spam

[John Munro]

If you're designing a replacement for email from the ground up,
shouldn't that include designing out spam? The only reason there's
email spam is because when email was designed they never imagined
something like spam would happen.

[Alex Esplin]

Even when email was invented, there was junk snail-mail, I just don't
think it was understood how easy it would become for spammers to mass-
abuse the system.

What Wave does to prevent this would be a very interesting
explanation, IMO.

[Tye]

Check the right end of the toolbar on the "Snapshots!" gadget.
http://www.crunchbase.com/assets/images/original/0004/5813/45813v1.png

There's a "Spam!" button :)

[Erick]

That's what I think the most exciting aspect of this is. I've been
saying for years that the inherent security flaws in SMTP alone are
reason enough to make it a non-viable communication standard for the
future. Its good to see that someone is finally taking up the helm in
the search for a replacement. I haven't gotten the chance to fully
review the protocol spec yet, but lets hope that the final version
takes security into consideration from the ground up.

[Edgar doiron]

well since you invite people to your wave.... spam shouldn't be able
to get in, no?

[Sylvain Munaut]

On May 28, 2:52 pm, Edgar doiron <edgar.doi...@gmail.com> wrote:
> well since you invite people to your wave.... spam shouldn't be able
> to get in, no?

A spammer could 'invite' you to one of his wave.
But he would need a Wave server and that server would most likely be
the master for that wave. (i.e. the spammer has to keep states
himself !) which may get impractical if you want to invite millions of
people in a wave (or create millions of wave) ...

OTOH, maybe that state could be regenerated each time and not stored.

(from what I understand ... I haven't read the federation protocol
docs just yet)

Sylvain

[sshaukat]

In the press conference right after the keynote, a reporter asked
about spam prevention. Lars Rasmussen responded that it hasn't been
given much thought yet, since it is a closed developer's preview for
now, but also mentioned that most likely Wave would use a whitelist
option, where you'd have to add a friend/coworker before they could
send/invite you to Waves.

[Kiran Mudiam]

The spam filtering in Gmail has come a long way, I am pretty sure that
there can be a bot that will filter spam out using the API(if exposed)
You could perhaps add the spambot to your wave as well, to remove spam
automagically!

[aris...@googlemail.com]

I think that spam prevention has to be absolutely key - as others have
said gmail has an excellent anti-spam system - could they not
collaborate?

[JB]

Well, this is a reason why I think systems like Twitter is superior.
It is inherently more spam resistant as you "sub" to the conversation
instead of a free for all like email where anyone can send you a
message.

On May 29, 2:39 am, "ariste...@googlemail.com"

[Daniel]

It is exciting to see a potential replacement for email as the SMTP
protocol is fundamentally flawed and the sooner we all ditch our email
clients and upgrade to a new system, the better. But for my (SME)
clients, the two biggest problems with email are authenticity and
spam. I can send an email from barack...@whitehouse.gov and it is
almost impossible for an average user to prove who that email actually
came from. This makes spam possible. The second problem is that there
is a very low cost for sending email, which makes the low conversion
rate of spam (from emails sent to sales won) viable, when you can send
1,000,000 emails for only a few cents.

I implore this burgeoning engineering community to make it an
objective to design out Spam at the protocol level. If the release
version of Wave has a "report spam" or "white-list" button in the user
interface, then this objective will have failed.

In the SME space, it is real benefits to productivity like not having
to deal with spam and scams every day, that will convince the
carpenters, retailers and dentists of this world to move to a new
shiny system, not real-time collaborative editing (as cool as that
is!)

Errata: D.J Bernstein wrote a good analysis of the flaws of SMTP his
Internet Mail 2000 concept, before getting fed up and abandoning it:
http://cr.yp.to/im2000.html
I am sure many engineers (like me) have dreamed of one day rewriting
the SMTP protocol, but can't afford the time off work to do it! Now is
the chance to collaborate as a community to do this properly. Thanks
Google for taking the first step with Wave.


- Daniel Larsen
http://twitter.com/DanielLarsenNZ

[kdegussem]

Am I the only one here that has critics on the whitelist feature.

I see e-mail in a way where random person A can contact random person
B with an address. Just like telephones work.
There is not restriction whatsoever which is a good thing on its own.
But for example when you order something online, a company will send
you an email (or wave) but it is not in the whitelist, so the e-mail
will not be visible.
Example 2: Some long lost friend tries to get back in touch with you
but here e-mails will not come through because of the same reason
above.

I believe that if they implement whitelists against spam, they would
just adapt the same idea that is now being used for ordinary e-mail. I
also believe that people will only change protocols (and I mean the
normal folks) when there is an essential advantage. SMTP and current e-
mail is outdated, we all agree on that. But the biggest problem is
spam. So if google wave has an awnser against spam people would be
more motivated to change protocol. A more embedded and more up to date
protocol is nice and is appealing (also to me) but I can't stop
wondering whether this enough to move people for change. Google is
like the Obama on the Internet but I don't know if they can create the
same effect.

Also, when looking into coorporate life, will there be an interaction
between ordinary mail and wave? 99% users use e-mailservices. Will
companies be able to contact wave people through mail?

The way I see it, wave is more the next generation IM with advanced
features including ordinary mail on a different platform. Do I think
it is a full replacement for e-mail? No.

On 29 mei, 13:49, Daniel <daniellarse...@gmail.com> wrote:
> It is exciting to see a potential replacement for email as the SMTP
> protocol is fundamentally flawed and the sooner we all ditch our email
> clients and upgrade to a new system, the better. But for my (SME)
> clients, the two biggest problems with email are authenticity and

> spam. I can send an email from barack.ob...@whitehouse.gov and it is

> almost impossible for an average user to prove who that email actually
> came from. This makes spam possible. The second problem is that there
> is a very low cost for sending email, which makes the low conversion
> rate of spam (from emails sent to sales won) viable, when you can send
> 1,000,000 emails for only a few cents.
>
> I implore this burgeoning engineering community to make it an
> objective to design out Spam at the protocol level. If the release
> version of Wave has a "report spam" or "white-list" button in the user
> interface, then this objective will have failed.
>
> In the SME space, it is real benefits to productivity like not having
> to deal with spam and scams every day, that will convince the
> carpenters, retailers and dentists of this world to move to a new
> shiny system, not real-time collaborative editing (as cool as that
> is!)
>
> Errata: D.J Bernstein wrote a good analysis of the flaws of SMTP his
> Internet Mail 2000 concept, before getting fed up and abandoning it:http://cr.yp.to/im2000.html
> I am sure many engineers (like me) have dreamed of one day rewriting
> the SMTP protocol, but can't afford the time off work to do it! Now is
> the chance to collaborate as a community to do this properly. Thanks
> Google for taking the first step with Wave.
>

> - Daniel Larsenhttp://twitter.com/DanielLarsenNZ

[Andrew]

Perhaps separating the communication into different classes would
resolve the issue with spam.

So let's say that you have your whitelist with *trusted* friends who
don't send you spam--hopefully anyway. That's great, fine, whatever.

But let's say you give out your business card to an imporatant contact
with a Wave address on it. How to make sure that you don't go on
someone's spam list? One way to avoid this problem is to have the
application generate one time use access keys, (something short like
'Ax7tg',) to print out with your business cards. That way even if a
spammer got hold of it, they could only send one message.

I think that could pretty much solve this whole spam mess in one fell
swoop. How about that? I'm pretty smart, me thinks!

> > > send/invite you to Waves.- Hide quoted text -
>
> - Show quoted text -

[Don-Martin]

> But let's say you give out your business card to an imporatant contact
> with a Wave address on it. How to make sure that you don't go on
> someone's spam list? One way to avoid this problem is to have the
> application generate one time use access keys, (something short like
> 'Ax7tg',) to print out with your business cards. That way even if a
> spammer got hold of it, they could only send one message.
>
> I think that could pretty much solve this whole spam mess in one fell
> swoop. How about that? I'm pretty smart, me thinks!

Either spam is solved or not. Spammer don't care about you, they care
about many. If they know for a fact that using this one-time Wave
address will get through, they will make a very good spam, scam, or
phishing message. I.e. problem not solved.

There is no such thing as more secure or less secure. Either it is
secure or not.

Instead of making it harder to spammer, it most be impossible for
spammers.

There is no such thing as too much work, when it comes to spammers.
Spam email is a billion dollar business.

The only way to solve spam is by cryptographically secure algorithms
and hashes.

* Google most set up a Wave Certificate Server, where each Wave server
most get a certificate from.

* Each user on a Wave server will get a certificate too.

* When a Wave is released to its destination Wave server, the server
will then check if this Wave is in fact written by who it clames, and
in fact comes from the Wave server, that it clames.

* What you get for free, is that the Wave haven't been tempered with
on the way.

[Joseph]

I pretty much was going to suggest what Don-Martin already has.

If every wave server has to be certified in some manner or another,
even if the certificates themselves are easy to get a hold of, these
can easily be blacklisted if a centralized system is implemented.
Problems with SPF and others like is that they are not mandatory and
therefore cannot be used as a broad measure against all servers.

While I can understand that spam could still emerge at some point a
system to block the rest of the connections once a mass mailout has
been detected would make spam 'capture' more efficient where a few
thousand would actually get the spam message instead of millions.

I highly encourage this to be given serious thought since it can make
or break the whole platform.

[John Munro]

I think you're on the right track.

I think both the spam button on the Wave UI and using GMail's spam
filtering technology are looking at this the wrong way - instead of
trying to filter out spam, Wave should be designed in such a way that
spam is fundamentally impossible.

I think the whitelist idea is flawed because there are so many use
cases for legitimately receiving unsolicited messages. You should be
able to publish Wave's equivalent to an email address freely without
fear of being inundated by spam.

I think two things Wave could do towards the goal of eliminating spam
are:

1. Eliminate impersonation - with email it's trivial to pretend to be
someone else by simply changing the "from" address. If I get a Wave
that says it came from my bank there should be no question that it
came from my bank. This may be where Don's certificate ideas come in.

2. Eliminate automation - no spammer is going to manually send out a
million emails. More importantly no spammer is going to manually
create a million dummy accounts to avoid being blacklisted. Captcha's
don't work, but additional checks could be added to account creation
to ensure identity; for example by checking a phone or credit card
number. This will make it a bit more tedious to create an account but
that may not be a big deal since it's an action that most people will
only ever do once.

[Don-Martin]

On May 29, 7:05 pm, Joseph <JosephDeB...@gmail.com> wrote:
> If every wave server has to be certified in some manner or another,
> even if the certificates themselves are easy to get a hold of, these
> can easily be blacklisted if a centralized system is implemented.

Hacking a Wave server, and steeling the certificate wouldn't help you,
as you would need a valid certificate from a user from that Wave
server aswell. And user certificates are not stored on the Wave
server.

Even steeling a friends certificate won't help you, as that cert only
validates correctly when signed by the Wave server it was created on.
I.e. hacking Wave servers for spamming, not possible.

> While I can understand that spam could still emerge at some point a
> system to block the rest of the connections once a mass mailout has
> been detected would make spam 'capture' more efficient where a few
> thousand would actually get the spam message instead of millions.
>
> I highly encourage this to be given serious thought since it can make
> or break the whole platform.

Yes, but how to get in touch with key developers?

[Don-Martin]

> 2. Eliminate automation - no spammer is going to manually send out a
> million emails.  More importantly no spammer is going to manually
> create a million dummy accounts to avoid being blacklisted.

That is the argument made over and over again by Microsoft, Sony, and
so on in regards to DRM (Digital Restrictions Management), and have
been proved wrong every time.

There is no such thing as "too much work". There are thousands of
brilliant programmers out there with evil intentions. And when the
purpose is being able to earn millions of Dollars, it will get done.

Look at the XBox 1 or HDVD or BluRay. BluRay had the strongest DRM
known to date, AACS and BD+. Hacked and broken.

Look at XBox 360 on the other hand. Here Microsoft learned that they
had to make it impossible to hack it, and not hard. The state right
now is, that you can not get Linux on XBox 360, because MS have made
the security correct. (Sadly for us.)

[John Munro]

> There is no such thing as "too much work". There are thousands of
> brilliant programmers out there with evil intentions. And when the
> purpose is being able to earn millions of Dollars, it will get done.

Which is why I suggested a solution that involves tying the Weave
account to a verified human, without needing a clever technical
solution. If you could only have one Weave account per phone or
credit card number then no amount of work is going to get you a
million accounts, short of buying a million phones or credit cards.

[Norman Nunley]

Looks like this might be a good article to read, regarding the
SPAM/bad Wave server problem:

http://www.waveprotocol.org/whitepapers/wave-protocol-verification

[Don-Martin]

On May 29, 9:57 pm, John Munro <ghost...@gmail.com> wrote:
> Which is why I suggested a solution that involves tying the Weave
> account to a verified human, without needing a clever technical
> solution.  If you could only have one Weave account per phone or
> credit card number then no amount of work is going to get you a
> million accounts, short of buying a million phones or credit cards.

That wouldn't work. I e.g. have an email that I use online, one I use
for family and friends, and then I have the one where I work.

So one Wave per person, is not enough.

[Don-Martin]

On May 29, 10:00 pm, Norman Nunley <nnun...@gmail.com> wrote:
> Looks like this might be a good article to read, regarding the
> SPAM/bad Wave server problem:
>
> http://www.waveprotocol.org/whitepapers/wave-protocol-verification

Very interesting!! Thanks.

[Jorge Vargas]

On May 29, 3:20 pm, Don-Martin <d2x...@gmail.com> wrote:
> On May 29, 7:05 pm, Joseph <JosephDeB...@gmail.com> wrote:
>
> > If every wave server has to be certified in some manner or another,
> > even if the certificates themselves are easy to get a hold of, these
> > can easily be blacklisted if a centralized system is implemented.
>
> Hacking a Wave server, and steeling the certificate wouldn't help you,
> as you would need a valid certificate from a user from that Wave
> server aswell. And user certificates are not stored on the Wave
> server.
>

Create an account on the server and then hack the server?

[Richard Clark]

All of us who receive spam in their inbox every day appreciate the
fervor with which you're approaching this issue, but the basic truth
is you cannot "design out" spam without making the system useless.
Spam buttons are here to stay.

Imagine the changes that you might make to SMTP that would "design
out" spam. They don't exist. We can't even manage a 1:1 mapping
between Real Live Humans and Passports, let alone a 1:1 map between
real live humans who don't spam and email/wave accounts.

The major problem is that actions that you might take in that
direction are, in many cases, damaging to a large number of other
situations we find desirable - robots that run games, manage mailing
lists, feed rss feeds to us, alert us of system outages etc.

Those steps that can be taken reasonably have been taken. My read of
the protocol docs indicate that all servers will be cryptographically
verified - there will be no more faking of sender addresses like we
have in SMTP, so banning participants from a spammy server will work
as intended. In addition, the nature of the wave/wavelet means that,
in many cases, spam can be removed backwards in time.

Imagine it more like a forum in which any participant can identify
something in the stream that shouldn't be there and visibly get rid of
it, and everyone benefits from the change. Obviously there are
permissions issues (being discussed in another thread), but the
essential nature of the system is different from email and gives spam
fighters significantly more leverage than we have in the existing mail
system.

The spam/antispam thing is a fight that will go on until the end of
time, but the design of the wave as it stands appears to have the
features we need to gain the upper hand for now.

[Bob Oliver Bigellow XLII]

I agree. Some people think that SPAM is just a sign of the FLAW in
the SYSTEM. The reality is, SPAM is really a sign of UNSCRUPULOUS
PEOPLE in a TECHNOLOGICAL WORLD. That's it. Just as you cannot
create a perfectly virus-proof system which also allows unfettered
abilities to programmers for GOOD applications, you cannot create a
perfectly spam-proof system which also allows unfettered communication
capabilities.

I think the steps to make SPAM more identifiable and traceable have
already been taken. The problem with SMTP which makes spam EASIER is
the fact that mail servers automatically trust the information they
are given, there is little or no real authentication of users, etc...
So, it is very easy to spoof emails. With Waves, it looks like there
is more handshaking involved with authenticating the users and the
information.

Now, certainly, there is nothing which will stop a "spammer" from
using malicious software to create tons of fake Wave accounts just for
spamming... but since these accounts will be authenticated... the
moment they are detected as spamming, they can be blocked. With email
today, if someone spoofs spam in such a way that it appears to come
from my email address, a spam system shouldn't just block my email
address as spam, because it wasn't really me. With Wave, if the spam
is coming from my authenticated account, then it must be me. (Or
perhaps a bad robot or gadget that I installed, which would be another
problem on its own.) In any case, while this doesn't prevent spam, it
makes it much easier to trace spam to its root and to block spam from
an originator.

As someone else also pointed out, the other problem with spam is that
the email file gets transmitted from server to server until the
recipient reads it. In the case of Wave, the recipient is instead
invited to participate in the Wave... if it gets automatically marked
as spam, or if the individual marks it as spam, then the spam just
sits there on the originating server.

Really, I think that Gmail has great spam filters and I still deal
with more physical junk mail today than I deal with email spam. I am
willing to bet that if Wave systems start springing up everywhere,
many people will enjoy Google's Wave system because it will have an
equally decent spam system.

Unless someone can come up with a way to completely prevent spam
while:

1) Allowing mystery user A to email mystery user B without needing to
be pre-authorized (i.e., a long lost friend contacting you)
2) Allowing individuals to have more than one account (home account,
work account, I might also be a webmaster and might setup several
points of contact for different purposes)
3) Making Wave accounts just as easy to sign up for as email accounts
are today (requiring more hoops, credit card authentication, etc...
just makes it more difficult for the little people to participate on a
business-level on the Internet, which would be detrimental to what the
Internet represents today... a means to empower the Davids of the
Internet world to become just as important as the Goliaths in the
corporate world)

...then Wave spam is going to be here to stay in some form or
another... let's just hope we can keep it to a reasonably manageable
form.

[CobraA1]

Just because you can't design something perfectly doesn't mean you
should just give up and stop trying.

And there are some things we *can* do that make a lot of sense that we
should've been doing a long time ago.

End to end encryption and digital signatures, for example. Maybe they
won't kill spam completely, but the computational complexity of
generating new keys all the time would certainly place a larger burden
on spammers, and would make spoofing legitimate users impossible.

"I agree. Some people think that SPAM is just a sign of the FLAW in
the SYSTEM."

In some senses, yes it is. It is actually technologically feasible to
create a whitelist system that allows recipients to decide things like
"I only ever want to receive messages from my family." Instant
Messaging clients already allow this. Email can't. Hopefully Wave will
allow this.

Granted, not everybody would like such a system where new people are
automatically excluded, but right now many things, especially the
email system, don't even give you that choice.

Okay, we can't get rid of spam *completely* if we want a low barrier
of entry for new people to the system. But it's certainly possible to
make life much harder for the spammers.

[captnemo]

The basic answer to this is logically obvious: If I have a service,
email or otherwise, where I am able to print an "address" on a
business card, in the Yellow Pages, or on a web site, and any stranger
in the world can, at any time in the future, send a message to that
"address" then spam is a possibility.

There are a zillion ways to get rid of spam if you can control who is
able to send you a message. But the ability to publish a fixed email
address (or wave address) that the whole world can use as much or as
little as they want is one of the great powers of email. If Wave
allows this mode of operation, that is, the publishing of an "address"
that anyone can send messages to, anyone can add blips to a wave, then
spam is possible. And the only way to get rid of spam is by
sufficiently intelligent filtering.

Period. That's all there is to it. There's no way around this
problem and I cannot see why this is not absolutely obvious to people.

The same problem exists with snail mail. Anyone can mail junk to your
house.

The same problem exists with telephones (unless you use whitelisting
by means of caller ID, which defeats the purpose of a published phone
number). If anyone in the world can call you then phone spam is going
to happen.

So forget about "eliminating spam". You are wasting your time. You
can either whitelist or have open access. Open access will bring
spam. Effort should be placed on better and better spam filters
because spam is not going away, ever, as long as there are openly
publishable addresses.

Phil

[Damian Guppy]

Like i have stated befor, i agree spam will always exist. There is even spam on most IM systems, though viruses that infect valid users etc.

 

Wave's protocol by its nature will pose a problem for spammers, as instead of with email where messages are stored on the recipients server, with waves the messages are stored on the senders server, putting a massive load on spammers. If a spammer cant keep their server up, when it goes down all the messages(spam) hosted on that server go down with it. It also makes blacklists alot more effective, and puts more pressure on WSP's to make sure their users are doing the right thing and not spamming.

 

Beyond this i see not much point worrying about the spam situation. However there will still probably be a market for extensions and robots that filter and flag spam for users.

[Andrew Over (Google)]

Hi Joseph.

On May 29, 10:05 am, Joseph <JosephDeB...@gmail.com> wrote:

> I highly encourage this to be given serious thought since it can make
> or break the whole platform.

Spam is being taken very seriously and has been given a lot of
thought. As has been mentioned, wave has attributes of several forms
of communication, and if unprotected, will inherit many of the same
abuse vectors. Being added to spam waves, wiki-style content
vandalism, and blogger-style comment spam are just the tip of the
iceberg; there's also hostile API users to worry about :-)

We're fully aware that any platform that achieves critical mass will
be subject to abuse and are working to prevent this from troubling
users.

Cheers,
--Andrew

[Jorge Vargas]

Hello Phil,

I agree with you in everything except one thing. This is a huge deal
in the current system the bot and person are 100% equal there is no
way of knowing who is who. That combined with the ability to modify
every aspect of the conversation is a great power on for both the good
and bad. We need a way to make this not turn into usenet.

[Brett Morgan]

Ok, why did Usenet die? Why did IRC die?

My view on the death of Usenet, IRC and various boards was due to the
ability of spammers to hijack high S:N forums with the paid for
advertising.

We have seen various approaches to eliminating the ability of hijack,
be it in the forms of /. style karma, the splintering of Usenet into
blogs, and the re-invention of irc as twitter. Note, each of these is
a re-invention of the user interface, in each instance requiring
someone who wants attention of the community to earn it.

This is something that can be implemented atop wave, as we already
have reasonably strong identity, at least to the server level, of
someone (human or bot) using wave. This is all we require to build
karma systems that allow for reasonably easy filtering of noise.

Thoughts?

>>
>> On May 30, 9:21 pm, CobraA1 <jeremiah.m...@gmail.com> wrote:
>>> Just because you can't design something perfectly doesn't mean you
>>> should just give up and stop trying.
>>>
>>> And there are some things we *can* do that make a lot of sense that we
>>> should've been doing a long time ago.
>>>
>>> End to end encryption and digital signatures, for example. Maybe they
>>> won't kill spam completely, but the computational complexity of
>>> generating new keys all the time would certainly place a larger burden
>>> on spammers, and would make spoofing legitimate users impossible.
>>>
>>> "I agree.  Some people think that SPAM is just a sign of the FLAW in
>>> the SYSTEM."
>>>
>>> In some senses, yes it is. It is actually technologically feasible to
>>> create a whitelist system that allows recipients to decide things like
>>> "I only ever want to receive messages from my family." Instant
>>> Messaging clients already allow this. Email can't. Hopefully Wave will
>>> allow this.
>>>
>>> Granted, not everybody would like such a system where new people are
>>> automatically excluded, but right now many things, especially the
>>> email system, don't even give you that choice.
>>>
>>> Okay, we can't get rid of spam *completely* if we want a low barrier
>>> of entry for new people to the system. But it's certainly possible to
>>> make life much harder for the spammers.
>> >
>>
>
> >
>

--

Brett Morgan http://brett.morgan.googlepages.com/

[Jorge Vargas]

I kindly disagree freenode is probably one of the best resources out
there in it's kind.

> My view on the death of Usenet, IRC and various boards was due to the
> ability of spammers to hijack high S:N forums with the paid for
> advertising.
>

yes agreed, and this is specially scary with robots in wave, as they
are legit from the start. I reallly don't want to expend my days
filtering robots as we do with spam these days.

> We have seen various approaches to eliminating the ability of hijack,
> be it in the forms of /. style karma, the splintering of Usenet into
> blogs, and the re-invention of irc as twitter.

irc is actually jabber, in irc you have interesting discussions. in
fact I think Usenet is more like IRC and blogs are an invention over
forums.

> Note, each of these is
> a re-invention of the user interface, in each instance requiring
> someone who wants attention of the community to earn it.
>

agreed.

> This is something that can be implemented atop wave, as we already
> have reasonably strong identity, at least to the server level, of
> someone (human or bot) using wave. This is all we require to build
> karma systems that allow for reasonably easy filtering of noise.
>

which means someone will have to rebuild akismet.

[rak...@gmail.com]

Why not have two inboxes, one that uses a whitelist, and a second that
is questionable. In the questionable inbox it can be assumed that any
wave that is sent out to a bunch of people, and almost none of which
are in the whitelist are probably spam.

With waves being editable and all by any user, the first user to see a
questionable email can then decided to whitelist or spam it. If the
wave has been sent to 100 people, and none of those people had the
sender in the whitelist and the first 20 people then say it is spam,
then the other 80 will likely agree and it should be put into their
spam boxes as well and the sender will be considered a spammer. Now a
good spammer will then find a way to have the first 20 wave users be
some bots they own and they will be on a whitelist, but the next 30
might be real people who say this is spam and again it should be then
put into the spam filter as a spammer. If a wave is truly not spam it
will have a consistent amount of people agree that its spam and it’s
this that should be used to conclude its spam. Those who want to white
list a spammer should then be followed by Google, and if they have
only spammers and questionable spammers on the whitelist, then they
should be considered as spammers themselves or spam helpers.

[Brett Morgan]

On Tue, Jun 2, 2009 at 1:39 PM, Jorge Vargas <jorge....@gmail.com> wrote:
>
> On Mon, Jun 1, 2009 at 11:00 PM, Brett Morgan <brett....@gmail.com> wrote:

[chomp]

>> Ok, why did Usenet die? Why did IRC die?
>>
> I kindly disagree freenode is probably one of the best resources out
> there in it's kind.

I should spend some time on freenode =)

[Jorge Vargas]

I think everyone commenting in this thread (I say that from my own
experience acquired about 5 minutes ago) Should go read the
whitepapers. for example

"A wave is identified by a globally unique wave id, which is a pair of
a domain name and an id string. The domain names the wave provider
where the wave originated."

so this pretty much means 50% of the spam problem is solved. You know
WHO and created the spam. Now we only need to solve the problem where
a phishing scam happens and then you are inundated with spam.

[Bastian Hoyer]

> "A wave is identified by a globally unique wave id, which is a pair of
> a domain name and an id string. The domain names the wave provider
> where the wave originated."
>
> so this pretty much means 50% of the spam problem is solved. You know
> WHO and created the spam.

most of the spam will probably come from big wave providers... they
won't use their own servers because they know that won't work.

I like the idea of two inboxes... it could be easily combined with
your contact list. If you get invited to a wave from one of your
contacts it directly goes to the primary inbox, otherwise it goes to
the "unknown" box.

There will be definitly be a way to distinguish between human and
robot... currently you can safely assume that every <name>@appspot.com
attendee is a robot.
Guess this is the reason why they only allow robots on app engine right now...

[Damian Guppy]

I think i spotted a google employee stating that they intend to distinguish bots from people. It wont solve the problem of people writing traditionol bots who connect to wsp's like a human client would though.

Now for my next point.

If lots of spam starts comming from big WSP's then it would be fair to assume that traditional DNS blacklists will exist for waves, and thoes WSP's would find themselves quickly on thoes blacklists (just like with email), this will p*** off thier legit users as they wont be able to talk to everyone they want to, and they will move providers, damaging the WSP's business model, weather it is ad funded or subscription based. Therefor it would be putting the onus back on WSP's that provide services to spammers to clean up their act, and problem is (partially) solved.

[Steve Hammer]

From the cursory examination of the public protocol documentation, it
appears that faking the source address of a wave would be extremely
difficult. As such, a system where a user could block a specific
participant (or participants domain), combine with a heuristic system
(similar to what is provided by gmail currently), would be a better
solution.

Server wide blacklists could be a huge headache. There would be a lot
of overhead for maintaining the list and dealing with adding,
verifying, distributing it. A malicious individual might also be able
to get providers falsely inserted into the list (perhaps even Google),
which would be bad.

I think that allowing the block to occur at the user/participant level
would be a better method than to block at the server/protocol level
IMHO.

-Steve

[rak...@gmail.com]

> I think that allowing the block to occur at the user/participant level
> would be a better method than to block at the server/protocol level
> IMHO.

Most likely they will only ban a server, if they feel it is only for
spam. for the most part it should only be users only that get banned

[Steve Hammer]

On Tue, Jun 2, 2009 at 3:37 PM, rak...@gmail.com <rak...@gmail.com> wrote:
> Most likely they will only ban a server, if they feel it is only for
> spam. for the most part it should only be users only that get banned

True.

I wonder if there should be a method for a remote user to see if they
or their site is banned, but that is a discussion for the protocol
group.

[captnemo]

Hello Jorge,

Yeah, I was just reacting to statements along the line of "eliminate
spam" which I argue is an impossible goal if one is to have
publishable "email" addresses that anyone can send to. And I
attempted to point out the "laws of physics" that should make this
fact obvious.

In spite of the above, of course, we should do everything in our power
to make spam as difficult and as traceable as possible. The way that
Wave works is a nice step in that direction on a couple of fronts. I
expect that spammers will not be real happy when Wave starts to take
over.

As an aside, I have not looked at usenet much in a long time. I used
to use it a lot. But IRC is very much alive and well. Best resource
around for near instant answers to obscure technical questions
regarding programming, languages, Linux, etc. Recently I've become so
busy I have no time to hang out on IRC and help answer the questions
of others but I hope to do it again eventually.

Phil

[dion]

On Jun 3, 8:09 am, captnemo <1citizenofthewo...@gmail.com> wrote:
> Yeah, I was just reacting to statements along the line of "eliminatespam" which I argue is an impossible goal if one is to have

> publishable "email" addresses that anyone can send to.  And I
> attempted to point out the "laws of physics" that should make this
> fact obvious.

You are absolutely correct that it's impossible to prevent every
single spam wave from hitting a victim, but the point of spam is that
it's easy and cheap to hit millions of victims, so even with a very
low conversion rate you can make a profit.

With wave you have to keep the originating server on-line and off
blacklists for the entire lifetime of the wave and it's very hard to
spoof someone, so I'd argue that spamming will be much less of a
problem to begin with, however it's entirely possible that someone
takes over a client machine via a trojan of some sort and starts
abusing that users account, so spamming could still happen.

The hijacking of a legitimate account is a real problem, but by nature
spam needs to be sent to a very large number of victims to have any
chance of turning a profit, so all that's needed to get around this
hole is to make it slightly more expensive for someone to send a wave
to someone if they have never communicated before.

I think the best solution would be to build a hashcash payment system
into the protocol and have the recipient charge the sender a fee of
say 100 ms or so of CPU time if the sender isn't already in the
recipients whitelist, that would mean that the spammer would need to
peg a CPU for an entire day to spam just 1 M people.

The best part is that all of this would be completely transparent and
unobtrusive to the user, except when they first wave someone and they
have to heat their CPU for a 0.1 second while it churns out a hashcash
payment in the background.

It should be up to the recipient how much postage to charge the
sender, that way the postage can be cranked up in sync with faster
CPUs from Intel and new releases of V8.

Please have a peek at http://www.hashcash.org/

[Bastian Hoyer]

Well ... in the keynote they said that you can configure wave to only
accept waves from your contacts... that way you can be sure you don't
get any spam ;)

[David Fuelling]

Very timely post. 

I recently posted a very-draft-spec that formalizes something very close to your idea.  It's called OInvite (http://oinvite.net), and it provides a protocol for sending and receiving spam-free, cross-domain "friend requests". 

Basically, in a social-network-like system (like Wave), message spam is less of an issue because of the friend-list, which provides each user with the ability to "mute" a friend, or stop being friends with someone, thus eliminating the notion of spam (because we all "elect" to receive messages or waves from the people we receive them from, due to the friend list).

All that to say, there is still a serious spam threat at the point of "first-contact", which is where somebody you've never heard of wants to send/receive waves with you.  This is where the Hashcash idea, and Proof-of-Work (POW) in general, can be very powerful in preventing "friend request" spam.

That said, Google's own Ben Laurie has co-published a very provocative paper called "Proof of Work Proves not to Work" (http://oinvite.googlecode.com/files/ProofofWorkNoWork.pdf).  In it, the authors make a very persuasive case that POW mechanisms are not feasible in a general messaging system (like email) because in order to be effective, they harm legitimate user's ability to send/receive messages.

However, a follow-up paper (by different authors) outlines how Proof-of-Work could actually work if the "postage" was variable -- i.e., the first time a particular sender is encountered, postage is high, and later, as the sender proves his "un-spammy-ness", postage for him goes down (very similar to the idea suggested below, except postage calculation is more automatic based on your "reputation").  See here for that paper: http://oinvite.googlecode.com/files/ProofofWorkCanWorks.pdf.

Oinvite defines a way for invitation requests (i.e., friend requests) to contain a hash-cash compatible (though extended) token indicating a certain amount of postage.  I'm also working on a spec (an extension to OInvite code) that defines a way for individual user's to advertise the minimum "postage" required before they will even accept an OInvite from a user not on their friend list.

With these two pieces in place, the next step would be to setup some sort of economics system to make the OInvite Minimum Postage a variable number, based on the amount of spam a user is receiving.  If a user sets his "postage" at 20 bits, and is receiving Invitation spam, then a computer could automatically increase that user's advertised postage until the spam stops (measured by how many messages get flagged as spam).  In tandem, the computer could be updating that user's friends with increased advertised postage rates to help automatically reduce their spam, too.

All in all, the minimum postage required to get a "friend request" in front of my eyes should be controlled by me (the recipient).

Lastly, POW may not be the most effective mechanism to prevent "first-contact" spam.  So, OInvite allows for a pluggable "verification mechanism", so the community can decide what the best mechanism is.

I'd appreciate any input and participation!  The whole idea was released on Google Code about 3 days ago (ironic):  http://oinvite.net (forwards to google code).

Thanks!

David

[captnemo]

Exactly. I agree. That's what I was referring to somewhere above
where I said that Wave's design is inherently an improvement.

As you rightly point out, right now it's easy and effortless for a
spammer to jump online, pump out tons of spam for a few hours, then
disappear. That doesn't work with a Wave.

Since email spam works on razor thin margins and currently has no
ongoing "overhead cost" after emails are sent, unless I'm missing
something in the equation, Wave will pretty much ruin the game for
spammers. And botnets are no help either because Wave servers have to
stick around and be visible.

And unless I misunderstand the protocol, a Wave interaction is far
more complex and time-consuming than the "fire and forget" nature of
SMTP so this greatly hampers the speed at which spams can be sent.

So yeah, I expect that Wave pretty much ruins the game for spammers.
It cannot eliminate spam but it raises the cost of doing business way
above what spammers can afford.

Phil

[rak...@gmail.com]

I would argue that this method would not be a very good one, because
it is not the users computer that will be "paying the bill" it is the
server that is running/hosting the wave. If this comes out on servers
like myspace and facebook, then a spammer could just create a myspace
account, send out a lot of spam and then forget about the account.
They do not have to "pay the cost" it would be the myspace server
paying the cost for the waves. Now if wave was set up on individual
computers, then this method would be a much better solution.

> Google Code about 3 days ago (ironic):  http://oinvite.net(forwards to

[Jorge Vargas]

On Sat, Jun 6, 2009 at 8:39 PM, rak...@gmail.com<rak...@gmail.com> wrote:
>
> I would argue that this method would not be a very good one, because
> it is not the users computer that will be "paying the bill" it is the
> server that is running/hosting the wave. If this comes out on servers
> like myspace and facebook, then a spammer could just create a myspace
> account, send out a lot of spam and then forget about the account.
> They do not have to "pay the cost" it would be the myspace server
> paying the cost for the waves. Now if wave was set up on individual
> computers, then this method would be a much better solution.
>

That is true, but it is missing the cleanup effort. In current day
email. If I send 1000 spam emails that means 1000 people/programs/etc.
need to flag/delete/let sit forever that message. On a wave system
since all the waves are simply copies of the original wave, ones your
"provider" locates that account as a spammer and deletes it from his
system ALL copies of it are gone. Which means 1000 spam waves need to
be clean up by 1 person/program/organization.

So the only thing left is (and this happens in mmo with gold sellers)
is that they will create bots to create account to send spam and never
use that bot. But again that's too limited because all the waves are
connected. The moment one person/program/organization flags a wave as
spam everyone knows that is it, if all/most participants agree it's
spam, then it's gone from all systems.

Which together with all the other reasons posted here is another
direct stab at the business model of spammers.

So the fun question to ask now is, how are they going to evolve to be
more profitable :)

[dion]

On Jun 7, 2:39 am, "rak...@gmail.com" <rak...@gmail.com> wrote:
> ... because it is not the users computer that will be "paying the bill" it is the

> server that is running/hosting the wave.

Well, that would be up to the wave provider, I doubt any of them would
be willing to do that, so I'm pretty sure the work would be pushed to
the clients, no matter what,


> ... If this comes out on servers like myspace and facebook, then a spammer

> could just create a myspace account, send out a lot of spam

Correct, but nobody said that legacy messaging systems have to be easy
to use forever.

Email is close to useless today because of spam already so there will
have to be strong anti-spam measures in place for the email gateway.

As time goes on I'm sure that wave users who get too much spam from
legacy systems will add capchas and other bothersome checks for legacy
users who want to contact them from particular legacy systems.


In 2 years first time correspondents who email me will get a bounce
saying:

Hi, nice to hear from you, I don't believe we have ever talked.

To to protect my wave account from spam, please use one of the
following solutions:
* Use wave directly.
* Visit this webpage: http://dren.dk/spamtrap?id=42
* Simply re-send your original message with the result of 6*7
somewhere in the message.
* Set up your email client to add hashcash postage to every message
and resend your message.

[Bastian Hoyer]

2009/6/7 dion <dre...@gmail.com>:

> Email is close to useless today because of spam already so there will
> have to be strong anti-spam measures in place for the email gateway.

Is something wrong with your gmail account ? I think I get one spam
message in my inbox about once a week .)

An easy solution to keep wave clean from spam mails is to not import
mails into wave at all ;9

[injuryprone]

If Wave doesn't end communication to and from false or forged
identities then it wasn't designed for the 21st century.

[David Fuelling]

On Sun, Jun 7, 2009 at 4:44 AM, Jorge Vargas <jorge....@gmail.com> wrote:

On a wave system since all the waves are simply copies of the original wave, ones your
"provider" locates that account as a spammer and deletes it from his
system ALL copies of it are gone. Which means 1000 spam waves need to
be clean up by 1 person/program/organization.

I'm not convinced this is entirely accurate.  Remember that every server holds a "copy" of the original wave, so deleting the original wave will not automatically delete the "copies".  If it did, there would be some serious "control over my own information" issues with this protocol.  For example, what if my organization has different spam rules than your organization?  Does my organization really want to allow your organization to be able to "delete" things on my organization's servers?  Not to mention formal compliance issues (like keeping a backup of certain messages, no matter what).

My hunch is that each organization will make up its own mind about when/how to delete things from its own servers (though if any person in a wave marks the wave as spam, especially the creator of the wave, then other participants should at least take this information into consideration.

Formally speaking, I'm not entirely clear what's supposed to happen to a wave on a remote system when the "creator" on a different remote system deletes the wave.

[David Fuelling]

On Sun, Jun 7, 2009 at 12:39 AM, rak...@gmail.com <rak...@gmail.com> wrote:

I would argue that this method would not be a very good one, because
it is not the users computer that will be "paying the bill" it is the
server that is running/hosting the wave.

Not necessarily.  The burden of work could (like another poster mentioned) be pushed onto the client, though the "postage" only needs to be "paid" at first-contact, not with each message/wavelet.  Additionally, if a user's address is already in the recipient's "contact list" or friend-list, then no "postage" is required.   For a system like wave, my guess is many users will import their existing address-book/contact list, meaning messages could get through to their existing contacts at no extra "cost".  OInvite is only for the anonymous, first-contact case (which I would argue is the only remaining point of spam in a system like wave).
 

If this comes out on servers like myspace and facebook, then a spammer could just create a myspace account, send out a lot of spam and then forget about the account. They do not have to "pay the cost" it would be the myspace server paying the cost for the waves.

There are several problems with this line of thinking:

1. MySpace/Facebook won't allow a single account to send out millions of messages.  They already have _outbound_ throttling controls (technically, MySpace/Facebook are closed, so they have _internal_ throttling controls, which would function the same if they started letting _outbound_ requests go out).  In a federated system, it's the _in-bound_ throttling that becomes a problem.  OInvite basically says, a.) "Don't let any messages/waves through that come from a sender not on the recipient's friend-list" (i.e., there's no such thing as message-spam anymore); and b.) if a sender wants to get on a recipient's friend-list, check their POW token (or some other mechanism like the sender's reputation at the recieving provider.  e.g., how many other Facebook user's have this new sender on their friend list?  17?   Ok, then this sender is probably not a spammer since 17 other people added him as a "friend").
   
   
2. If a spammer tries to make millions of accounts, and send only a few "friend-requests" from each account, then this is a problem with or without OInvite, in Wave, GMail, or wherever, and demands a better registration process (e.g., Gmail's registration process, which requires you to receive an SMS to get a gmail account).  With OpenID registration, there could even be reputation information from a trusted identity provider that could help a MySpace/Facebook etc confirm that the user is a "regular" user, and not some user trying to send out 10 spam and disappear.
   
   
3. In federated systems like Google wave, __message spam__ will go away because only people on a friend list (i.e., invited to participate) will be allowed to send wave messages.  It's the "hey, add me to your friend list" spam problem that google wave isn't (apparetly) yet solving.
   

[d.brophy]

Come on people! Surely after 52 messages someone would have suggested
the obvious spam blocker?

If you're not on someone's whitelist, the first time you invite them
to a wave, their WSP sends back a message containing a CAPTCHA image
(http://www.captcha.net/). The wave invitation remains in a special
"Maybe spam" folder until you successfully complete the CAPTCHA.
Completing the CAPTCHA puts you on their whitelist, so subsequent
invites go straight to their inbox.

This system isn't new - it's been tried in email before. It didn't
really work well because it wasn't implemented in the mail reader
clients. The CAPTCHA requests where sent back as emails, which
regularly got marked as spam themselves. With tight integration into
the client software from day 1, this would be an excellent way to get
round spam.

The slight problem is that it's a real headache for companies like
amazon, who send thousands of legitimate messages every day to people
who make purchases. This could be mitigated by an easy "put me on your
whitelist" procedure that you're invited to do during your purchase.

[njoroge...@gmail.com]

I hate SPAM!!

On May 30, 10:40 pm, Bob Oliver Bigellow XLII <iam...@gmail.com>
wrote:

> I agree.  Some people think that SPAM is just a sign of the FLAW in

> the SYSTEM.  The reality is, SPAM is really a sign of UNSCRUPULOUS
> PEOPLE in a TECHNOLOGICAL WORLD.  That's it.  Just as you cannot
> create a perfectly virus-proof system which also allows unfettered
> abilities to programmers for GOOD applications, you cannot create a
> perfectly spam-proof system which also allows unfettered communication
> capabilities.
>
> I think the steps to make SPAM more identifiable and traceable have
> already been taken.  The problem with SMTP which makes spam EASIER is
> the fact that mail servers automatically trust the information they
> are given, there is little or no real authentication of users, etc...
> So, it is very easy to spoof emails.  With Waves, it looks like there
> is more handshaking involved with authenticating the users and the
> information.
>
> Now, certainly, there is nothing which will stop a "spammer" from
> using malicious software to create tons of fake Wave accounts just for
> spamming... but since these accounts will be authenticated... the
> moment they are detected as spamming, they can be blocked.  With email
> today, if someone spoofs spam in such a way that it appears to come
> from my email address, a spam system shouldn't just block my email
> address as spam, because it wasn't really me.  With Wave, if the spam
> is coming from my authenticated account, then it must be me.  (Or
> perhaps a bad robot or gadget that I installed, which would be another
> problem on its own.)  In any case, while this doesn't prevent spam, it
> makes it much easier to trace spam to its root and to block spam from
> an originator.
>
> As someone else also pointed out, the other problem with spam is that
> the email file gets transmitted from server to server until the
> recipient reads it.  In the case of Wave, the recipient is instead
> invited to participate in the Wave... if it gets automatically marked
> as spam, or if the individual marks it as spam, then the spam just
> sits there on the originating server.
>
> Really, I think that Gmail has great spam filters and I still deal
> with more physical junk mail today than I deal with email spam.  I am
> willing to bet that if Wave systems start springing up everywhere,
> many people will enjoy Google's Wave system because it will have an
> equally decent spam system.
>
> Unless someone can come up with a way to completely prevent spam
> while:
>
> 1) Allowing mystery user A to email mystery user B without needing to
> be pre-authorized (i.e., a long lost friend contacting you)
> 2) Allowing individuals to have more than one account (home account,
> work account, I might also be a webmaster and might setup several
> points of contact for different purposes)
> 3) Making Wave accounts just as easy to sign up for as email accounts
> are today (requiring more hoops, credit card authentication, etc...
> just makes it more difficult for the little people to participate on a
> business-level on the Internet, which would be detrimental to what the
> Internet represents today... a means to empower the Davids of the
> Internet world to become just as important as the Goliaths in the
> corporate world)
>
> ...then Wave spam is going to be here to stay in some form or
> another... let's just hope we can keep it to a reasonably manageable
> form.
>
> On May 30, 6:38 pm, Richard Clark <richard.cl...@gmail.com> wrote:
>
>
>
> > All of us who receive spam in their inbox every day appreciate the
> > fervor with which you're approaching this issue, but the basic truth
> > is you cannot "design out" spam without making the system useless.
> > Spam buttons are here to stay.
>
> > Imagine the changes that you might make to SMTP that would "design
> > out" spam. They don't exist. We can't even manage a 1:1 mapping
> > between Real Live Humans and Passports, let alone a 1:1 map between
> > real live humans who don't spam and email/wave accounts.
>
> > The major problem is that actions that you might take in that
> > direction are, in many cases, damaging to a large number of other
> > situations we find desirable - robots that run games, manage mailing
> > lists, feed rss feeds to us, alert us of system outages etc.
>
> > Those steps that can be taken reasonably have been taken. My read of
> > the protocol docs indicate that all servers will be cryptographically
> > verified - there will be no more faking of sender addresses like we
> > have in SMTP, so banning participants from a spammy server will work
> > as intended. In addition, the nature of the wave/wavelet means that,
> > in many cases, spam can be removed backwards in time.
>
> > Imagine it more like a forum in which any participant can identify
> > something in the stream that shouldn't be there and visibly get rid of
> > it, and everyone benefits from the change. Obviously there are
> > permissions issues (being discussed in another thread), but the
> > essential nature of the system is different from email and gives spam
> > fighters significantly more leverage than we have in the existing mail
> > system.
>
> > The spam/antispam thing is a fight that will go on until the end of
> > time, but the design of the wave as it stands appears to have the
> > features we need to gain the upper hand for now.
>
> > On May 30, 5:22 am, John Munro <ghost...@gmail.com> wrote:
>
> > > I think you're on the right track.
>
> > > I think both the spam button on the Wave UI and using GMail's spam
> > > filtering technology are looking at this the wrong way - instead of
> > > trying to filter out spam, Wave should be designed in such a way that
> > > spam is fundamentally impossible.
>
> > > I think the whitelist idea is flawed because there are so many use
> > > cases for legitimately receiving unsolicited messages.  You should be
> > > able to publish Wave's equivalent to an email address freely without
> > > fear of being inundated by spam.
>
> > > I think two things Wave could do towards the goal of eliminating spam
> > > are:
>
> > > 1. Eliminate impersonation - with email it's trivial to pretend to be
> > > someone else by simply changing the "from" address.  If I get a Wave
> > > that says it came from my bank there should be no question that it
> > > came from my bank.  This may be where Don's certificate ideas come in.
>
> > > 2. Eliminate automation - no spammer is going to manually send out a
> > > million emails.  More importantly no spammer is going to manually
> > > create a million dummy accounts to avoid being blacklisted.  Captcha's
> > > don't work, but additional checks could be added to account creation
> > > to ensure identity; for example by checking a phone or credit card
> > > number.  This will make it a bit more tedious to create an account but
> > > that may not be a big deal since it's an action that most people will
> > > only ever do once.- Hide quoted text -
>
> - Show quoted text -

[Bastian Hoyer]

or they just could continue to send emails. I don't see any advantage
for using wave for newsletter or confirmation mails where noone is
expecting any answer or collaboration

2009/6/8 d.brophy <da...@dontstayin.com>:

[rak...@gmail.com]

On Jun 7, 6:25 pm, David Fuelling <sappe...@gmail.com> wrote:

> On Sun, Jun 7, 2009 at 4:44 AM, Jorge Vargas <jorge.var...@gmail.com> wrote:
> > On a wave system since all the waves are simply copies of the original
> > wave, ones your
> > "provider" locates that account as a spammer and deletes it from his
> > system ALL copies of it are gone. Which means 1000 spam waves need to
> > be clean up by 1 person/program/organization.
>
> I'm not convinced this is entirely accurate.  Remember that every server
> holds a "copy" of the original wave, so deleting the original wave will not
> automatically delete the "copies".  If it did, there would be some serious
> "control over my own information" issues with this protocol.  For example,
> what if my organization has different spam rules than your organization?
> Does my organization really want to allow your organization to be able to
> "delete" things on my organization's servers?  Not to mention formal
> compliance issues (like keeping a backup of certain messages, no matter
> what).

A wave will not likely be deleted. wave does already have a spam
folder as well as an inbox, i did suggest earlier that it should add a
third box for questionable spam as well. In a case that you bring up
on your server it would then be indicated in the questionable box
while the server that considered it spam will place it in spam. Of
course even if it is in the spam box you can still read it; this is
necessary since its a collaboration tool as well.

[rak...@gmail.com]

On Jun 8, 7:41 am, "d.brophy" <d...@dontstayin.com> wrote:
> If you're not on someone's whitelist, the first time you invite them
> to a wave, their WSP sends back a message containing a CAPTCHA image
> (http://www.captcha.net/). The wave invitation remains in a special
> "Maybe spam" folder until you successfully complete the CAPTCHA.
> Completing the CAPTCHA puts you on their whitelist, so subsequent
> invites go straight to their inbox.

I think you miss the point of the "Maybe spam" folder, its not up to
the sender of the wave to decided this is or is not spam. If i send a
wave invite to 50 people, and none have me on their white list, then
my wave goes into the "Maybe spam" folder for all of them. I have no
control over this and no way of getting it out of that folder, other
then contacting the people and letting them know that i am sending
them a wave for the first time please add me to your white list. If
someone does not care to add me to their white list then all my
subsequent waves that I invite them to will also go into their "Maybe
spam" folder unless they decided to black list me, at which point all
my waves to them will be in a spam folder; this would not mean every
person I sent a wave to will not receive it though, as only they will
consider me spam.

> This system isn't new - it's been tried in email before. It didn't
> really work well because it wasn't implemented in the mail reader
> clients. The CAPTCHA requests where sent back as emails, which
> regularly got marked as spam themselves. With tight integration into
> the client software from day 1, this would be an excellent way to get
> round spam.
>
> The slight problem is that it's a real headache for companies like
> amazon, who send thousands of legitimate  messages every day to people
> who make purchases. This could be mitigated by an easy "put me on your
> whitelist" procedure that you're invited to do during your purchase.

Amazon is not a big problem with waves, remember if they really want
amazon can be their own wave server. As their own wave server they do
not have to worry about being black listed, it is the same as Tom on
myspace being blacklisted on myspace...

[Dmitry Unkovsky]

Having amazon on mind. Why it would need to "send" wavelet? It has a
collection of wavelets. You subscribe for those of your interest. You
see new entries, updates etc. Some implementation of paging will cope
with ones having long history.
No spam, no confirmation. All commercials you get are only those you
are interested in. If not, you drop that wavelet. What more to wish
for?
Quite the same applies to services like facebook - their accounts or
topics are considered as permanent wavelets that you can subscribe and
receive updates of or send your updates to them.

So, let me state that all spam topic is really about spam-invitations
and misbehaving accepted participants. Isn't it?

If agreed with that, "maybe spam" seems to be the first and simplest
thing to do.
May be, next - some kind of distributed spam-participants database.
Not sure. Have to think more on it.

[David Fuelling]

On Mon, Jun 8, 2009 at 3:31 PM, rak...@gmail.com <rak...@gmail.com> wrote:

On Jun 8, 7:41 am, "d.brophy" <d...@dontstayin.com> wrote:
> If you're not on someone's whitelist, the first time you invite them
> to a wave, their WSP sends back a message containing a CAPTCHA image
> (http://www.captcha.net/). The wave invitation remains in a special
> "Maybe spam" folder until you successfully complete the CAPTCHA.
> Completing the CAPTCHA puts you on their whitelist, so subsequent
> invites go straight to their inbox.

I think you miss the point of the "Maybe spam" folder, its not up to
the sender of the wave to decided this is or is not spam. If i send a
wave invite to 50 people, and none have me on their white list, then
my wave goes into the "Maybe spam" folder for all of them. I have no
control over this and no way of getting it out of that folder, other
then contacting the people and letting them know that i am sending
them a wave for the first time please add me to your white list. If
someone does not care to add me to their white list then all my
subsequent waves that I invite them to will also go into their "Maybe
spam" folder unless they decided to black list me, at which point all
my waves to them will be in a spam folder; this would not mean every
person I sent a wave to will not receive it though, as only they will
consider me spam.

I agree -- a "Maybe Spam" folder will get treated exactly the same way the "Spam" folder gets treated, and most people will never bother to look.  We need a system that doesn't really involve changing _how_ the user interacts with their "Inbox" (whatever that box is), yet at the same time is effective at reducing/eliminating invitation spam (since message spam will likely go away in Wave).
 

> This system isn't new - it's been tried in email before. It didn't
> really work well because it wasn't implemented in the mail reader
> clients. The CAPTCHA requests where sent back as emails, which
> regularly got marked as spam themselves. With tight integration into
> the client software from day 1, this would be an excellent way to get
> round spam.

I think another reason it might not work is that many CAPTCHA systems have been cracked by spammers (probably)
http://blogs.zdnet.com/security/?p=1418

[David Fuelling]

On Mon, Jun 8, 2009 at 5:20 PM, Dmitry Unkovsky <oil.c...@gmail.com> wrote:

So, let me state that all spam topic is really about spam-invitations
and misbehaving accepted participants. Isn't it?

Exactly.  And misbehaving participants won't last long.

[dion]

On Jun 8, 1:41 pm, "d.brophy" <d...@dontstayin.com> wrote:
> If you're not on someone's whitelist, the first time you invite them
> to a wave, their WSP sends back a message containing a CAPTCHA image

No, that's a terrible idea!

There are tons of reasons for programs to send wave invites, putting
something as broken as a CAPTCHA in the loop will break in that case.

... not to mention that humans fail CAPTCHAs all the time :)

A much better solution would be to demand a little hashcash postage to
be affixed to the invite, that would make it possible to send invites
for programs, but it would be very expensive to mass-invite a huge
amount of victims as a spammer would.

A global whitelisting infrastructure should be used to allow
legitimate non-spamming mass-invites to take place without burning CO2
for the hashcash.

The whitelisting should probably be in the form of a reputation system
for the senders, so if a sender is found to be consistently hammy over
a long period of time, then they would be allowed to pay less hashcash
for invites and possibly even get to do invites for free.

[d.brophy]

No, you misunderstand. Once the sender correctly completes the
CAPTCHA, the wave automatically moves into the recipients Inbox.

Or you could do away with this Maybe Spam folder entirely.

[Dmitry Unkovsky]

Having web of waves with invites, and web of sites with emails, we see
that invitations functionality is isomorphic mapping of email system
functionality, only exclude "permanent" conversations from last.
Well, not sure, but looks like truth.

Considered that accounts, or entities, can be created in any number,
we have email spam problem projected as invites problem, with all
usual ways to fight with it.

The difference between two is level of interaction.
And the need of "anonymous" invites - can we avoid them?

We can partially do that granting "permission to invite" (subscribing)
to sources of interest.

But there is very general case -- "subscription" to wave by non-wave
action, e.g. by telling account name to friend, publishing address in
whitepaper or on public wave etc. -- we can't transfer subscription to
subscriber's wave account, or it becomes permission to very general
class of participants.

Let's think of groups. Can they help? Granting option to invite to any
group participant. Then, how can anonymous(spammer) or program became
participant of this group? The same problem with unlimited number of
anonymous accounts here.

That's just another turn on thinking. Looks like having "anonymous"
accounts is alike flaw as allowing sender spoofing in smtp.

[David Fuelling]

On Mon, Jun 8, 2009 at 9:33 PM, Dmitry Unkovsky <oil.c...@gmail.com> wrote:

Having web of waves with invites, and web of sites with emails, we see
that invitations functionality is isomorphic mapping of email system
functionality, only exclude "permanent" conversations from last.
Well, not sure, but looks like truth.

Considered that accounts, or entities, can be created in any number,
we have email spam problem projected as invites problem, with all
usual ways to fight with it.

My feeling is that you have much less "fighting" when the problem is reduced to trying to fight "invitation spam".  There are _many_ more messages sent on the web than their are invitations (I could be wrong about this prediction).

Anyway, a possible aid in this battle might be to separate my "Inbox" from my "Invitation Requests" box.  That way, you don't have your attention diverted to "invitations" as often as you would if they were coming into your regular "Inbox".

The difference between two is level of interaction.
And the need of "anonymous" invites - can we avoid them?

We could avoid them, but they serve an important purpose, even in email.
 

That's just another turn on thinking. Looks like having "anonymous"
accounts is alike flaw as allowing sender spoofing in smtp.

I don't think it's a flaw, so long as their exists a "friend-list" mechanism so that message spam cannot come through.
There _should_ be way less "invitation spam" than there will be "message spam", especially if we have a generic way to throttle invitation requests in a cross-domain fashion.

[Jorge Vargas]

ok then please read "delete" as flagged as spam. Instead of 1000000
people marking the wave spam in their inbox you will have 1000 servers
doing it. That's much better control over what is spam.

[Joakim Gândara]

Doesn't a Wave spammer open himself up to DDoS attacks? If he sends
out messages to millions of Wave users from his rogue Wave server, he
not only has to allow them to fetch the wave (or more probably waves
if he's sending individual waves to each recipienct) from his server,
the recipients can also edit the wave.

This allows every user to not only change the spam into whatever
message he likes, but also invite others to use the rogue Wave server.
The spammer can unwittingly become the host of angry flame wars,
taxing his rogue server without giving him a penny in return.

A rogue server may attempt to protect itself from repercussions by not
responding to invitations or edits, but then it would be easy to
automate blacklisting based on whether a Wave server doesn't implement
those features.

[Joel Dietz]

My guess is DDoS attacks of the kind you mention are illegal, which would be one drawback.

[Bastian Hoyer]

the spammer could easily modify his server to not accept any changes
to the wave from other servers.

[Justin Johnson]

Couldn't the rogue server respond as if it did edit the wave or accept the invitation without actually doing it to prevent being blacklisted in that manner?

On Wed, Jul 22, 2009 at 1:04 AM, Joakim Gândara <joaquim...@gmail.com> wrote: