okay, here's where lots of VoIP things go wrong: they think it's okay to use the same line for normal internet access as well as VoIP (i'm assuming you have a broadband line with an upload speed of max 256k but this also even applies - if you load it enough - if you have e.g. 1MB SDSL).
given that the MTU has to be slammed up so far (in order for ISPs to compete on "bandwidth" rather than "latency") to ridiculous levels (1400-1500) it leaves very little options at _your_ end even if you _do_ do QoS tricks.
so, your only _sensible_ option is: get a second broadband line, and use it _exclusively_ for VoIP.
and if you are going to do _that_ then make sure that you get a fixed IP address and put the damn ADSL card _in_ the asterisk [or SIP] server.
the reason is quite simple: NAT on SIP is a _complete_ bitch to set up, especially due to RTP (the audio) and you can avoid an awful lot of hassle by putting the ADSL card into your server, so it is a direct interface on the server. this assumes, of course, that you're not running windows!
also - make sure you use 8k CODECs like GSM, because you very quickly run out of bandwidth on a 256k upload if you use 32k CODECs.
oh - another one i forgot (but it's probably slash-dot-too-late...)
in debian (testing/unstable), hotplug has recently been superceded by udev and hal.
this is GREAT news because hotplug is a pig that adds oh, what... 20 seconds to your boot time.
i recently set up a desktop system that now takes only 15-20 seconds to boot up (and i moved the/etc/rc2.d priority of xserver and xfs to 16 and 15 respectively to make their startup overlap the services that _so_ won't be needed whilst x is starting and someone's logging in).
all you have to do is add testing and unstable to your/etc/apt/sources.list, and then:
obviously, KDE 3.4 isn't faster than say fluxbox, icewm or fvwm2.
gnome is a sick puppy that should have been shot at birth for its own good.
personally i love fvwm2, with the default debian configuration from when it had a 3x3 desktop grid thing, so i've resisted upgrading ever since in order to avoid losing that default config: i just _love_ the 1 second startup time and the fact that it runs xterm, and i then type the name of programs to run (mozilla-firefox, gaim, mutt) in the xterm and go from there.
for personal use/preference: bugger KDE and Gnome:) p.s. i _love_ superkaramba and kroller, though (see kde-look.org).
1) install KDE 3.4 (it's faster, it's better all-round).
2) run prelink -v --conserve-memory -q -a
but first add/usr/lib/mozillaNNNNN and/usr/lib/kde3 to
prelink.conf (and any other software such as openoffice)
3) on debian, edit/etc/default/rcS and replace
FSCKFIX="n" with FSCKFIX="y"
4) on debian, install hal, dbus-1 and udev, and then edit/etc/default/hal and make sure DROP_DAEMON_PRIVS is
commented out (this will make it possible for you to
mount auto-detected USB drives etc.)
5) cd to/etc/hal/device.d and do this:
ln -s/usr/bin/fstab-sync 50-fstab-sync.hal
5) edit/etc/profile and add this:
export KDE_IS_PRELINKED="1"
these simple things will make your system faster, more robust in the face of complete technically incompetent blithering idiots who would otherwise blindly press ctrl-d when faced with a prompt saying "your filesystem is corrupted. give root password for maintenance or press ctrl-d", and also provide automatic access to USB devices that is otherwise bloody inconvenient.
"Safe Harbour" conflict with Anti-Spam/Virus!
on
DMCA Abuse Widespread
·
· Score: 3, Interesting
A service provider must satisfy the following critical elements in order to qualify for the "safe harbor" or protection from liability provided by subsection 512(a) (note that subsection 512(k)(1)(A) defines "service provider" as used in subsection 512(a)):
(e) The service provider must not modify the communication selected by the Internet user [512(a)(5)];
so, if you "modify" the email to put "X-Spam" tags in it, you no longer qualify for the "safe harbor" provisions.
in fact, if you put ANY headers with the message, then the communication is "modified".
oh! oh! that explains george bush and tony blair's behaviour _thank_ you for clearing that up. i was worried i was being paranoid but now i know it's just demonic alien nation possession, *whew*, i can _handle_ that _no_ problem.
the model of selling boxed software at a fixed price is fundamentally flawed as a guaranteed and continuous revenue stream.
you only have to look at the dilberts to know that this is true: namely that because software is never perfect and that its deployment requires updates, for security or simply to keep up with user expectations, you can never ever have a company that survives purely on the sale of software for a fixed, one-off price.
pretty soon, the company will have its money come in; pretty soon, that money will be spent. what then?
well, here's the possible solutions:
* support - you pay money for a support contract. but if the software works, why should you bother?
* expansion of capabilities. but if there are competitor products out there, again: why bother? just buy the competitor product.
* deliberate introduction of flaws or deliberately shipping before product is really ready for production use: this is the most commonly taken tack, because it guarantees that people will "upgrade" - pay more money for _yet another_ boxed product, or for the "upgrade" license.
* software as a service. this is the more "honest" approach. you pay a little per month, and that goes directly towards the developers pockets and the sales/marketing and the directors, and everyone's happy.
trouble is, the "software as a service" model is drastically undercut by the "software as a boxed product", except for the largest software development projects.
and of all but the really _really_ large software development projects for desktop usage (such as those which as MSRPC-based applications which are based on approx 100-man-years of coding such as Exchange) you can get free software alternatives - OpenOffice, Apache, Samba.
i gave one uk department a very polite form of hell, and, amazingly, it worked.
i basically explained to them that what they were doing was forcing people to fork out £500 for windows software plus another £500 for a computer capable of _running_ the windows software.
i then liked this into "discrimination", for which they could quite seriously get into trouble.
to their credit, they actually responded, sorted out the web site (and stopped publishing things in.doc format).
... developed to speed up the deployment and computation of algorithms when you needed results quicker than a single Computer could handle, were lost due to the introduction of the microprocessor.
also:
there are sci-fi books and short stories about this sort of thing.
one of them is "the end of eternity", by isaac asimov - Computer Harkan (Computer as in a title like Doctor) is the main character.
also there's a short story - by greg bear, i believe - about a space expedition that got lost in deep space: the entire crew learned how to do sines, cosines etc. it took months, and by the time they were far enough along to plot a reasonable course home, the crew were doing the calculations in their heads.
fascinating but only enough material for a _short_ story...
would be a story about how KDE can be cross-compiled on linux to run on wine or natively on windows.
at present, the only way to compile KDE for windows is by compiling it under windows (which requires that you have windows).
and that means you have to utilise a specialised forked version of kde source code, which has a bastardised version of configure, called configure.bat.
aluminium comprises EIGHT PERCENT of the earth's crust - it is one of the most abundant metals on the planet, and one that can be used to store the most energy.... i _can_ think of a reason why the DoE would not want aluminium cells to be developed: a small cell weighing a few ounces would be capable of delivering well over 1,000 amps. with the right kind of shorting out, i feel reasonably confident in speculating that the cell itself would make a very effective bomb.
i've spoken to them, and i've been tracking the company for some time (nearly two years).
they have a prototype cell already working.
they are presently being funded by toyota and a few other companies to produce prototype "large" cells sufficient to power electric vehicles.
they are paranoid as hell because the potential for this product is for it to become a BILLION dollar industry.
they ALSO don't want to get ripped off (the patent is in swedish).
the production process requires nanotechnology
[usually, because nanotech is in such an early stage of development, that means they need a supply of materials in the form of somewhere from 10 to 40nm equally sized diameter balls of compound "X", "Y" or "Z"]
hope this helps explain why their site looks like it's nothing more than "give me money".
shit - if i had had £10,000, **I** would have given them money, before the jan 1 2005 deadline for personal investors ran out.
then i would have been able to receive a license for the manufacture of their technology once they had the mass-production factory blueprints worked out.
Slashdot Mirror
okay, here's where lots of VoIP things go wrong: they think it's okay
to use the same line for normal internet access as well as VoIP (i'm
assuming you have a broadband line with an upload speed of max 256k
but this also even applies - if you load it enough - if you have e.g.
1MB SDSL).
given that the MTU has to be slammed up so far (in order for ISPs to
compete on "bandwidth" rather than "latency") to ridiculous levels
(1400-1500) it leaves very little options at _your_ end even if
you _do_ do QoS tricks.
so, your only _sensible_ option is: get a second broadband line,
and use it _exclusively_ for VoIP.
and if you are going to do _that_ then make sure that you get a fixed
IP address and put the damn ADSL card _in_ the asterisk [or SIP] server.
the reason is quite simple: NAT on SIP is a _complete_ bitch to set up,
especially due to RTP (the audio) and you can avoid an awful lot of hassle by putting the ADSL card
into your server, so it is a direct interface on the server. this assumes,
of course, that you're not running windows!
also - make sure you use 8k CODECs like GSM, because you very quickly run out of bandwidth
on a 256k upload if you use 32k CODECs.
oh - another one i forgot (but it's probably slash-dot-too-late...)
/etc/rc2.d priority of xserver and xfs to 16 and 15 respectively to make their startup overlap the services that _so_ won't be needed whilst x is starting and someone's logging in).
/etc/apt/sources.list, and then:
in debian (testing/unstable), hotplug has recently been superceded by udev and hal.
this is GREAT news because hotplug is a pig that adds oh, what... 20 seconds to your boot time.
i recently set up a desktop system that now takes only 15-20 seconds to boot up (and i moved the
all you have to do is add testing and unstable to your
apt-get install udev hal
hurrah.
KDE 3.4 is faster... than KDE 3.2 or KDE 3.3.
:) p.s. i _love_ superkaramba and kroller, though (see kde-look.org).
obviously, KDE 3.4 isn't faster than say fluxbox, icewm or fvwm2.
gnome is a sick puppy that should have been shot at birth
for its own good.
personally i love fvwm2, with the default debian configuration from when it had a 3x3
desktop grid thing, so i've resisted upgrading ever since in order to avoid losing that default config: i just _love_ the 1 second startup time and the fact that it runs xterm, and i then
type the name of programs to run (mozilla-firefox, gaim, mutt) in the xterm and go from
there.
for personal use/preference: bugger KDE and Gnome
1) install KDE 3.4 (it's faster, it's better all-round).
/usr/lib/mozillaNNNNN and /usr/lib/kde3 to
/etc/default/rcS and replace
/etc/default/hal and make sure DROP_DAEMON_PRIVS is
/etc/hal/device.d and do this: /usr/bin/fstab-sync 50-fstab-sync.hal
/etc/profile and add this:
2) run prelink -v --conserve-memory -q -a
but first add
prelink.conf (and any other software such as openoffice)
3) on debian, edit
FSCKFIX="n" with FSCKFIX="y"
4) on debian, install hal, dbus-1 and udev, and then edit
commented out (this will make it possible for you to
mount auto-detected USB drives etc.)
5) cd to
ln -s
5) edit
export KDE_IS_PRELINKED="1"
these simple things will make your system faster, more robust in the face of complete technically incompetent blithering idiots who would otherwise blindly press ctrl-d when faced with a prompt saying "your filesystem is corrupted. give root password for maintenance or press ctrl-d", and also provide automatic access to USB devices that is otherwise bloody inconvenient.
A service provider must satisfy the following critical elements in order to qualify for the "safe harbor" or protection from liability provided by subsection 512(a) (note that subsection 512(k)(1)(A) defines "service provider" as used in subsection 512(a)):
(e) The service provider must not modify the communication selected by the Internet user [512(a)(5)];
so, if you "modify" the email to put "X-Spam" tags in it, you no longer qualify for the "safe harbor" provisions.
in fact, if you put ANY headers with the message, then the communication is "modified".
oh! oh! that explains george bush and tony blair's behaviour _thank_ you for clearing that up. i was worried i was being paranoid but now i know it's just demonic alien nation possession, *whew*, i can _handle_ that _no_ problem.
this is a _copy_ of a _copy_. the geocities site on which the copy
l ouisette.lanteigne.html
is hosted is overloaded. so i made a copy of the google cache, here:
http://hands.com/~lkcl/activa.holdings.report.by.
http://www.rabble.ca/babble/ultimatebb.php?ubb=get _topic&f=5&t=001759
someone may have tracked down activa holdings.
activa's web site is "under construction"...
could someone tell me if the activa corporation that is suing this
e mentdisplay_id=5216
person is the same company involved in _this_:
http://www.labour.gov.za/media/statement.jsp?stat
here's the difference: can you pay microsoft - or any of its employees - to work on windows or office to get them to fix bugs or add new features?
.... you mean... like... nobody noticed the poor spelling of the word
"occasionally" in the orignal article for lik.... svern yeers?
pttoeey, i spit upon you geeks with your poor spellin attenshun spans.
the model of selling boxed software at a fixed price is fundamentally flawed
:)
as a guaranteed and continuous revenue stream.
you only have to look at the dilberts to know that this is true:
namely that because software is never perfect and that its deployment
requires updates, for security or simply to keep up with user
expectations, you can never ever have a company that survives purely
on the sale of software for a fixed, one-off price.
pretty soon, the company will have its money come in; pretty soon, that
money will be spent. what then?
well, here's the possible solutions:
* support - you pay money for a support contract. but if the software
works, why should you bother?
* expansion of capabilities. but if there are competitor products
out there, again: why bother? just buy the competitor product.
* deliberate introduction of flaws or deliberately shipping before
product is really ready for production use: this is the most commonly
taken tack, because it guarantees that people will "upgrade" - pay
more money for _yet another_ boxed product, or for the "upgrade"
license.
* software as a service. this is the more "honest" approach. you
pay a little per month, and that goes directly towards the developers
pockets and the sales/marketing and the directors, and everyone's
happy.
trouble is, the "software as a service" model is drastically undercut
by the "software as a boxed product", except for the largest software
development projects.
and of all but the really _really_ large software development projects
for desktop usage (such as those which as MSRPC-based applications which
are based on approx 100-man-years of coding such as Exchange) you can
get free software alternatives - OpenOffice, Apache, Samba.
so things are a little odd, right now...
http://erp5.org/ is missing from the list of ERP solutions in the article.
need someone to carry your code?
HA. onedollaaa! i carry loooong time
i gave one uk department a very polite form of hell, and, amazingly, it worked.
.doc format).
i basically explained to them that what they were doing was forcing people to fork out £500 for windows software plus another £500 for a computer capable of _running_ the windows software.
i then liked this into "discrimination", for which they could quite seriously get into trouble.
to their credit, they actually responded, sorted out the web site (and stopped publishing things in
... it only went away...
hey, there's a decnet sourceforge project.
...
AT LAST! we can marry freedce, a little used, totally misunderstood and strategically critical project back with DECnet, a little used
DCE 1.1 (from which freedce is derived) shows signs of support for DECnet having been pulled (from the "commercial" version(s)) in rather a hurry.
the neat thing about dce/rpc is that it supports literally _any_ transport and i do mean any.
... developed to speed up the deployment and computation of algorithms when you needed results quicker than a single Computer could handle, were lost due to the introduction of the microprocessor.
also:
there are sci-fi books and short stories about this sort of thing.
one of them is "the end of eternity", by isaac asimov - Computer Harkan (Computer as in a title like Doctor) is the main character.
also there's a short story - by greg bear, i believe - about a space expedition that got lost in deep space: the entire crew learned how to do sines, cosines etc. it took months, and by the time they were far enough along to plot a reasonable course home, the crew were doing the calculations in their heads.
fascinating but only enough material for a _short_ story...
would be a story about how KDE can be cross-compiled on linux to run on wine or natively on windows.
at present, the only way to compile KDE for windows is by compiling it under windows (which requires that you have windows).
and that means you have to utilise a specialised forked version of kde source code, which has a bastardised version of configure, called configure.bat.
yerk.
the most significant thing about this is that linux, running in a XEN Domain 0, will be able to run (and manage) MAC-x86 in a "guest" domain.
and also, vmware will be able to host MAC-x86 under both windows and linux.
that's the best bit about MAC-x86 being available.
muyhahaha *cackle* *dribble, leer*
is anyone here concerned about the parallels with the latest episode of Dr Who?
the game-shows where losers are beamed up to the Dalek Mothership?
EXTERMINATE!
EXTERMINATE!
12 years ago, why did the DoE's requirements for funding specify that the cells must be above 1.5 volts?
... i _can_ think of a reason why the DoE would not want aluminium cells to be developed: a small cell weighing a few ounces would be capable of delivering well over 1,000 amps. with the right kind of shorting out, i feel reasonably confident in speculating that the cell itself would make a very effective bomb.
that made DAMN sure that aluminium cells (see http://europositron.com/ would not be researched.
aluminium comprises EIGHT PERCENT of the earth's crust - it is one of the most abundant metals on the planet, and one that can be used to store the most energy.
i've spoken to them, and i've been tracking the company for some time (nearly two years).
they have a prototype cell already working.
they are presently being funded by toyota and a few other companies to produce prototype "large" cells sufficient to power electric vehicles.
they are paranoid as hell because the potential for this product is for it to become a BILLION dollar industry.
they ALSO don't want to get ripped off (the patent is in swedish).
the production process requires nanotechnology
[usually, because nanotech is in such an early stage of development, that means they need a supply of materials in the form of somewhere from 10 to 40nm equally sized diameter balls of compound "X", "Y" or "Z"]
hope this helps explain why their site looks like it's nothing more than "give me money".
shit - if i had had £10,000, **I** would have given them money, before the jan 1 2005 deadline for personal investors ran out.
then i would have been able to receive a license for the manufacture of their technology once they had the mass-production factory blueprints worked out.
:)
pops another one...