Re: [libreplanet-discuss] Tackling Network Effect

[wayne, steve]

I agree that we should try something that already exists. This is something I found while perusing Diaspora's federation protocol:

http://code.google.com/p/webfinger/wiki/WebFingerProtocol

This protocol is intended to allow one to find the public services that another person uses by using their email address as a key. I think earlier in this thread someone mentioned using our hypothetical directory protocol to find email addresses...using what? Their names? I can foresee problems with not knowing how to spell a person's name, uniqueness (which Mike/Mikael/Michael Anderson's email should a service implementing this protocol give out?), etc. 

On the other hand, if you know a person well enough then you may know one of their email addresses which is already a unique identifier. I propose that we use this as our primary lookup key.

The advantage of using Webfinger to create a federated directory lookup service is that it is already used by Diaspora, and it is already in draft form...we would just be doing the groundwork of understanding what open network services are available (diaspora, jabber, ...?) and figuring out how to connect with servers implementing those services, then writing a small reference implementation of our federated directory lookup service.

Does that make sense?

wayne

On Wed, May 16, 2012 at 12:12 PM, Ramana Kumar <ramana.kumar@gmail.com> wrote:

On Tue, May 15, 2012 at 12:18 AM, Michał 'rysiek' Woźniak
<rysiek@fwioo.pl> wrote:
> Dnia poniedziałek, 14 maja 2012 o 15:16:38 Ramana Kumar napisał(a):
>> On Mon, May 14, 2012 at 12:42 PM, Michał 'rysiek' Woźniak
>> > We could create a *simple* directory protocol (no, LDAP won't do)
>> > for finding out people's e-mails, JIDs, etc (if, of course,
>> > those people would provide such data); integrate it into
>> > available solutions (client-side: Pidgin, mail clients,
>> > Diaspora, etc.; server-side: ejabberd, mail servers, Diaspora,
>> > etc.) - when providing profile data (like e-mail addr) users
>> > could select to make that info available for federated access by
>> > other servers and clients.
>>
>> Are there any existing directory protocols that would serve this
>> purpose?
>
> Not sure, haven't done the research.
>
>> What's wrong with LDAP? (It has the advantage of being
>> already standard.)
>
> It has the huge disadvantage of being extremely complicated,
> cumbersome, difficult to implement and use. We need something much,
> much simpler, that can, however, be built upon LDAP. Maybe a subset?

I think a subset of LDAP would be a good starting point.
In particular, just using LDAP will confirm whether it really is too
complicated and focus our attention on what we would want from a
simpler protocol.

>
>> Are we reinventing FNS (XFN)?
>
> No. FNS AFAIK was used for naming machines, not keeping human contact
> info.
>
>> If a new simple protocol is required, let's make a short spec of it
>> here and now!
>
> Sure, as soon as I get some minimal sleep... ;)

(Or alternatively, pick out the bits of LDAP to use; or maybe I can
try to do that based on what we come up with for the simple protocol.)

>
>> Michał, I hope you will oblige to start on that, or otherwise
>> explain a bit more what you mean.
>
> Yes. I would love to, and I am very happy this seems to be picking up
> steam on this list. I will try to do some preliminary specs during the
> next few days. Shall we use a pad?
> http://pad.telecomix.org/simple-contact-federation

Ah, yes, sure!

>
>> > We could create a website for "leave your proprietary
>> > network/service day", hold such a day say once per two months.
>> > People could log-in with their proprietary network logins
>> > (Facebook/Twitter/Google+) and pledge leaving the network for a
>> > libre one (like Diaspora) or at least setting up an account on a
>> > libre network  on a given day, or when the number of their
>> > friends pledging the same reaches a given number.
>>
>> LibrePlanet, is there anyone here, or do you know of anyone, who
>> would be able and willing to design this website?
>> Also, to host it?
>
> I can provide for some minimal hosting, either via my lab or the FLOSS
> Foundation here in Poland.
>
>> I am willing to pay for domain registration and hosting if nobody
>> else volunteers, or no organisation can donate it, if there is a
>> good design. (Of course, donations would be welcome to help spread
>> the cost!)
>
> I would suggest using an established name for starters. LibrePlanet
> seems a great place, but I don't want to impose my ideas on anyone
> here. A sub-domain maybe?

So what would you call it?

>
>> > This would obviously be sent to their walls, timelines or
>> > whatever it's called in their proprietary social network of
>> > choice. their friends would get the message and see that maybe
>> > they are not alone in their idea of leaving Facebook.
>> >
>> > This is obviously just a preliminary idea, it would need to be
>> > properly thought through and blueprinted.
>>
>> I thank you for taking the time to sketch the preliminary idea! It
>> sounds great.
>> I suggest we keep working out the details on list.
>
> Indeed, I'd love to see it taking off.
>
>> > Second idea that I might start realising soon myself is firing up
>> > cloud service providers that integrate many libre software
>> > services (like XMPP, e-mail, Diaspora, StatusNet, OpenID,
>> > Mozilla Sync, ownCloud, etc.) so that people would have a chance
>> > of using a *single* service for all their communication needs,
>> > while retaining control over their data (using standard, free
>> > software projects would mean switching the provider would be
>> > much easier than leaving Facebook).
>> >
>> >
>> > These two ideas should for obvious reasons be put into practice
>> > about the same time.
>>
>> Yes. Michał, would you be interested in making a prototype of such
>> a service with me?
>
> In fact, similar service (many integrated services, including
> eGroupWare, ownCloud, StatusNet, e-mail, webmail, XMPP, Mozilla Sync)
> is already set-up in the lab I run here at Warsaw University of
> Technology. It still needs a lot of work to be prime-time ready.
>
> I am also thinking of either spinning it off as a separate company, or
> doing it under the aegis of the Polish FLOSS Foundation.

Do you use it?
What kind of work is necessary to get it ready?
Can I try it out?

>
>> > > Describe the world where I find out Skype is nonfree and then
>> > > [go where] ([why there?]) and [do what] to get a free
>> > > replacement, that anyone can easily do.
>> >
>> > http://podupti.me is just such a place to go for "Facebook
>> > replacement". It needs better exposure, and much better
>> > information on-site, but it's the right idea.
>>
>> Presumably it can be advertised on that website for "leave
>> proprietary network day".
>> What about for Skype replacement?
>
> I am talking generally, Skype is just one of the building blocks. For
> that I would like to see either JID/Jingle or SIP.
>
>> > Problem is, in my opinion, that we need to refocus.
>> >
>> > We - the FLOSS Community - are by and large great technicians,
>> > programmers, we have the manpower and will to tackle *technical*
>> > issues.
>> >
>> > However, our problems currently are not purely technical; Free
>> > Software is often technically superior to proprietary/centralised
>> > solutions, yet users still use the proprietary platforms (ICQ/AIM
>> > being one of the examples).
>> >
>> > Why? Reasons are many. Sometimes it's the usability (like with
>> > Skype); sometimes it's the network effect - already vast amounts
>> > of users use the proprietary solution, so the incentive to use
>> > it is much stronger, as the aim is to keep in touch (Skype,
>> > Facebook, ICQ/AIM, etc.).
>>
>> Excellent point.
>
> Thanks.
>
>> > These two problems are not easy to solve for us, as we are not
>> > that competent in those areas. Especially usability would need
>> > new sort of people joining in and helping out - UX designers and
>> > the like. This is one of the reasons why I feel that while many
>> > things Canonical does are really bad (please, let's not get into
>> > a "Ubuntu is Evuhl" flame now, m'kay? ;) ), we are still - as a
>> > community - getting a lot of good stuff from the fact they are
>> > employing UX designers, for example.
>>
>> One method might be to spread the ideal of freedom in the Ubuntu
>> community, to leverage the existing talent there.
>> This might essentially be: promote gNewSense over Ubuntu. But it
>> might be more than that.
>
> It will be hard to promote gNewSense to Ubuntu users. I do not
> envision them moving to gNewSense in droves.
>
> However, we *should* point out blatant mistakes on Ubuntu's part. Like
> calling both Skype and Ekiga "Free" in Software Centre. There should
> obviously be two different categories, and that is something both easy
> to do and high-profile.

Point that out where?

>
>> > Now, the network effect thing is more "tackleable" by us. It's
>> > really a question of focusing on creating ways for people to
>> > feel that they are in a group. they are connected, easy ways to
>> > find other people, etc. Diaspora, as I wrote already, does it
>> > well. We should simply move this up on our priority list, so
>> > that our techie community would find that important to work on.
>> >
>> > Specifically, I strongly believe that tackling the network effect
>> > problem is more important than Gnash, Google Earth, FLOSS network
>> > router drivers. I think it should (after finding a suitable way
>> > of describing it) make it to the priority projects list.
>>
>> FSF, your response?
>
> I am also extremely interested in FSF's reply here.

They do read this list, right?

>
> --
> Pozdrawiam
> Michał "rysiek" Woźniak
>
> Fundacja Wolnego i Otwartego Oprogramowania