IRC log of dig on 2013-03-29

Timestamps are in UTC.

00:48:07 [rszeno]
rszeno has quit (Quit: Leaving.)
02:23:39 [scor]
scor (~scor@c-98-216-39-127.hsd1.ma.comcast.net) has joined #dig
02:23:39 [scor]
scor has quit (Changing host)
02:23:40 [scor]
scor (~scor@drupal.org/user/52142/view) has joined #dig
02:56:45 [scor]
scor has quit (Quit: scor)
05:53:35 [bblfish]
bblfish (~bblfish@AAubervilliers-651-1-301-37.w83-114.abo.wanadoo.fr) has joined #dig
07:50:32 [melvster]
melvster (~melvin@89.176.108.70) has joined #dig
08:11:45 [melvster]
melvster has quit (Ping timeout: 255 seconds)
08:12:52 [trueg_away]
trueg_away is now known as trueg
08:18:58 [melvster]
melvster (~melvin@ip-94-112-34-93.net.upcbroadband.cz) has joined #dig
09:42:56 [jmvanel]
jmvanel has quit (Ping timeout: 260 seconds)
10:04:28 [rszeno]
rszeno (~rszeno@79.114.105.37) has joined #dig
10:15:29 [melvster]
hi bblfish: may be faster to discuss cert : key on irc rather than mail?
10:15:54 [bblfish]
could be.
10:16:17 [bblfish]
not sure. I'd rather not change cert:key relation I have a bad feeling about it.
10:16:38 [melvster]
do you really think a change in the ontology would then trigger multiple changes elsewhere? is this to do with reasoning or IFP etc.?
10:16:50 [melvster]
sure i just want to try and understand the logic
10:17:23 [melvster]
it would be nice if the cert ontology could be used in payment too, otherwise we have to write another almost identical ontology one for agents one for payments
10:17:38 [bblfish]
who says it cannot be used in payments?
10:18:01 [melvster]
payments work with accounts, rather than agents
10:18:05 [melvster]
agents have an account
10:18:11 [melvster]
it's pretty confusing
10:18:21 [bblfish]
do the bitcoin people have an ontology or are you just making it up?
10:18:24 [melvster]
but basically that's how accountancy has worked for a longtime
10:19:06 [melvster]
they use dsa keys tied to an account
10:19:16 [melvster]
so the foaf and cert ontologies can be reused
10:19:16 [bblfish]
have they got an ontology?
10:19:21 [melvster]
and also the label
10:19:37 [melvster]
im unsure who you mean by 'they'
10:19:43 [bblfish]
the bitcoing people
10:19:51 [melvster]
sure
10:19:56 [bblfish]
where?
10:19:59 [melvster]
multiple
10:20:01 [melvster]
foaf isone
10:20:05 [bblfish]
which is the official one?
10:20:06 [melvster]
dcterms
10:20:30 [bblfish]
where is the official bitcoin ontology spec that we can refer to to see what their problem is?
10:20:51 [melvster]
let's forget bitcoin for the moment
10:20:58 [melvster]
i think that's distracted the conversation
10:21:23 [bblfish]
ok. So you want to just use the bucket analogy
10:21:34 [bblfish]
you have a bank account and the bank account is a bucket
10:21:34 [melvster]
sure ... bucket would be like an account
10:21:38 [melvster]
exactly!
10:21:58 [melvster]
now associated with this bucket would be a key
10:22:03 [bblfish]
A bank account may even be like an ldp:Container
10:22:05 [melvster]
much like assoicated with a safe is a key
10:22:16 [bblfish]
you can POST new moeny to it, and you can remove money from it
10:22:43 [melvster]
im unsure LDP can model distributed buckets very well
10:22:57 [melvster]
because the problem bitcoin solves is a race condition known as 'double spend'
10:23:04 [bblfish]
ah you are back to bitcoin
10:23:12 [melvster]
yes
10:23:20 [melvster]
but you can think of it as distrubted buckets
10:23:20 [bblfish]
but you said above it confuses things
10:23:30 [melvster]
i can generalize
10:23:54 [melvster]
in the situation of a distributed bucket system
10:24:15 [bblfish]
you mean you have a bucket and I have a bucket.
10:24:15 [melvster]
you need to be able to increment and decrement buckets
10:24:23 [melvster]
ok good example
10:24:27 [melvster]
but it extends to N buckets
10:24:35 [bblfish]
timbl has a bucket
10:24:41 [melvster]
for example
10:24:53 [melvster]
now in a decentralized environment
10:24:55 [bblfish]
ok. So you can put money into my bucket but not remove it.
10:25:01 [melvster]
it's a hard problem to determine what is in every bucket
10:25:24 [bblfish]
look I don't have 20hours to discuss this.
10:25:41 [melvster]
ill try and be succinct
10:25:43 [bblfish]
The thing is you can see that you are going way into hypotheticals here.
10:26:10 [melvster]
but the summary is that accounts need to be able to have keys (independent of bitcoin, webid etc.)
10:26:20 [bblfish]
I don't think you'll have the resources to explain why you need cert:key's domain to be rdfs:Resource
10:26:38 [bblfish]
you mean public keys?
10:26:46 [webr3]
webr3 has quit (Ping timeout: 246 seconds)
10:26:47 [melvster]
pub/priv key pairs
10:27:02 [melvster]
but keys
10:27:10 [bblfish]
ok. Who operates the private key?
10:27:11 [melvster]
i didnt say public keys, i said keys
10:27:20 [bblfish]
the bucket itself?
10:27:23 [melvster]
out of scope
10:27:27 [bblfish]
no
10:27:30 [melvster]
it's just a key associated with a bucket
10:27:38 [bblfish]
the cert:key relates the owner of the private key to the public key
10:27:45 [bblfish]
so why not create a new relation
10:28:00 [melvster]
sure i could fork the whole ontology
10:28:06 [bblfish]
bucket wac:accessibleBy key
10:28:15 [bblfish]
no you don't have to fork the whole ontology
10:28:20 [bblfish]
you just need one new relation
10:28:59 [melvster]
i think we essentially have 2 concepts here:
10:29:10 [melvster]
1) a WebID public key
10:29:14 [melvster]
2) a public key
10:29:35 [bblfish]
cert:key relates an agent to a key.
10:29:45 [bblfish]
you can then create a relation to relate a bucket to a key
10:29:46 [melvster]
yes my point is that that is a bug
10:29:54 [bblfish]
why not a new relation?
10:30:06 [melvster]
a new relation in the cert ontology?
10:30:11 [melvster]
that could work
10:30:16 [bblfish]
or another ontology
10:30:22 [bblfish]
the bitbucket ontology
10:30:30 [melvster]
it's generally good practice to reuse stuff tho
10:30:59 [melvster]
otherwise there's double to maintain and double to implement
10:31:16 [bblfish]
yes, but if you can't re-use it then you create a new relation. We have one in WAC. The realtion between an ACL and agents, resource, and methods
10:31:48 [melvster]
im only interested in publickey at this point, not in WAC, agents, TLS or anything else
10:32:16 [bblfish]
you mention bitbucket. And we don't know what the best way to model their use case is
10:32:17 [webr3]
webr3 (~nathan@host213-122-100-71.range213-122.btcentralplus.com) has joined #dig
10:32:26 [melvster]
it didnt mention bitbucket, you did
10:32:49 [melvster]
oh, you mean bitcoin?
10:33:03 [melvster]
ok let me quickly describe that
10:33:23 [melvster]
a bitcoin account is a foaf : OnlineAccount with a label, and a DSA key pair
10:33:29 [bblfish]
http://lists.w3.org/Archives/Public/public-webid/2013Mar/0074.html
10:33:34 [bblfish]
I meant bitcoin sorry
10:33:53 [melvster]
ah ok
10:34:07 [bblfish]
where is that specified?
10:34:11 [melvster]
one sec
10:37:47 [melvster]
e.g. http://en.wikipedia.org/wiki/Bitcoin#Addresses there's lots of info in the bitcoin wiki about the exact details ...
10:38:12 [melvster]
btw bitcion is completely decentralized
10:38:16 [melvster]
so there's not really a 'they
10:38:23 [melvster]
it's not a corporation or w3c member etc
10:38:28 [melvster]
it's a technology like git or the web
10:38:57 [bblfish]
That is not an ontology. There is no mention of foaf:Account there.
10:38:59 [melvster]
but it's the biggest distributed computing project on the planet now
10:39:11 [rszeno]
melvster, imo you need to build a specific ontology for bitcoin
10:39:28 [bblfish]
yes, but you are assuming that they need a relation between an account and a key. They may not need that.
10:39:45 [bblfish]
they may be ok with a WebID relation to an key.
10:39:47 [melvster]
yes that's my assumption
10:39:54 [bblfish]
but it may be flawed
10:39:57 [melvster]
it may
10:40:15 [melvster]
however, it has lead me to uncover what i think is a bug in the cert ontology
10:40:28 [melvster]
two separate conversations, in a way
10:40:36 [bblfish]
it can't be a bug until it is clear that your ontology is a good one
10:40:42 [bblfish]
and you have no ontology
10:40:55 [melvster]
sure it can
10:41:03 [melvster]
bugs are not dependent on external ontologies
10:41:42 [melvster]
the question lies as to A) what should be the domain of cert : key B) what is the cost of changing
10:42:00 [bblfish]
you are saying: I think one could model X this way, if it is modelled this way then you have a bug. But you have not proven that your way of modelling it is correct. So it is only a hypothetical bug. How can we fix hypothetical bugs
10:42:38 [melvster]
you examine the semantics
10:42:39 [bblfish]
?
10:42:57 [bblfish]
You have no semantics yet. You have a hypothetical semantics
10:43:05 [melvster]
the :key relation has a Domain
10:43:09 [bblfish]
yes.
10:43:59 [bblfish]
it has foaf:Agent as a domain
10:43:59 [melvster]
that is the possibility space of who can have a Key
10:44:27 [bblfish]
no it defines the cert:key relation
10:44:27 [trueg]
trueg is now known as trueg_away
10:44:27 [bblfish]
you want to define access control to bitcoin "buckets". It may be that an access control ontology is the way to go.
10:44:37 [bblfish]
that is what WAC does.
10:45:22 [melvster]
question
10:45:25 [bblfish]
in your logic we could have argued that resources require access to people who can prove keys. So perhaps you would want to say <> putAccess key .
10:45:30 [melvster]
is a bank account an Agent?
10:45:52 [bblfish]
not if you think of it as a bucket.
10:46:04 [bblfish]
where you can put money in
10:46:10 [melvster]
yet a bank account could be protected with PKI
10:46:15 [bblfish]
yes.
10:46:16 [melvster]
it could have public and private keys
10:46:20 [bblfish]
no
10:46:28 [bblfish]
you can model it with WAC
10:46:52 [bblfish]
with WAC you can say this agent can PUT money into the bucket, and this one can GET money out of it
10:47:25 [melvster]
ok well it may not be an efficient use of time to explain how modern accountancy works ... could we quickly just talk about ( B) what would be the implications of changing Agent to Resource
10:47:56 [melvster]
1. why would it be impossible leave implementations untouched
10:47:58 [bblfish]
I don't think there is much about accountance relevant to this problem that I don't understand
10:48:20 [bblfish]
that is the question you asked in the e-mail
10:48:22 [melvster]
2. why would it be impossible to leave other specs untouched
10:48:24 [melvster]
yes
10:48:35 [bblfish]
so you see the chat on irc did not get us further.
10:48:57 [melvster]
i dont see that
10:49:03 [bblfish]
the quesiton I have is why do we need to extend the range of cert:key
10:49:07 [melvster]
but you are entitled to think so
10:49:11 [melvster]
ok
10:49:17 [melvster]
that i can explain
10:49:52 [bblfish]
well above you said you have a hypothetical case,
10:50:01 [bblfish]
and that is not a good or solid argument
10:50:11 [melvster]
i would like to start 'webizing' mature distributed computing problems in the payments space, but not limited to that, it extends to all robots that will be able to use digital or legacy cash
10:50:25 [bblfish]
because I can show that we can do access control to resources in a fine grained manner without your extension
10:50:31 [melvster]
for this, i would like to model PKI pub/pri key pairs
10:50:46 [bblfish]
ok you can use the cert ontology for that
10:51:46 [melvster]
id like to reuse cert ontology
10:52:16 [bblfish]
( anyway it's good to know that bitcoin uses DSA - in fact they use ECDSA - which is eliptic curve DSA - which I am not sure is related directly to DSA btw )
10:52:27 [bblfish]
who says you cannot?
10:52:29 [melvster]
so for example an agent could have multiple accounts, each with a key/pair, all capable of payments
10:53:07 [bblfish]
why do you think that the bitcoin should be modelled as an foaf:Account?
10:53:12 [cheater__]
cheater__ (~cheater@p57AEB110.dip.t-dialin.net) has joined #dig
11:14:24 [DIGlogger]
DIGlogger (~dig-logge@groups.csail.mit.edu) has joined #dig
11:14:24 [adams.freenode.net]
topic is: Decentralized Information Group @ MIT http://dig.csail.mit.edu/
11:14:24 [adams.freenode.net]
Users on #dig: DIGlogger RalphS cheater__ webr3 rszeno melvster bblfish bergi trueg_away betehess Yudai___ sandro_ ericP_ presbrey tyteen4a03 manu1 manu-db mattl
11:15:03 [melvster]
a file can have a public key
11:15:08 [melvster]
but the agent can use that key
11:16:27 [melvster]
i think you're suggesting having something like cert : hasKey to have a wider domain, but to my mind it seems more logical just to let cert : key do that job, you've not explained why WAC needs changing
11:16:36 [bblfish]
DigLogger: the logs you missed are here: http://pastebin.com/DwfXaRMh
11:16:36 [bblfish]
I'm logging. I don't understand 'the logs you missed are here: http://pastebin.com/DwfXaRMh', bblfish. Try /msg DIGlogger help
11:17:04 [bblfish]
melvester why do you put spaces between the prefix and the name?
11:17:17 [bblfish]
cert : hasKey instead of cert:hasKey ?
11:17:19 [melvster]
because in some client is turns into a smiley
11:17:23 [melvster]
on IRC
11:17:30 [melvster]
:P
11:17:36 [melvster]
is : P
11:17:47 [melvster]
so :PublicKey is a smiley for me
11:17:53 [melvster]
is less readable
11:18:01 [bblfish]
ah, you should turn off smileys
11:18:14 [melvster]
sure but i cant turn off smiley's for everyone
11:18:50 [bblfish]
melvster if you have cert:key have a domain of anything, then we start having to explain what the relation of that anything is to the private key of the public key
11:19:06 [bblfish]
is there an owner of that anything that has control of the private key?
11:19:30 [melvster]
in webid sure there is, nothing changes there
11:19:34 [bblfish]
so now we need two relations. The relation of ownership to an anything, that may have a cert key.
11:19:59 [melvster]
a webid is already an agent
11:20:07 [trueg_away]
trueg_away is now known as trueg
11:20:29 [melvster]
im NOT suggesting change what a webid is
11:20:35 [melvster]
that stays exactly the same
11:20:43 [bblfish]
ok, so now we need to always check that a thing is a foaf:Agent too.
11:20:55 [bblfish]
so we need to do twice the work
11:21:36 [bblfish]
because your cert:key relation is something has a relation to a key. What is that relation? Where is the agent that has the private key?
11:22:21 [melvster]
wait, remember what a key is, its just a number
11:22:22 [bblfish]
you need to get the triangle of the private key in there.
11:22:28 [bblfish]
exactly
11:22:40 [bblfish]
so cannot everything have a cert:key relation to my key?
11:23:27 [melvster]
i think this is what Mo meant when he said rdf : Resource ... btw we could ask him on #swig he is nevali there
11:24:22 [bblfish]
rdf:Resource means anything.
11:24:26 [melvster]
ok
11:24:36 [melvster]
so it means we dont restrict it
11:25:43 [bblfish]
yes
11:26:16 [bblfish]
which means that in those cases where it is an agent, the relation between the public key and the private key gets lost
11:27:00 [melvster]
do you mean "the relation between the public key and the private key"
11:27:12 [melvster]
or "the relation between the resource and the public key"
11:28:06 [melvster]
so instead of having
11:28:09 [melvster]
agent -> key
11:28:11 [melvster]
we have
11:28:16 [melvster]
agent -> key
11:28:17 [melvster]
AND
11:28:21 [melvster]
resource -> key
11:28:22 [melvster]
and
11:28:29 [melvster]
agent -> resource -> key
11:28:40 [melvster]
in particular, it's useful when that resource is an account
11:28:56 [bblfish]
it seems useful, but it is bad modelling.
11:29:11 [bblfish]
it won't work.
11:29:34 [melvster]
why?
11:29:59 [melvster]
webid will still work
11:30:02 [melvster]
wac will still work
11:30:09 [melvster]
implementations will still work
11:30:17 [melvster]
and you get more for free
11:30:58 [melvster]
a key is just a number
11:31:09 [melvster]
it seems odd to say that only an agent can have a number
11:31:26 [melvster]
and you rule out whole classes of things such as bank accounts
11:31:45 [bblfish]
you don't rule anything out at all
11:31:55 [bblfish]
you just model it differently
11:32:23 [melvster]
i mean in terms of using the natural turn which is in the cert ontology of ": key"
11:32:52 [melvster]
one way to model it would be
11:33:01 [melvster]
like we used to have cert : identity
11:33:07 [melvster]
we could have cert : account
11:33:08 [melvster]
but
11:33:26 [melvster]
after long discussions people felt it was more natural to go the other direction
11:37:23 [melvster]
reusing : key i can start building webid bots with payment ability today
11:37:46 [trueg]
trueg is now known as trueg_away
11:37:47 [melvster]
maybe i should just do that and we can change the ontology later
11:37:54 [melvster]
if at all
11:38:21 [melvster]
showing a useful implementation always is a good argument towards consensus
11:38:29 [melvster]
then it's less hypothetical
11:43:12 [trueg_away]
trueg_away is now known as trueg
11:47:33 [bblfish]
not just an implementation, an ontology.
11:47:41 [bblfish]
and you need to have it reviewed.
11:49:39 [melvster]
bblfish: what's the way forward, should we continue on IRC / on email / or should I raise an issue? I still claim you have not made any reasonable objections that I can understand ...
11:49:57 [melvster]
i claim it's abug
11:49:57 [bblfish]
I am trying to respond to your mail.
11:50:01 [melvster]
ok sure
11:50:09 [melvster]
and one that's easy to fix
11:50:23 [bblfish]
no it is not easy to fix
11:50:26 [bblfish]
there is no bug.
11:50:36 [bblfish]
your improvement in fact introduces bugs
11:50:45 [bblfish]
so I am trying to argue that.
11:52:42 [melvster]
ok well i guess the WG will decide in time ... im just trying to understand better
11:53:58 [melvster]
so far two are in favour rdfs : Resource one is in favour of foaf : Agent
11:54:36 [melvster]
that expressed opinions on the thread
11:55:20 [melvster]
one way to model it
11:55:31 [melvster]
i could say bitcoin:1d8ds8hdioahihd is an Agent
11:55:36 [melvster]
what's an agent?
11:55:42 [melvster]
"things that do stuff"
11:56:01 [melvster]
then the agent has the key
11:56:07 [melvster]
tho is NOT a webid
11:56:11 [melvster]
officially
11:56:34 [melvster]
can an account be an agent i wonder
12:00:04 [scor]
scor (scor@nat/acquia/x-vegtuyulblmhxewj) has joined #dig
12:00:04 [scor]
scor has quit (Changing host)
12:00:04 [scor]
scor (scor@drupal.org/user/52142/view) has joined #dig
12:07:23 [trueg]
trueg is now known as trueg_away
12:52:50 [melvster]
bblfish: im not sure the Range of cert : key should not be cert : PublicKey either -- see the description, "key - relates an agent to a key - most often the public key. "
13:12:29 [bblfish]
yes, I would be ok with that change.
13:12:35 [trueg_away]
trueg_away is now known as trueg
13:12:56 [bblfish]
it would require a relation privateKey from public to private key
13:14:32 [bblfish]
here is the answer I was thinking of above http://lists.w3.org/Archives/Public/public-webid/2013Mar/0083.html
13:15:03 [bblfish]
and here is the second answer http://lists.w3.org/Archives/Public/public-webid/2013Mar/0085.html
13:53:03 [melvster]
melvster has quit (Ping timeout: 260 seconds)
13:55:23 [melvster]
melvster (~melvin@ip-94-112-34-93.net.upcbroadband.cz) has joined #dig
14:15:46 [jmvanel]
jmvanel (~jmvanel@139.239.24.109.rev.sfr.net) has joined #dig
14:25:46 [timbl]
timbl (~timbl@30-5-66.wireless.csail.mit.edu) has joined #dig
14:37:07 [scor]
scor has quit (Ping timeout: 264 seconds)
14:37:43 [scor]
scor (scor@nat/acquia/x-fykeaanlhxsxzquc) has joined #dig
14:37:44 [scor]
scor has quit (Changing host)
14:37:44 [scor]
scor (scor@drupal.org/user/52142/view) has joined #dig
15:00:35 [timbl]
timbl has quit (Quit: timbl)
15:04:39 [timbl]
timbl (~timbl@30-5-66.wireless.csail.mit.edu) has joined #dig
15:07:54 [melvster]
bblfish: you could just add both keys to the range if you want an easy solution (pub/pri)
15:08:38 [melvster]
just reading your reply
15:35:41 [melvster]
bblfish: i guess we have quite different views on the complexity of this, we can just let the CG or WG reach a consensus and hopefully make a quick change, as necessary
15:36:05 [melvster]
if it stays agent, ill make a new predicate
15:36:16 [melvster]
if it is resource, ill reuse
15:37:19 [melvster]
it might make more sense just to start fresh, tho i thought it might be nice to see webid based robots making payments in the wild
15:38:45 [melvster]
there's also: http://payswarm.com/specs/source/vocabs/security
15:38:52 [melvster]
which is designed for payments
15:49:50 [trueg]
trueg is now known as trueg_away
15:54:52 [trueg_away]
trueg_away is now known as trueg
16:04:44 [timbl]
timbl has quit (Quit: timbl)
16:18:06 [timbl]
timbl (~timbl@30-5-66.wireless.csail.mit.edu) has joined #dig
16:27:30 [scor]
scor has quit (Quit: scor)
17:14:01 [scor]
scor (scor@drupal.org/user/52142/view) has joined #dig
17:27:43 [jmvanel]
jmvanel has quit (Ping timeout: 260 seconds)
17:36:39 [melvster]
melvster has quit (Quit: Leaving.)
17:40:41 [cheater_1]
cheater_1 (~cheater@p57AEAF15.dip.t-dialin.net) has joined #dig
17:42:55 [trueg]
trueg is now known as trueg_away
17:44:04 [cheater__]
cheater__ has quit (Ping timeout: 260 seconds)
19:13:58 [jmvanel]
jmvanel (~jmvanel@188.197.139.88.rev.sfr.net) has joined #dig
20:05:19 [ericP_]
ericP_ is now known as ericP
20:05:49 [ericP]
ericP is now known as Guest58435
20:18:37 [Guest58435]
Guest58435 is now known as ericP
20:19:20 [RalphS]
RalphS has quit ()
21:14:46 [scor]
scor has quit (Quit: scor)
21:28:20 [trueg_away]
trueg_away is now known as trueg
21:37:48 [timbl]
timbl has quit (Quit: timbl)
21:46:54 [timbl]
timbl (~timbl@30-5-66.wireless.csail.mit.edu) has joined #dig
21:50:29 [timbl]
timbl has quit (Client Quit)
22:10:19 [trueg]
trueg is now known as trueg_away
22:38:04 [jmvanel]
jmvanel has quit (Ping timeout: 260 seconds)
22:45:03 [timbl]
timbl (~timbl@c-24-62-225-11.hsd1.ma.comcast.net) has joined #dig
23:13:48 [betehess]
betehess has quit (Quit: Leaving)
23:31:16 [melvster]
melvster (~melvin@89.176.108.70) has joined #dig