Re: D-CENT: state of the art - not

Hello Harry, long time no see.

Harry Halpin wrote:
> > Quick note - as you may know, there is over a year in the between
> > project proposal writing (in 2012) and the project review/begin date
> > (2013).

I don't think it's normal for EU proposals to age so fast.
They should still be sensible two years later.

> > W3C will make sure DCENT is up to speed on standards, with Neo
> > doing implementation work. Neo is a lean development firm, so we haven't
> > made any choices architecturally yet, as we're interviewing users and
> > figuring out their needs before jumping on board with any protocols.

Good. So basically the EU grant was given on the idea of getting
people together to do something? Well ok, here I am. I do things.

> > Andreas - also, there has been exceedingly little work in the field
> > actually since 2012, so I think your use of the term "irresponsible" is
> > irresponsible :) Regardless, the only open-source I've seen some
> > progress as regards interop testing was OStatus.

No surprise since interop isn't currently a useful goal. Why work
on the interoperability of solutions that do not scale and thus do
not achieve what is requested from them? The last thing we need is
another XMPP - a standard that does not fulfil the requirements
because it has lack of scalability built-in at the foundation.

I mean lack of scalability concerning large subscriber numbers to
social feeds - the Twitter use case. A Twitter based on XMPP is
impossible. I presume it stops working reasonably after some hundred
subscribers. Unfortunately I haven't seen any numbers published -
projects that run into these hard limits simply turn quiet or disappear.
Would be more scientific if they were to publish their findings so
we can actually stop betting on technologies that didn't consider
the basics of scalability design when they were created.

On Mon, Feb 24, 2014 at 11:47:05AM +0100, Andreas Kuckartz wrote:
> The XMPP community has been active all the time. But that does not imply
> that XMPP is or should be the future of the Social Web. I think that we
> have reached a point where previous efforts to create distributed social
> media can be re-evaluated and a new start is possible based on that -
> and based on more recent developments, including the progress of Linked
> Data and mass surveillance
> 
> In any case it likely would not be that easy to connect Elgg instances
> using XMPP because PHP does not seem to be the best language for XMPP.
> It _might_ make sense for D-CENT to stick with OStatus, but then it
> should not be sold as _the_ future of the Social Web.

It may work as an isolated platform, but from my understanding an Elgg-
based platform provides neither scalability nor the confidentiality a
group of activists would need. There are reasons why Lorea developers
are running Elgg in maintenance mode and looking out for more appropriate
solutions.

> My main concern was and is not OStatus but all the projects which were
> already dead for several years when the proposal was written. I have
> seen something like that before and unfortunately sometimes carelessness
> regarding such details is neither restricted to a slide nor only the
> result of simple ignorance (which can be corrected) but of indifference.
> I would be happy to see that this fear turns out to be unfounded now.

It indeed looks a bit marketer style to mention the hype names of the
time.. or even of previous times.

> > The IndieWeb work is making some progess now around WebMention,
> 
> For those who are interested:
> https://indiewebcamp.com/webmention

Which is just an optimization of the already existing pingback hack
on top of HTTP. It solves no new problems, it is barely more efficient
than with XML inside and most of all it is no tool to implement a
scalable distributed social web.

Received on Wednesday, 26 February 2014 14:59:05 UTC