One thing that confuses me about Jabber is that people seem to forget that good old SMTP solves many of the same problems, and in fact solves them better.
For example, many years of work have gone into making sure that email never gets lost. SMTP mailers just don't lose email anymore. Jabber messages, on the other hand, are not really reliable. If the user to whom you are targeting a message is not online, the server may queue the messages, but the policy is not clear as to how long they will be stored, or if the server is rquired to store them at all.
This makes me worry about the idea of using Jabber to build infrastructure where you rely on messages to always be delivered.
It seems to me that many of the issues that Jabber solves have been solved using existing technology such as SMTP, and mailer and mailing list services built on top of it, like qmail, mailman, etc.
I paid $2000 / month for 1/4 T1 in 1995. Since then, PC's have gone from 90 MhZ to 2 GHz. RAM has dropped in price by a factor of 20 or more. Disk drives by a factor of 100. Bandwidth inside of CMOS chips is up by a factor of 100.
So... the technology used for switching digital signals is now cheaper than any analog phone technology. Why should a T1 line be any more expensive than a regular voice line?
The thing stinks of monopoly practices.
I used a "washlet" for a couple of years
on
Best High-Tech Toilet?
·
· Score: 2, Interesting
IN Japan, they refer to the washing toilets as "washlets". I had one in our apartment when living in Japan for two years.
Simply put, after using the washlet for two years, coming back to the US and using only
toilet paper now is like wiping my ass with dried leaves.
As someone else pointed out, the biggest obstacle to adoption of these things in the US is probably the lack of AC power next to the toilet.
Otherwise, I would love to have washlets in my house in the US. The heated seat is great on those cold mornings, and the warm water washing
is much cleaner and healthier and more comfortable than dry toilet paper (yuck!).
Just like with mobile phones and healthy food, the Japanese are ahead of us in this area.
I just got back from Japan last Fall, and had spent the last two years developing iMode applications.
What's cool about iMode for me was:
1) Google's iMode gateway - search Google, and it converts anything (including frames!) into cHTML pages in 5 kbyte chunks, with navigation links. I used to read Roger Ebert's movie reviews, on my phone on the train.
2) cHTML is not WAP. You could write basic HTML pages and they would display on the phones. No syntax errors, no WML, no gateways barfing on you. It simply worked, just like the real web used to. It showed that WAP was completely unneccessary for accessing the web over small low bandwidth devices. Plain old forms work better than "cards", and encourage a more consistent user interface.
3) Email -- the phone's email app was nicely integrated with the browser - you could read your mail and send it, real SMTP mail, and if you got a hyperlink in your email, you could click to follow it in your browser.
4) iAppli -- you could write little Java apps that run in the phone, and can talk to the network via HTTP. Sound, animation, little scratchpad memory on the phone.
5) Nice TFT color displays, FM synthesized polyphonic high quality sound.
6) Latency to get pages was very low, maybe a few seconds the first time you hit a site, and then less than 2 seconds for each subsequent page.
7) You could use plain GIF images in your pages. They didn't handle JPEG at that time though:-(
Anyway, the thing simply worked. No fooling around with broken phones and broken emulators and incompatible protocols. NTT just leaned on the handset makers and made them comply with the (simple) standards.
It is the direct opposite of the WAP situation in the US. WAP should just hurry up and die, and make room for something that works.
If you look at the actual products sold at WalMart, a large fraction are made in China.
Some fraction of that is essentially slave labor.
Americans may vote with their pocketbook, but
that does not absolve people of the responsibility
for understanding where cheap stuff really comes from.
I didn't say there's no upper limit to the size
of an IP packet, only that you should optimize it
at 'runtime' when you're actually talking to
another host end to end, rather than tossing something
arbitrary into the network and letting all the routers along the way pick up the mess. Start with
something standard, say 512 bytes maximum or something, and if you get greedy, try upping it, if you can't get through, try making it smaller.
Packet switching is great, but it turns out that
the IP mechanism of IP packet fragmentation and
reassembly was a big mistake. In other words, you
don't want to break up packets. The original idea was that the MTU (minimum transmission unit) was
unknown, so that you should be able to have
routers which break a single IP packet into smaller fragments and have someone reassemble them elsewhere in the network. It added loads of hair and performance penalties, and really
broke the whole "end to end" model. Here in the future, I believe that people now think you can use end-to-end MTU discovery, and just keep notching the packet size down until you get reliable transmission. Anyone know if IP packet fragmentation is still part of IPV6?
The problem is that the information producer is
paying for bandwidth, not the consumer.
If people could pay per packet to access web sites, then it would not cost anything for a popular provider to serve millions of gigabytes.
This is orthogonal to DDOS attacks, but actually, if the infrastructure were in place for micro-payments for packet billing, then tracking
DDOS attacks would be easier.
You can easily use FEC for burst errors, you just use a cross-interleaved encoding on top of the basic encoding. That's what CD players do and they can eat a 4000 bit long burst error without a hiccup. And CD-ROM encodes another layer on top of CD Audio, and thus can sustain even larger burst errors.
One of the most well known forward error correction algorithms is the Reed-Solomon coding, developed in the 1960's. It is the encoding used
for CD Audio and CD ROM discs, as well as being widely used in disk drive controllers.
Phil Karn (KA9Q) also has an even better implementation available for download someplace.
I had the idea, along with many other people, of distributed filesystems using FEC encoding. My original idea was to spread files over a large number of anonymous FTP servers, and then "harvest" them later. I also was interested in using FEC for UDP audio streams, back in 1994 or so. Anyway, this is old news, and I would be
alarmed if any patents were granted on the idea of
using error correction coding to transfer files.
computers are basically free now, so the price
is determined by the display. A CRT is ugly
as hell, in my opinion, so you want an LCD. An
obsolete laptop may still have a perfectly acceptable display (You wanted 10"). Look on ebay for
laptops with missing CDROM or cracked case or
other cosmetic problems. Then open it up, mount the display, and maybe fold the keyboard underneath. Wireless 802.11 card will make it a
great web-ified picture frame.
I use an old 486 75 Mhz thinkpad with 16 mB of RAM, hooked to
a new 15" flat panel I bought explicitly as a picture frame. The advantage is that it contacts a web server in my house, which selects pictures and lets my friends upload new photos or send them as attachments by email, and they are displayed in our living room.
The price of my system was $0 for the old laptop, and $500 for the beautiful display. But LCD panels
are coming down in price. I didn't use the display on the laptop because its only 640x480, and
doesn't have enough colors.
I don't want to have an electric fan in my
CPU heatsink. Is there a totally passive
heatsink for the Duron which will cool adequately with
no fan at all?
The Timothy McVeigh / gun show / Turner diaries / militia crowd, to me is just as big a threat as the Islamic guys.
Why don't we hear anything about cracking down on domestic right wing terror groups - I feel more threatened by these racist paramilitary thugs practicing their "militia" excercises with real guns than I do about the
Arab hijackers; after all, McVeigh blew up a big building too, and he came from a subculture of violence and racism and ignorance which is happily and healthily living in our own USA.
Please, while we are picking off Arabs across the world, let's have our new secret police assasinate some of our own racist redneck nutjobs as well.
First, work should start immediately to rebuild the World Trade Center. This time they should make it a little higher.
Second, we should invade Afghanistan, and
kick out the Taliban and install a democratic government. I don't care if they are responsible or not, I just hate those guys.
Third, we should take the Palestinians and move them to Saudi Arabia, and declare Saudi Arabia the new Jerusalem. Mideast problem is now solved: Palestinians are rich and don't have anything to complain about anymore, Saudi corrupt Islamic theocracy is gone, Saudi royal family are out on their butts, Israel is now just a patch of desert that no one wants.
Also check out
http://www.axis.com I've used their embedded linux ETRAX100LX boards, they work great. Single
ethernet, but it comes with entire cross compiler
and flash boot loader that works over ethernet
In the 1970's my father made a visit to Brazil, where
as you know they are serious about coffee. He
invented a coffee strength tool using the same principle, but
somewhat simpler, no computer required -- you just take a plastic ruler
and dip it into the coffee, and then read off the
strength by seeing where the last tick on the ruler is that you can still see. A manual optical
strength meter.
The writing seems, well, somehow it just seems like a hoax. The details are too neat, the
style of writing a little too offhand, the interviews with people don't ring true. I dunno, it just doesn't have the "ring of truth" to it.
I think one thing that people haven't quite realized yet is how very very easy it is for a company like Microsoft to take a "public" standard, such as Kerberos, DNS, NFS, SMTP, or whatever, and make it incompatible. It takes just one little change to the protocol, in the guise of "innovation" to make it incompatible.
This is one of their real sources of leverage. They can claim they are "innovating", while
magically making a proprietary product which is a parasite off of existing standards. In other words, today's Internet protocols are sort of like "GPL" code, implicitly, because you need to
make your code obey their "license", in order to interoperate. When someone takes a protocol private, they can get a free ride from the existing implementations, i.e.., their code seems to interoperate, but the user becomes dependent on some new feature, and is suddenly locked in to the proprietary product.
I think that the public, who has not had intimate experience with building and using network protocols and APIs, has little idea of how easy it is to "pollute" an existing standard. Indeed, many protocols, such as SMTP, are subject to "accidental" pollution, where implementors with good intentions still try to stick in a new "feature" into their code.
In the past Microsoft has, and does, make slightly incompatible versions of such basic protocols as SMTP. I have personally experienced plenty of failures as my own legally formatted SMTP messages are chewed up and spit out garbled by Microsoft mail handling products (just "bugs" in their MIME handling, or "embrace and extend"?).
Anyway, I think this is another silent front in Microsoft's war on humanity, and people should realize that the real damage is happening without loud obnoxious announcements by Microsoft management.
How about this idea- no case at all:
Take a big piece of plywood, and attach the motherboard and disk drives, and take the power supply, remove it's sheet metal case and fan, and just hang the whole thing on the wall.
The only thing making noise will be the disk drive.
The magic of modern semiconductor lithography is
that the same area of silicon that contained
a 4 bit CPU in 1980 can contain a x386 compatible
PC and all motherboard functions plus network controller. At some point, as the chip circuit densities
increase the packaging and
bonding for the semiconductor become the
dominant cost, not the chip area.
So, as devices require more sophisticated networking
and other features, it makes sense to use
a controller with enough oomph for the job. Bumming instructions in assembly language is
getting to be less and less necessary, and is
actually a waste of developer's time, if the
silicon is cheap enough.
Hey, I'm living in the future already; I keep
my home directory under CVS, and have it updated
every evening to a remote server, and then
replicated to all my other devices (laptops,
pc's at work, etc).
I don't need sun NFS, I a sure as hell don't need
Microsoft.NET. I just need TCP/I-fucking-P that
works and free software. If it were up to Microsoft, they would extend TCP/IP to have
some incompatible Microsoft extensions. No wait, they already did!
I am proposing we implement social mechanisms
to regulate the use of this information.
Imagagine this scenario: In your neighborhood everyone has cable modem. All your neighbors hook
up some web cams pointing out their windows, to
form the "neighborhood net".
Are you now going to say this is illegal, and
send the police to take away their cameras?
Wouldn't that be an even worse violation of
civil rights than what I propose?
You have proposed no specific regulations which
will help preserve individual rights. Here are a couple of proposals off the top of my head:
You may not make use of publically gathered surveillance information to single out, harass or stalk
an individual.
The police may not use the public surveillance
networks to track and individual without a court order.
No government worker may use the
public surveillance network for purposes of tracking an individual, without a court order.
These are probably flawed, but at least I am trying. What is your proposal? Private citizens can already have a"neigborhood watch" group. Are you going to outlaw that as well?
Slashdot Mirror
One thing that confuses me about Jabber is that
people seem to forget that good old SMTP solves many of the same problems, and in fact solves them better.
For example, many years of work have gone into making sure that email never gets lost. SMTP mailers just don't lose email anymore. Jabber messages, on the other hand, are not really reliable. If the user to whom you are targeting a message is not online, the server may queue the messages, but the policy is not clear as to how long they will be stored, or if the server is rquired to store them at all.
This makes me worry about the idea of using Jabber to build infrastructure where you
rely on messages to always be delivered.
It seems to me that many of the issues that Jabber
solves have been solved using existing
technology such as SMTP, and mailer and mailing list services built on top of it, like qmail, mailman, etc.
I paid $2000 / month for 1/4 T1 in 1995.
... the technology used for switching
Since then, PC's have gone from 90 MhZ to 2 GHz.
RAM has dropped in price by a factor of 20 or more. Disk drives by a factor of 100. Bandwidth inside of CMOS chips is up by a factor of 100.
So
digital signals is now cheaper than any
analog phone technology. Why should a T1 line be
any more expensive than a regular voice line?
The thing stinks of monopoly practices.
Simply put, after using the washlet for two years, coming back to the US and using only toilet paper now is like wiping my ass with dried leaves.
As someone else pointed out, the biggest obstacle to adoption of these things in the US is probably the lack of AC power next to the toilet.
Otherwise, I would love to have washlets in my house in the US. The heated seat is great on those cold mornings, and the warm water washing is much cleaner and healthier and more comfortable than dry toilet paper (yuck!).
Just like with mobile phones and healthy food, the Japanese are ahead of us in this area.
I just got back from Japan last Fall, and
:-(
had spent the last two years developing
iMode applications.
What's cool about iMode for me was:
1) Google's iMode gateway - search Google, and
it converts anything (including frames!) into
cHTML pages in 5 kbyte chunks, with navigation
links. I used to read Roger Ebert's movie reviews, on my phone on the train.
2) cHTML is not WAP. You could write basic
HTML pages and they would display on the
phones. No syntax errors, no WML, no gateways
barfing on you. It simply worked, just like
the real web used to. It showed that
WAP was completely unneccessary for
accessing the web over small low bandwidth
devices. Plain old forms work better than "cards",
and encourage a more consistent user interface.
3) Email -- the phone's email app was nicely
integrated with the browser - you could read your mail and send it, real SMTP mail, and if you got
a hyperlink in your email, you could click to follow it in your browser.
4) iAppli -- you could write little Java
apps that run in the phone, and can talk to
the network via HTTP. Sound, animation,
little scratchpad memory on the phone.
5) Nice TFT color displays, FM synthesized polyphonic high quality sound.
6) Latency to get pages was very low, maybe
a few seconds the first time you hit a site,
and then less than 2 seconds for each subsequent page.
7) You could use plain GIF images in
your pages. They didn't handle JPEG at that
time though
Anyway, the thing simply worked. No fooling around
with broken phones and broken emulators and
incompatible protocols. NTT just leaned on the
handset makers and made them comply with the
(simple) standards.
It is the direct opposite of the WAP situation in the US. WAP should just hurry up and die, and
make room for something that works.
If you look at the actual products sold at WalMart, a large fraction are made in China.
Some fraction of that is essentially slave labor.
Americans may vote with their pocketbook, but
that does not absolve people of the responsibility
for understanding where cheap stuff really comes from.
I didn't say there's no upper limit to the size
of an IP packet, only that you should optimize it
at 'runtime' when you're actually talking to
another host end to end, rather than tossing something
arbitrary into the network and letting all the routers along the way pick up the mess. Start with
something standard, say 512 bytes maximum or something, and if you get greedy, try upping it, if you can't get through, try making it smaller.
Packet switching is great, but it turns out that
the IP mechanism of IP packet fragmentation and
reassembly was a big mistake. In other words, you
don't want to break up packets. The original idea was that the MTU (minimum transmission unit) was
unknown, so that you should be able to have
routers which break a single IP packet into smaller fragments and have someone reassemble them elsewhere in the network. It added loads of hair and performance penalties, and really
broke the whole "end to end" model. Here in the future, I believe that people now think you can use end-to-end MTU discovery, and just keep notching the packet size down until you get reliable transmission. Anyone know if IP packet fragmentation is still part of IPV6?
The problem is that the information producer is
paying for bandwidth, not the consumer.
If people could pay per packet to access web sites, then it would not cost anything for a popular provider to serve millions of gigabytes.
This is orthogonal to DDOS attacks, but actually, if the infrastructure were in place for micro-payments for packet billing, then tracking
DDOS attacks would be easier.
You can easily use FEC for burst errors, you just use a cross-interleaved encoding on top of the basic encoding. That's what CD players do and they can eat a 4000 bit long burst error without a hiccup. And CD-ROM encodes another layer on top of CD Audio, and thus can sustain even larger burst errors.
I have a GPL implementation at http://rscode.sourceforge.net/
Phil Karn (KA9Q) also has an even better implementation available for download someplace.
I had the idea, along with many other people, of distributed filesystems using FEC encoding. My original idea was to spread files over a large number of anonymous FTP servers, and then "harvest" them later. I also was interested in using FEC for UDP audio streams, back in 1994 or so. Anyway, this is old news, and I would be alarmed if any patents were granted on the idea of using error correction coding to transfer files.
computers are basically free now, so the price
is determined by the display. A CRT is ugly
as hell, in my opinion, so you want an LCD. An
obsolete laptop may still have a perfectly acceptable display (You wanted 10"). Look on ebay for
laptops with missing CDROM or cracked case or
other cosmetic problems. Then open it up, mount the display, and maybe fold the keyboard underneath. Wireless 802.11 card will make it a
great web-ified picture frame.
I use an old 486 75 Mhz thinkpad with 16 mB of RAM, hooked to
a new 15" flat panel I bought explicitly as a picture frame. The advantage is that it contacts a web server in my house, which selects pictures and lets my friends upload new photos or send them as attachments by email, and they are displayed in our living room.
The price of my system was $0 for the old laptop, and $500 for the beautiful display. But LCD panels
are coming down in price. I didn't use the display on the laptop because its only 640x480, and
doesn't have enough colors.
I don't want to have an electric fan in my
CPU heatsink. Is there a totally passive
heatsink for the Duron which will cool adequately with
no fan at all?
The Timothy McVeigh / gun show / Turner diaries / militia crowd, to me is just as big a threat as the Islamic guys.
Why don't we hear anything about cracking down on domestic right wing terror groups - I feel more threatened by these racist paramilitary thugs practicing their "militia" excercises with real guns than I do about the
Arab hijackers; after all, McVeigh blew up a big building too, and he came from a subculture of violence and racism and ignorance which is happily and healthily living in our own USA.
Please, while we are picking off Arabs across the world, let's have our new secret police assasinate some of our own racist redneck nutjobs as well.
First, work should start immediately to rebuild the World Trade Center. This time they should make it a little higher.
Second, we should invade Afghanistan, and
kick out the Taliban and install a democratic government. I don't care if they are responsible or not, I just hate those guys.
Third, we should take the Palestinians and move them to Saudi Arabia, and declare Saudi Arabia the new Jerusalem. Mideast problem is now solved: Palestinians are rich and don't have anything to complain about anymore, Saudi corrupt Islamic theocracy is gone, Saudi royal family are out on their butts, Israel is now just a patch of desert that no one wants.
http://www.plathome.co.jp/products/openblocks/inde x.html
Also check out
http://www.axis.com I've used their embedded linux ETRAX100LX boards, they work great. Single
ethernet, but it comes with entire cross compiler
and flash boot loader that works over ethernet
The CRT usually stopped working because a big capacitor in the drive circuit on the board blew out, you can often just replace this.
In the 1970's my father made a visit to Brazil, where
as you know they are serious about coffee. He
invented a coffee strength tool using the same principle, but
somewhat simpler, no computer required -- you just take a plastic ruler
and dip it into the coffee, and then read off the
strength by seeing where the last tick on the ruler is that you can still see. A manual optical
strength meter.
My Mitubishi i-mode phone runs more than 400 hours on a charge (full charge in two hours).
The writing seems, well, somehow it just seems like a hoax. The details are too neat, the
style of writing a little too offhand, the interviews with people don't ring true. I dunno, it just doesn't have the "ring of truth" to it.
I think one thing that people haven't quite realized yet is how very very easy it is for a company like Microsoft to take a "public" standard, such as Kerberos, DNS, NFS, SMTP, or whatever, and make it incompatible. It takes just one little change to the protocol, in the guise of "innovation" to make it incompatible.
This is one of their real sources of leverage. They can claim they are "innovating", while
magically making a proprietary product which is a parasite off of existing standards. In other words, today's Internet protocols are sort of like "GPL" code, implicitly, because you need to
make your code obey their "license", in order to interoperate. When someone takes a protocol private, they can get a free ride from the existing implementations, i.e.., their code seems to interoperate, but the user becomes dependent on some new feature, and is suddenly locked in to the proprietary product.
I think that the public, who has not had intimate experience with building and using network protocols and APIs, has little idea of how easy it is to "pollute" an existing standard. Indeed, many protocols, such as SMTP, are subject to "accidental" pollution, where implementors with good intentions still try to stick in a new "feature" into their code.
In the past Microsoft has, and does, make slightly incompatible versions of such basic protocols as SMTP. I have personally experienced plenty of failures as my own legally formatted SMTP messages are chewed up and spit out garbled by Microsoft mail handling products (just "bugs" in their MIME handling, or "embrace and extend"?).
Anyway, I think this is another silent front in Microsoft's war on humanity, and people should realize that the real damage is happening without loud obnoxious announcements by Microsoft management.
http://nooper.co.jp/showcase/gallery.php?s=15&l=en
I thought this was already posted on slashdot last week though...
How about this idea- no case at all: Take a big piece of plywood, and attach the motherboard and disk drives, and take the power supply, remove it's sheet metal case and fan, and just hang the whole thing on the wall. The only thing making noise will be the disk drive.
The magic of modern semiconductor lithography is
that the same area of silicon that contained
a 4 bit CPU in 1980 can contain a x386 compatible
PC and all motherboard functions plus network controller. At some point, as the chip circuit densities
increase the packaging and
bonding for the semiconductor become the
dominant cost, not the chip area.
So, as devices require more sophisticated networking
and other features, it makes sense to use
a controller with enough oomph for the job. Bumming instructions in assembly language is
getting to be less and less necessary, and is
actually a waste of developer's time, if the
silicon is cheap enough.
Hey, I'm living in the future already; I keep
.NET. I just need TCP/I-fucking-P that
my home directory under CVS, and have it updated
every evening to a remote server, and then
replicated to all my other devices (laptops,
pc's at work, etc).
I don't need sun NFS, I a sure as hell don't need
Microsoft
works and free software. If it were up to Microsoft, they would extend TCP/IP to have
some incompatible Microsoft extensions. No wait, they already did!
to regulate the use of this information.
Imagagine this scenario: In your neighborhood everyone has cable modem. All your neighbors hook
up some web cams pointing out their windows, to
form the "neighborhood net".
Are you now going to say this is illegal, and
send the police to take away their cameras?
Wouldn't that be an even worse violation of
civil rights than what I propose?
You have proposed no specific regulations which
will help preserve individual rights. Here are a couple of proposals off the top of my head:
an individual.
networks to track and individual without a court order.
public surveillance network for purposes of tracking an individual, without a court order.
These are probably flawed, but at least I am trying. What is your proposal? Private citizens can already have a"neigborhood watch" group. Are you going to outlaw that as well?