measured in terms of CALs. Better
yet, get license agreements that eliminate CALs althogether.
SQL Server Enterprise, Windows 2000 Data Center Edition, and
certain other "middle-ware-free" implementations may be exempt
from CALS (for a price tag equivalent to about 200 CALS).
Unfortunately, the CALs are nickles and dimes when you consider the
costs of stress testing, regression testing, and capacity management
for even moderately busy servers. With server competent consultants
running $200/hour in NYC, and the averege Win2K engagement lasting
3-6 months for 3-5 people, the costs can range from $1/4 million
to $1 million.
Of course, if Microsoft decides to play "Back Office" rules (peak 1/2
hour usage), the CALs can add up to several million as well.
Get it in writing in language that everyone can understand clearly.
> > > Nope, the client is the IIS Server A single client, Or you license
> > SQL
> > > Server by number of processors it is running on. Note these
> > processors can
> > > either be in the same machine or on separate ones.
> >
> > Quoted from Microsoft EULA for IIS:
> > <quote>
> > FOR MICROSOFT WINDOWS NT SERVER -- CLIENT ACCESS
> > In addition to the Client Access requirements currently set forth in the
> > applicable EULA,
> > <b> you need a separate Client Access License for Windows NT Server in
> > order to access or otherwise utilize the following Windows NT Server
> > basic network/application services or Server Software components:</b>
> > Microsoft Message Queue Server (sending or receiving messages from
> > Microsoft Message Queue Server), Microsoft Transaction Server (invoking
> > component-based applications managed by Microsoft Transaction Server),
> > and Remote Access Service (accessing the server from a remote location
> > through a communications link). Note: Remote Access Service includes the
> > use of Internet Connection Services, including Internet Authentication
> > Services (validation or transference of a remote access request) or
> > Connection Point Services (remotely configuring the Microsoft Connection
> > Manager Client with new phone numbers or other data). Performance or
> > Benchmark Testing. You may not disclose the results of any benchmark
> > test of either the Server Software or Client Software for Microsoft
> > Message Queue Server, Microsoft Transaction Server or Microsoft Internet
> > Information Server to any third party without Microsoft's prior written
> > approval. Installation on a Single Server. The Server Software
> > components that make up the applicable SOFTWARE PRODUCT may only be
> > installed together for use on one Server and may not be separated,
> > unless otherwise provided herein. Note on Microsoft Site Server Express.
> > You may freely copy and distribute Microsoft Site Server Express for
> > your use on any computer within your organization.
> > </quote>
>
> The phrasing in this section of the EULA is "Internet Authenticaion
> Services". If you authenticate a remote login over the internet, then
it's
> a CAL, if it's a non-authenticated connection than it's covered under the
> unlimited connection internet license.
>
>
--
Rex Ballard - VP I/T Architecture
Linux Advocate, Internet Pioneer
http://www.open4success.com
Linux - 60 million satisfied users worldwide
and growing at over 9%/month! (recalibrated 10/23/00)
Sent via Deja.com
http://www.deja.com/
------------------------------
From: "Ayende Rahien" <[EMAIL PROTECTED]>
Crossposted-To: alt.linux.sux
Subject: Re: linux does NOT suck (oh yes it does)
Date: Tue, 26 Dec 2000 18:34:38 +0200
"Kyle Jacobs" <[EMAIL PROTECTED]> wrote in message
news:ddW16.67214$[EMAIL PROTECTED]...
> Just because you've had a picture perfect Linux experience doesn't mean
all
> the people I test for also have perfect experiences.
>
> I don't use Linux in a professional environment. I hope to, one day. But
> now, Linux is just a crappy plaything. My personal problem is it's not
> getting better. On the desktop, it sucks. As a server? Only a few
> distro's are even worth my 10 minutes. And god help me if I call tech
> support.
As a server, I find all the distributions of Linux useless, as I can get a
BSD for the same price, get a higher quality product and lose nothing in the
process.
------------------------------
From: [EMAIL PROTECTED] (Bob Hauck)
Subject: Re: New to Linux, and I am not satisfied.
Reply-To: bobh{at}haucks{dot}org
Date: Tue, 26 Dec 2000 16:43:44 GMT
On Sun, 24 Dec 2000 15:45:28 GMT, T. Max Devlin <[EMAIL PROTECTED]> wrote:
>Said Bob Hauck in comp.os.linux.advocacy on Sat, 23 Dec 2000 23:43:42
>>On Sat, 23 Dec 2000 19:40:31 GMT, T. Max Devlin <[EMAIL PROTECTED]> wrote:
>>>Said Bob Hauck in comp.os.linux.advocacy on Sat, 23 Dec 2000 15:43:43
>>>GMT;
> The primary reason for avoiding an ISP-based news service is
> they require that you be connected through their dial-up box in order to
> access news. Since I always am going to be using a variety of business
> and personal systems and dial-up accounts, I'm not satisfied with this
> arrangement.
Ok, that's legitimate. Your ISP could provide authenticated NNTP so you
could log in from anywhere, but if they don't then they have to restrict
by IP otherwise they spammers will be pounding them into the ground in a
week.
In the "old days", there used to be a lot of open NNTP servers. The
spammers and software pirates and idiots have pretty much killed off
that breed. That's too bad, but I think it is the price you pay for the
growth of the Internet.
>>> So how come you had to pay by the number of simultaneous users?
>>
>>How else would you price it? You could do flat-rate, but that would
>>mean the little customers subsidize the big ones.
>
>Actually, the big customers subsidize the little ones; that's what
>economies of scale are supposed to provide. Get it?
No, I don't, because we are comparing two different pricing models with
different effects regarding cross-subsidy.
*If* they charge a flat rate, ISP's with 10 simultaneous users pay the
same dollar amount as ones with 500. That sounds to me like the small
guy is subsidizing the big guy, since at least some of the NNTP service's
costs are related to per-user resources.
OTOH, *if* they charge per-user, then it might be as you suggest where
the big guy subsidizes the little guy, depending on what kind of volume
breaks are given.
>>The economies of scale I spoke about are related to the fact that a
>>small ISP doing their own news server will be carrying thousands of
>>groups that literally none of their customers read.
>
>Actually, they won't be carrying any groups. How is that economies of
>scale?
I write "the small ISP doing their own news server". They do exist you
know. That is, ISP's that handle their own news feeds (and you normally
want more than one to get reliability and full coverage). They get
feeds from a couple of other ISP's, and maybe feed some too. That's how
it used to be done, and still is by a shrinking minority. You might ask
yourself why they are a minority and why there are fewer and fewer of
them doing it themselves and more are contracting with Supernews and the
like.
>I know, I know. You're trying to say its cheaper for a large company to
>run servers for bunches of ISPs. Which leaves my question, "How is it
>'economies of scale' that they charge the ISPs per simultaneous user?"
>You asked "how else", and that's really beside the point. However they
>do it, apparently they aren't benefitting from 'economies of scale'.
There are economies of scale if the big service can provide NNTP for
less than what an ISP would pay for equipment, bandwidth, labor, and the
cost of the feeds to do it on their own. There can still be per-user
costs in this scenario, but they would need to be lower than what they
would be on a smaller scale.
>> So they are paying for disk space and bandwidth to download groups
>> that nobody reads. And that's a lot of disk space and a lot of
>> bandwidth these days.
>Disk space and bandwidth are both entirely replenishable.
Disk space is cheap, but not free, and the high-performance variants
needed for a good news server are especially non-free. Bandwidth is not
particularly cheap to an ISP either because they can't (well...shouldn't)
buy a consumer pipe that's already been oversold 30 to 1. And labor is
very expensive.
> Your argument seems to ignore the fact that the entire basis of Usenet
> is servers which deal with these issues precisely. Perhaps it is too
> much work to handle it well, but its really a matter of how short your
> aging is.
You seem to forget that I actually ran a Usenet server with a full fead
for about four years. I know what I'm talking about.
How short your aging is affects how much disk space you need, which does
indirectly affect how hard you work (if you try to run close to capacity
in order to make maximum use of your disks, sometimes you'll get hit by
a blizzard of spam or Grateful Dead tunes and run out of space). How
much work you do is also affected by how reliable your equipment is, how
reliable your upstream feeds are, and how often you need to upgrade your
hardware in order to keep up with the ever-growing volume of binary crap
and spam.
News admins spend a lot of time tuning things and recovering from
various failures (e.g. one of your feeds went down for a day and is now
spewing gigabytes of old news at you as fast as it can). They are also
continuously spending money on more of those cheap LVD SCSI disks.
Being a full-time news admin isn't so bad, actually, but trying to do it
part time is a pain in the ass and most small ISP's can't afford a
full-time admin for news. They would rather spend the money on things
like tech support or web admins, things that more of their customers
care about.
> I don't like what Deja or Supernews have become, and I don't like my
> ISP "giving me" NNTP access only when I'm dialed in through their
> links.
Yes, that is one complaint a few of our more sophisticated users had
about Supernews. The alternative is to use authenticated NNTP, which
old news clients don't support. That is less of a problem now, and
maybe they should change their policy.
Most of the users who complained about this were also clued enough to
understand how to read news with slrn from a shell account. Giving
these users shell accounts solved the majority of the complaints.
> Turns out, its not very popular, even though I know it is the most
> reliable, powerful, scalable, and flexible way to run discussion
> forums and article archives and such. That sucks.
Yes, it is the most scalable way to run forums. Yes, it is more
flexible than web-based ones. Yes, it does suck that it seems to be
getting shunted aside. There are lots of reasons for what's happening
though, only a few of which have to do with the avarice of Supernews.
One thing is that the spammers discovered Usenet. Yes, the cancelbots
sort of keep it under control, but at the cost of further increasing the
already insane volume of messages. Even a few years ago I was getting
tens of thousands of cancels a day.
Another thing is piracy and porn. Not that they exist, but the sheer
volume they create. The volumes on binary groups are just out of
control. Worse, since the volume in binary groups is created by
relatively few irresponsible people, it changes wildly from day to day,
which makes it hard to optimize disk space and creates a lot more work
and expense for admins. Many smaller ISP's simply don't carry binaries
any more, but too much of that results in binaries showing up in other
groups, creating a new problem. A lot of this could be eliminated if
everyone used authenticated NNTP, but that goes against the grain of
many Usenetters.
It is the volume of messages that's driving the consolidation toward
providers like Supernews among small to medium ISP's.
Another problem with NNTP is the "web mentality" of new users aided and
abetted by the "lock in mentality" of vendors. New users generally
don't "get" Usenet and since it isn't a proprietary technology most
software vendors aren't interested in explaining it. They are more
interested in selling proprietary versions like Exchange and Domino or
closed web-based solutions. Here's where the avarice of Supernews comes
in.
> >> A bigger operation avoids this waste, but they still have to
> >> provide per-user resources too.
> >What per-user resources? That's the fallacy; there are no per-user
> >resources.
Yes there are. Each NNTP reader creates a connection and a process on
the server. This uses memory and cpu time and creates IO traffic to the
disks and network. All of these represent costs in that the systems
that support readers must be sized according to the expected number of
users (as opposed to feeder machines that must be sized by the volume of
traffic and expiration policy).
--
-| Bob Hauck
-| To Whom You Are Speaking
-| http://www.haucks.org/
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list by posting to comp.os.linux.advocacy.
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Advocacy Digest
******************************
