[opencog-dev] Re: AtomSpace on IPFS !

2019-10-29 Thread Linas Vepstas
For your entertainment: the AtomSpace on top of IPFS is now "code complete"!

The code is here:

https://github.com/opencog/atomspace-ipfs
> 
>

It works.  (Six of seven unit tests pass; the seventh unit test is a
multi-user unit test and it fails because the current design requires
coordination between users.)

It's slow. Approximately very slow, as compared to the SQL backend.  But it
works and its feature complete.

There are design issues. They're described in the README file. I don't see
any particular way of "fixing" them, without asking for and getting new
features in IPFS.  So if you are bored of opencog, and want to hack on IPFS
internals instead, come talk to me.

-- Linas

-- 
cassette tapes - analog TV - film cameras - you

-- 
You received this message because you are subscribed to the Google Groups 
"opencog" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to opencog+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/opencog/CAHrUA3530HrabnAnHqkcjN9UeuO_4uOmXUjOEO4bS1ja6%2B_XJA%40mail.gmail.com.


Re: [opencog-dev] Re: AtomSpace on IPFS !

2019-10-21 Thread Linas Vepstas
On Mon, Oct 21, 2019 at 5:46 AM Roland Coeurjoly 
wrote:

> When I click in the link you posted:
>
>
> https://explore.ipld.io/#/explore/QmT9tZttJ4gVZQwVFHWTmJYqYGAAiKEcvW9k98T5syYeYU
> 
>
> On another browser I don't get that error but it stays forever loading,
> without nothing displaying.
>

It won't be "forever", but it might be a long time. The speed at which IPFS
distributes requests seems to be problematic.  I don't understand why.
There is a DHT, and clearly, various peers have to talk to one-another to
figure out where some particular entry is. From experience, this seems to
take minutes.

That is -- that particular entry is on my server, and accessing it locally,
its instant. But for explore.ipld.io -- it needs to query the DHT to find
my server (or some other server that has a cached copy) and again, that
slow, I don't know why.

-- Linas
-- 
cassette tapes - analog TV - film cameras - you

-- 
You received this message because you are subscribed to the Google Groups 
"opencog" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to opencog+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/opencog/CAHrUA36f796OpP9L9KH7uodsgRuX9OCP%2B9En_c9U9etDRDQESQ%40mail.gmail.com.


Re: [opencog-dev] Re: AtomSpace on IPFS !

2019-10-21 Thread Ivan V.
I'm just trying to be useful: the link

works fine on Ubuntu 18.04.3 LTS / Firefox and Ubuntu 18.04.3 LTS /
Chrommium. It shows a hierarchical atom structure I can navigate through.

pon, 21. lis 2019. u 12:46 Roland Coeurjoly 
napisao je:

> When I click in the link you posted:
>
>
> https://explore.ipld.io/#/explore/QmT9tZttJ4gVZQwVFHWTmJYqYGAAiKEcvW9k98T5syYeYU
> 
>
> On another browser I don't get that error but it stays forever loading,
> without nothing displaying.
>
> On Monday, October 21, 2019 at 12:15:55 PM UTC+2, linas wrote:
>
>>
>> On Mon, Oct 21, 2019 at 3:55 AM Roland Coeurjoly 
>> wrote:
>>
>>> I get this error:
>>>
>>> AbortController is not defined
>>>
>>
>> From where? Its not an error message that my code generates.
>>
>>>
>>> On Sunday, October 20, 2019 at 6:35:06 AM UTC+2, linas wrote:

 Oh, I found one:  This should work:


 https://explore.ipld.io/#/explore/QmT9tZttJ4gVZQwVFHWTmJYqYGAAiKEcvW9k98T5syYeYU

 -- linas

 On Sat, Oct 19, 2019 at 11:19 PM Linas Vepstas 
 wrote:

> Pursuant to a recent question from Jim Rutt, I thought it might be
> interesting to see if its possible to layer the atomspace on top of IPFS.
> Rather than speculate, I thought I'd try to do it.   It's here:
>
> https://github.com/linas/atomspace-ipfs
>
> So far you can only store the AtomSpace to IPFS; you cannot yet load
> from IPFS.  I've published a demo, which you can find at
> /ipfs/QmT9tZttJ4gVZQwVFHWTmJYqYGAAiKEcvW9k98T5syYeYU
>
> Unfortunately, you need to find a data explorer to view it, and I
> haven't found any good public websites for this, yet.  You'd need the
> built-in http://localhost:5001/ browser to see it clearly.
>
> Issues and challenges:
> * Supporting Values will be difficult and confusing. I have no good
> ideas for this yet.
> * Likewise, I don't yet know how to search the incoming set.
> * I'm concerned that it's slow. I think we'll be lucky to pump a few
> hundred atoms/second through it. But I  don't really know.
> * I'm using IPNS for mutability, and IPNS has ... well-known issues.
>
> But .. what works does work. So that's encouraging. I invite everyone
> to work on it.  I don't particularly want to be a long-term maintainers on
> this.
>
> -- Linas
>
> --
> cassette tapes - analog TV - film cameras - you
>


 --
 cassette tapes - analog TV - film cameras - you

>>> --
>>> You received this message because you are subscribed to the Google
>>> Groups "opencog" group.
>>> To unsubscribe from this group and stop receiving emails from it, send
>>> an email to ope...@googlegroups.com.
>>> To view this discussion on the web visit
>>> https://groups.google.com/d/msgid/opencog/159b1662-6f29-4583-be45-26e8b62dc0cd%40googlegroups.com
>>> 
>>> .
>>>
>>
>>
>> --
>> cassette tapes - analog TV - film cameras - you
>>
> --
> You received this message because you are subscribed to the Google Groups
> "opencog" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to opencog+unsubscr...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/opencog/b5869166-3712-4568-bac1-a54841e91191%40googlegroups.com
> 
> .
>

-- 
You received this message because you are subscribed to the Google Groups 
"opencog" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to opencog+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/opencog/CAB5%3Dj6VFv4k0dizDW8hu1gnmgsTdZW6QMtgunuWfUkNMOFxaJw%40mail.gmail.com.


Re: [opencog-dev] Re: AtomSpace on IPFS !

2019-10-21 Thread Roland Coeurjoly
When I click in the link you posted: 

https://explore.ipld.io/#/explore/QmT9tZttJ4gVZQwVFHWTmJYqYGAAiKEcvW9k98T5syYeYU
 


On another browser I don't get that error but it stays forever loading, 
without nothing displaying.

On Monday, October 21, 2019 at 12:15:55 PM UTC+2, linas wrote:

>
> On Mon, Oct 21, 2019 at 3:55 AM Roland Coeurjoly  > wrote:
>
>> I get this error:
>>
>> AbortController is not defined
>>
>
> From where? Its not an error message that my code generates. 
>
>>
>> On Sunday, October 20, 2019 at 6:35:06 AM UTC+2, linas wrote:
>>>
>>> Oh, I found one:  This should work:
>>>
>>>
>>> https://explore.ipld.io/#/explore/QmT9tZttJ4gVZQwVFHWTmJYqYGAAiKEcvW9k98T5syYeYU
>>>
>>> -- linas
>>>
>>> On Sat, Oct 19, 2019 at 11:19 PM Linas Vepstas  
>>> wrote:
>>>
 Pursuant to a recent question from Jim Rutt, I thought it might be 
 interesting to see if its possible to layer the atomspace on top of IPFS.  
 Rather than speculate, I thought I'd try to do it.   It's here:

 https://github.com/linas/atomspace-ipfs

 So far you can only store the AtomSpace to IPFS; you cannot yet load 
 from IPFS.  I've published a demo, which you can find at 
 /ipfs/QmT9tZttJ4gVZQwVFHWTmJYqYGAAiKEcvW9k98T5syYeYU

 Unfortunately, you need to find a data explorer to view it, and I 
 haven't found any good public websites for this, yet.  You'd need the 
 built-in http://localhost:5001/ browser to see it clearly.

 Issues and challenges:
 * Supporting Values will be difficult and confusing. I have no good 
 ideas for this yet.
 * Likewise, I don't yet know how to search the incoming set.
 * I'm concerned that it's slow. I think we'll be lucky to pump a few 
 hundred atoms/second through it. But I  don't really know.
 * I'm using IPNS for mutability, and IPNS has ... well-known issues.

 But .. what works does work. So that's encouraging. I invite everyone 
 to work on it.  I don't particularly want to be a long-term maintainers on 
 this.

 -- Linas

 -- 
 cassette tapes - analog TV - film cameras - you

>>>
>>>
>>> -- 
>>> cassette tapes - analog TV - film cameras - you
>>>
>> -- 
>> You received this message because you are subscribed to the Google Groups 
>> "opencog" group.
>> To unsubscribe from this group and stop receiving emails from it, send an 
>> email to ope...@googlegroups.com .
>> To view this discussion on the web visit 
>> https://groups.google.com/d/msgid/opencog/159b1662-6f29-4583-be45-26e8b62dc0cd%40googlegroups.com
>>  
>> 
>> .
>>
>
>
> -- 
> cassette tapes - analog TV - film cameras - you
>

-- 
You received this message because you are subscribed to the Google Groups 
"opencog" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to opencog+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/opencog/b5869166-3712-4568-bac1-a54841e91191%40googlegroups.com.


Re: [opencog-dev] Re: AtomSpace on IPFS !

2019-10-21 Thread Linas Vepstas
On Mon, Oct 21, 2019 at 3:55 AM Roland Coeurjoly 
wrote:

> I get this error:
>
> AbortController is not defined
>

>From where? Its not an error message that my code generates.

>
> On Sunday, October 20, 2019 at 6:35:06 AM UTC+2, linas wrote:
>>
>> Oh, I found one:  This should work:
>>
>>
>> https://explore.ipld.io/#/explore/QmT9tZttJ4gVZQwVFHWTmJYqYGAAiKEcvW9k98T5syYeYU
>>
>> -- linas
>>
>> On Sat, Oct 19, 2019 at 11:19 PM Linas Vepstas 
>> wrote:
>>
>>> Pursuant to a recent question from Jim Rutt, I thought it might be
>>> interesting to see if its possible to layer the atomspace on top of IPFS.
>>> Rather than speculate, I thought I'd try to do it.   It's here:
>>>
>>> https://github.com/linas/atomspace-ipfs
>>>
>>> So far you can only store the AtomSpace to IPFS; you cannot yet load
>>> from IPFS.  I've published a demo, which you can find at
>>> /ipfs/QmT9tZttJ4gVZQwVFHWTmJYqYGAAiKEcvW9k98T5syYeYU
>>>
>>> Unfortunately, you need to find a data explorer to view it, and I
>>> haven't found any good public websites for this, yet.  You'd need the
>>> built-in http://localhost:5001/ browser to see it clearly.
>>>
>>> Issues and challenges:
>>> * Supporting Values will be difficult and confusing. I have no good
>>> ideas for this yet.
>>> * Likewise, I don't yet know how to search the incoming set.
>>> * I'm concerned that it's slow. I think we'll be lucky to pump a few
>>> hundred atoms/second through it. But I  don't really know.
>>> * I'm using IPNS for mutability, and IPNS has ... well-known issues.
>>>
>>> But .. what works does work. So that's encouraging. I invite everyone to
>>> work on it.  I don't particularly want to be a long-term maintainers on
>>> this.
>>>
>>> -- Linas
>>>
>>> --
>>> cassette tapes - analog TV - film cameras - you
>>>
>>
>>
>> --
>> cassette tapes - analog TV - film cameras - you
>>
> --
> You received this message because you are subscribed to the Google Groups
> "opencog" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to opencog+unsubscr...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/opencog/159b1662-6f29-4583-be45-26e8b62dc0cd%40googlegroups.com
> 
> .
>


-- 
cassette tapes - analog TV - film cameras - you

-- 
You received this message because you are subscribed to the Google Groups 
"opencog" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to opencog+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/opencog/CAHrUA34-f9pvLNMC3dLdwJb19DY6U2RfbYY93HH0GGqYK5NNhA%40mail.gmail.com.


[opencog-dev] Re: AtomSpace on IPFS !

2019-10-21 Thread Roland Coeurjoly
I get this error:

AbortController is not defined

On Sunday, October 20, 2019 at 6:35:06 AM UTC+2, linas wrote:
>
> Oh, I found one:  This should work:
>
>
> https://explore.ipld.io/#/explore/QmT9tZttJ4gVZQwVFHWTmJYqYGAAiKEcvW9k98T5syYeYU
>
> -- linas
>
> On Sat, Oct 19, 2019 at 11:19 PM Linas Vepstas  > wrote:
>
>> Pursuant to a recent question from Jim Rutt, I thought it might be 
>> interesting to see if its possible to layer the atomspace on top of IPFS.  
>> Rather than speculate, I thought I'd try to do it.   It's here:
>>
>> https://github.com/linas/atomspace-ipfs
>>
>> So far you can only store the AtomSpace to IPFS; you cannot yet load from 
>> IPFS.  I've published a demo, which you can find at 
>> /ipfs/QmT9tZttJ4gVZQwVFHWTmJYqYGAAiKEcvW9k98T5syYeYU
>>
>> Unfortunately, you need to find a data explorer to view it, and I haven't 
>> found any good public websites for this, yet.  You'd need the built-in 
>> http://localhost:5001/ browser to see it clearly.
>>
>> Issues and challenges:
>> * Supporting Values will be difficult and confusing. I have no good ideas 
>> for this yet.
>> * Likewise, I don't yet know how to search the incoming set.
>> * I'm concerned that it's slow. I think we'll be lucky to pump a few 
>> hundred atoms/second through it. But I  don't really know.
>> * I'm using IPNS for mutability, and IPNS has ... well-known issues.
>>
>> But .. what works does work. So that's encouraging. I invite everyone to 
>> work on it.  I don't particularly want to be a long-term maintainers on 
>> this.
>>
>> -- Linas
>>
>> -- 
>> cassette tapes - analog TV - film cameras - you
>>
>
>
> -- 
> cassette tapes - analog TV - film cameras - you
>

-- 
You received this message because you are subscribed to the Google Groups 
"opencog" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to opencog+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/opencog/159b1662-6f29-4583-be45-26e8b62dc0cd%40googlegroups.com.


[opencog-dev] Re: AtomSpace on IPFS !

2019-10-19 Thread Linas Vepstas
Oh, I found one:  This should work:

https://explore.ipld.io/#/explore/QmT9tZttJ4gVZQwVFHWTmJYqYGAAiKEcvW9k98T5syYeYU

-- linas

On Sat, Oct 19, 2019 at 11:19 PM Linas Vepstas 
wrote:

> Pursuant to a recent question from Jim Rutt, I thought it might be
> interesting to see if its possible to layer the atomspace on top of IPFS.
> Rather than speculate, I thought I'd try to do it.   It's here:
>
> https://github.com/linas/atomspace-ipfs
>
> So far you can only store the AtomSpace to IPFS; you cannot yet load from
> IPFS.  I've published a demo, which you can find at
> /ipfs/QmT9tZttJ4gVZQwVFHWTmJYqYGAAiKEcvW9k98T5syYeYU
>
> Unfortunately, you need to find a data explorer to view it, and I haven't
> found any good public websites for this, yet.  You'd need the built-in
> http://localhost:5001/ browser to see it clearly.
>
> Issues and challenges:
> * Supporting Values will be difficult and confusing. I have no good ideas
> for this yet.
> * Likewise, I don't yet know how to search the incoming set.
> * I'm concerned that it's slow. I think we'll be lucky to pump a few
> hundred atoms/second through it. But I  don't really know.
> * I'm using IPNS for mutability, and IPNS has ... well-known issues.
>
> But .. what works does work. So that's encouraging. I invite everyone to
> work on it.  I don't particularly want to be a long-term maintainers on
> this.
>
> -- Linas
>
> --
> cassette tapes - analog TV - film cameras - you
>


-- 
cassette tapes - analog TV - film cameras - you

-- 
You received this message because you are subscribed to the Google Groups 
"opencog" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to opencog+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/opencog/CAHrUA36kE8a7dJmUuX80BTJowz77mWXLD1RGKa_QpLLSBv5KoA%40mail.gmail.com.