Discussion:
[GNU/consensus] [Social] More internal use of ActivityStreams?
hellekin (GNU Consensus)
2013-01-02 20:32:17 UTC
Permalink
Coding for N different protocols on N different servers is not a better
use of time than just building support for the leading protocol.
*** I agree that OStatus has a hand in the game, and it's mentioned in
the GNU/consensus manifesto as the main protocol to follow. But I
disagree that the leadership position is enough to ignore other
alternatives.

Tent.io, to mention only one, seems promising. OStatus itself is a set
of protocols to solve a number of issues, and could as well integrate
new ones. Consider SMTP, IRC, and XMPP: they all do some sort of
federation. There is no single truth, and not considering diversity
won't lead anywhere, IMO.

==
hk
Melvin Carvalho
2013-01-02 20:33:26 UTC
Permalink
Post by hellekin (GNU Consensus)
Coding for N different protocols on N different servers is not a better
use of time than just building support for the leading protocol.
*** I agree that OStatus has a hand in the game, and it's mentioned in
the GNU/consensus manifesto as the main protocol to follow. But I
disagree that the leadership position is enough to ignore other
alternatives.
Tent.io, to mention only one, seems promising. OStatus itself is a set
of protocols to solve a number of issues, and could as well integrate
new ones. Consider SMTP, IRC, and XMPP: they all do some sort of
federation. There is no single truth, and not considering diversity
won't lead anywhere, IMO.
Yes, exactly!
Post by hellekin (GNU Consensus)
==
hk
Mikael Nordfeldth
2013-01-03 03:21:51 UTC
Permalink
Post by hellekin (GNU Consensus)
Coding for N different protocols on N different servers is not a better
use of time than just building support for the leading protocol.
*** I agree that OStatus has a hand in the game, and it's mentioned in
the GNU/consensus manifesto as the main protocol to follow. But I
disagree that the leadership position is enough to ignore other
alternatives.
I'll just break this out and continue it here on the Consensus list.
Original thread archived at:
http://lists.gnu.org/archive/html/social/2013-01/msg00000.html
Post by hellekin (GNU Consensus)
Tent.io, to mention only one, seems promising. OStatus itself is a set
of protocols to solve a number of issues, and could as well integrate
new ones.
I'm somewhat curious about Tent.io. Given that it's an API and not a
protocol as well as tightly tied to HTTP. How does it intend to
interoperate with XMPP etc? I haven't read anything about such ideas at
all when it comes to Tent, but then again I stopped caring about Tent.io
when I realised they were working _against_ established standards and
methods (i.e. mapping their own kind of activitystreams or whatever).

For example smacking some kind of arbitrary HTTP-like layer on top of
XMPP is likely to be bulky and brutal (I care about all the efforts on
social networking with XMPP as base). In these cases a lot of work would
have to go into standardising the behaviour and thus reinventing a wheel
or two.
--
Mikael Nordfeldth
http://blog.mmn-o.se/
***@hethane.se
+46705657637
Melvin Carvalho
2013-01-03 11:44:17 UTC
Permalink
Post by Mikael Nordfeldth
Post by hellekin (GNU Consensus)
Coding for N different protocols on N different servers is not a better
use of time than just building support for the leading protocol.
*** I agree that OStatus has a hand in the game, and it's mentioned in
the GNU/consensus manifesto as the main protocol to follow. But I
disagree that the leadership position is enough to ignore other
alternatives.
I'll just break this out and continue it here on the Consensus list.
Original thread archived at: http://lists.gnu.org/archive/**
html/social/2013-01/msg00000.**html<http://lists.gnu.org/archive/html/social/2013-01/msg00000.html>
Tent.io, to mention only one, seems promising. OStatus itself is a set
Post by hellekin (GNU Consensus)
of protocols to solve a number of issues, and could as well integrate
new ones.
I'm somewhat curious about Tent.io. Given that it's an API and not a
protocol as well as tightly tied to HTTP. How does it intend to
interoperate with XMPP etc? I haven't read anything about such ideas at all
when it comes to Tent, but then again I stopped caring about Tent.io when I
realised they were working _against_ established standards and methods
(i.e. mapping their own kind of activitystreams or whatever).
I appreciate your world view, and I understand that you hold your
convictions strongly.

tent is a well regarded and popular initiative, and it needs no defence.
However, I will comment that http has a reasonable track record of building
same origin social networking sites, such as friendster, myspace, facebook
and google plus. To take that approach to cross origin federation is at
least reasonable, and at best logical. There have been instances of social
web sites ineroperating with XMPP too.

In this particular forum I think it would be more appropriate not to
belittle projects that you feel, do not think the same as you, but rather,
try and operate from a perspective of tolerance, and see what you can learn
from other projects. We have had major problems in the past with people
taking the "my way or the high way" approach. Indeed we have lost valuable
partners, such as friendica from this. I think if you can take time to
look at things from the perspectives of other projects, you may gain a gain
a great deal. Working together we can do more.
Post by Mikael Nordfeldth
For example smacking some kind of arbitrary HTTP-like layer on top of XMPP
is likely to be bulky and brutal (I care about all the efforts on social
networking with XMPP as base). In these cases a lot of work would have to
go into standardising the behaviour and thus reinventing a wheel or two.
--
Mikael Nordfeldth
http://blog.mmn-o.se/
+46705657637
hellekin (GNU Consensus)
2013-01-03 16:09:00 UTC
Permalink
I stopped caring about Tent.io when I realised they were working
_against_ established standards and methods
*** Mikael, I hope this misunderstanding can be cleared when the Tent
people join the conversation. As far as I can tell, they thought the
lack of privacy in OStatus' scope was enough of a reason to take another
approach. I can't blame them for that choice. There are other
occurrences of "forks" that were beneficial to the community: look at
how Merb created a better Rails, by focusing on different features, that
ended up being integrated into the core of Ruby on Rails in the next
major version.
record of building same origin social networking sites
*** Indeed Melvin, but that is not comparable: building a single Web
site is a simple thing to do. Making it work seamlessly with others is
another story.

If we look at OAuth, the specification sounds great, but the
implementation disastrous: why would I give a third-party write access
to my profile if it only needs to access a feature (my status updates)
to read it! That entirely nullifies the intent of the specification.
That should be addressed as a blocker bug, because privacy is more
important than convenience.

Next iterations of OAuth will certainly address that mistake, but it
prompts the question of the importance of ethics in programming:
unintended consequences can be terrible in that context.
There have been instances of social web sites ineroperating with XMPP too.
*** I started a list of potential partners on the wiki: it indeed
contains XMPP-based projects. It needs some love: some categories,
templates, etc. Please reply to that point by mentioning [wiki] in the
Subject.
try and operate from a perspective of tolerance, and see what
you can learn from other projects.
*** Although I agree with your idea, I would recommend against using the
word "tolerance". Historically, a "tolerance" was a permission granted
by the religious authority to allow infidels to access Church-controlled
areas, for e.g., trading. Therefore, unlike the common understanding of
that word today, it proceeds from a dominant perspective that is far
from our purpose of integrity and transclusion, and from your idea.
Indeed we have lost valuable partners, such as friendica from this.
*** Who is "we"? I listed Friendica as a stakeholder, and a potential
partner on the wiki. In the next weeks, I will contact all projects
individually if they don't come by themselves. A consensus needs all
parties to emerge.
if you can take time to look at things from the perspectives of other
projects, you may gain a gain a great deal. Working together we can do
more.
*** +1

Time is a rare resource, making tools that can save time and enable
understanding of complex issues, to raise the bar above "us" and "them",
is a critical part of reaching a consensus.

==
hk
Mikael Nordfeldth
2013-01-03 16:50:37 UTC
Permalink
Post by hellekin (GNU Consensus)
*** Mikael, I hope this misunderstanding can be cleared when the Tent
people join the conversation. As far as I can tell, they thought the
lack of privacy in OStatus' scope was enough of a reason to take another
approach. I can't blame them for that choice.
I was thinking more of them mapping out a unique list of activity definitions rather than extending existing definitions, which will inevitably cause confusion when writing software to bridge networks of Tent/everything else.

However I believe they are still very early in their work and have a lot of possibility to change some of the "not invented here" philosophy.
Melvin Carvalho
2013-01-03 18:03:56 UTC
Permalink
Post by Mikael Nordfeldth
Post by hellekin (GNU Consensus)
*** Mikael, I hope this misunderstanding can be cleared when the Tent
people join the conversation. As far as I can tell, they thought the
lack of privacy in OStatus' scope was enough of a reason to take another
approach. I can't blame them for that choice.
I was thinking more of them mapping out a unique list of activity
definitions rather than extending existing definitions, which will
inevitably cause confusion when writing software to bridge networks of
Tent/everything else.
However I believe they are still very early in their work and have a lot
of possibility to change some of the "not invented here" philosophy.
Mikael, I think it would be helpful is you did not continue with such an
accusatory tone. Tent is not an NIH system, they have a well designed JSON
framework. There have been data standards for this kind of thing for at
least 12 years and everything has constantly evolved. Both tent and your
personal preferences are slightly unorthodox, wrt to the main stream, or
the standards world, but no one should be criticized for trying to
innovate. As it happens tent have done a great job, imho.
Mikael Nordfeldth
2013-01-03 20:04:51 UTC
Permalink
Post by Melvin Carvalho
Post by Mikael Nordfeldth
However I believe they are still very early in their work and have a
lot of possibility to change some of the "not invented here"
philosophy.
Mikael, I think it would be helpful is you did not continue with such an
accusatory tone.  Tent is not an NIH system, they have a well designed
JSON framework.
I'm sorry if I've had a too harsh tone, as I don't intend to offend anyone. I hope however that people on this list are used to the shortcomings of text communication. Otherwise I may have to pardon myself in future discussions as well, though I'll of course try to keep a more gentle tone.

Either way I think this has come a bit off-topic especially given the subject of the thread .)

I think we have explained our respective opinions with the mentioned protocols/APIs quite well and I hope from now on hellekin can get some Tent and Friendica folks over here, as well as we can discuss the proposed consensus manifesto with many developers from a wide variety of projects.

I'll spend some more time with reading the manifesto in the comings days for sure.
Melvin Carvalho
2013-01-03 17:51:55 UTC
Permalink
Post by hellekin (GNU Consensus)
I stopped caring about Tent.io when I realised they were working
_against_ established standards and methods
*** Mikael, I hope this misunderstanding can be cleared when the Tent
people join the conversation. As far as I can tell, they thought the
lack of privacy in OStatus' scope was enough of a reason to take another
approach. I can't blame them for that choice. There are other
occurrences of "forks" that were beneficial to the community: look at
how Merb created a better Rails, by focusing on different features, that
ended up being integrated into the core of Ruby on Rails in the next
major version.
record of building same origin social networking sites
*** Indeed Melvin, but that is not comparable: building a single Web
site is a simple thing to do. Making it work seamlessly with others is
another story.
If we look at OAuth, the specification sounds great, but the
implementation disastrous: why would I give a third-party write access
to my profile if it only needs to access a feature (my status updates)
to read it! That entirely nullifies the intent of the specification.
That should be addressed as a blocker bug, because privacy is more
important than convenience.
Next iterations of OAuth will certainly address that mistake, but it
unintended consequences can be terrible in that context.
Sure OAuth is not ideal, in fact the main author left and called it a
fail. Personally as far as web 2.0 tech goes I think it's not bad, and has
a role. There are many ways to do authentication. But you start first
with identification.
Post by hellekin (GNU Consensus)
There have been instances of social web sites ineroperating with XMPP
too.
*** I started a list of potential partners on the wiki: it indeed
contains XMPP-based projects. It needs some love: some categories,
templates, etc. Please reply to that point by mentioning [wiki] in the
Subject.
try and operate from a perspective of tolerance, and see what
you can learn from other projects.
*** Although I agree with your idea, I would recommend against using the
word "tolerance". Historically, a "tolerance" was a permission granted
by the religious authority to allow infidels to access Church-controlled
areas, for e.g., trading. Therefore, unlike the common understanding of
that word today, it proceeds from a dominant perspective that is far
from our purpose of integrity and transclusion, and from your idea.
Appreciate your point of view but I am not using the religious connotation,
but rather, tolerance is one of the axioms of the web:

http://www.w3.org/DesignIssues/Axioms.html

Perhaps you would be more comfortable with the term 'respect'. That each
project or tech should show respect for the view points of others.
Post by hellekin (GNU Consensus)
Indeed we have lost valuable partners, such as friendica from this.
*** Who is "we"? I listed Friendica as a stakeholder, and a potential
partner on the wiki. In the next weeks, I will contact all projects
individually if they don't come by themselves. A consensus needs all
parties to emerge.
FSW community group. It's great if you are able to bring MIke or any of
the others back to the conversation. I'm a big fan.
Post by hellekin (GNU Consensus)
if you can take time to look at things from the perspectives of other
projects, you may gain a gain a great deal. Working together we can do
more.
*** +1
Time is a rare resource, making tools that can save time and enable
understanding of complex issues, to raise the bar above "us" and "them",
is a critical part of reaching a consensus.
+1

There are mailing lists for many project specific works all over. I joined
this list in the hope that we could foster a spirit of interop, that has
failed to happen in other groups. It's not about portraying one technology
as the 'best', it's about working together to achieve meaningful milestones.
Post by hellekin (GNU Consensus)
==
hk
Mikael Nordfeldth
2013-01-03 16:57:43 UTC
Permalink
Post by Melvin Carvalho
In this particular forum I think it would be more appropriate not to
belittle projects that you feel, do not think the same as you, but
rather, try and operate from a perspective of tolerance, and see what
you can learn from other projects.
Tent.io absolutely has a lot of ideas to bring to the table. However I strongly believe that to be a good, content and network agnostic protocol for social networking it must be reasonably compatible with other transport layers than HTTP as well.
Post by Melvin Carvalho
We have had major problems in the
past with people taking the "my way or the high way" approach.  Indeed
we have lost valuable partners, such as friendica from this.
Really? I have seen you write this several times now but haven't bothered yet to question it. But what makes you draw the conclusion that Friendica has "left" fedsocweb ambitions?

I can only see increased activity on Friendica's end and lots of interest in maintaining good compatibility with other federated social services.
Loading...