From danbri at danbri.org Tue Jun 16 02:32:39 2009 From: danbri at danbri.org (Dan Brickley) Date: Tue, 16 Jun 2009 09:32:39 +0200 Subject: [Social] Opera Unite: webserver in a browser Message-ID: <4A374A97.3040504@danbri.org> Opera just announced this - http://dev.opera.com/articles/view/an-introduction-to-opera-unite/ http://labs.opera.com/news/2009/06/16/ http://unite.opera.com/ Basically they expose webserver via the browser, and it seems also offer some proxying of this into public URIs like http://mymac.chrismills.operaunite.com/ I'm crossposting this to the XMPP Social list and the W3C SocialWeb XG list, since the intro in http://labs.opera.com/news/2009/06/16/ has some interesting motivation re social network and data portability, and I've lately been wondering about design decisions where I'm setting up personal/domestic computing APIs and feel drawn to XMPP rather than HTTP mainly due to NAT/Firewall traversal issues: XMPP services on a laptop can be universally addressed, unlike HTTP services. So I wanted to ask - is there a XEP spec for proxying HTTP over XMPP? Would this be relevant to Opera Unite scenarios such as the following? """Social networking is important, but who owns it ? the online real estate and all the content we share on it? How much control over our words, photos, and identities are we giving up by using someone else?s site for our personal information? How dependent have we become? I imagine that many of us would lose most of our personal contacts if our favorite Web mail services shut down without warning. Also, many of us maintain extensive friend networks on sites like MySpace and Facebook, and are, therefore, subject to their corporate decisions via ?Terms of Service? and click-through agreements. Furthermore, what does it mean anyway to be connected to hundreds of our ?closest? friends? What about our real social networks, the people we want to interact with on a regular basis (like once a week, or even every day)? Why are online solutions to help us with our real-world social needs so few and far between? We are connected to a Web that has democratized much and is an amazing source of information. However, "the wisdom of the crowd," along with the notion that our data ought to live on other people's computers that we don't control, has contributed to making the Internet more impersonal, anonymous, fragmented, and more about "the aggregate" than the individual. In fact, quite the opposite of the original promise. For too long, we?ve been going online to connect to each other, but sacrificing intimacy as a result.""" thanks for any thoughts, Dan From dave at cridland.net Tue Jun 16 05:34:04 2009 From: dave at cridland.net (Dave Cridland) Date: Tue, 16 Jun 2009 11:34:04 +0100 Subject: [Social] Opera Unite: webserver in a browser In-Reply-To: <4A374A97.3040504@danbri.org> References: <4A374A97.3040504@danbri.org> Message-ID: <27893.1245148444.945048@puncture> On Tue Jun 16 08:32:39 2009, Dan Brickley wrote: > I'm crossposting this to the XMPP Social list and the W3C SocialWeb > XG list, since the intro in http://labs.opera.com/news/2009/06/16/ > has some interesting motivation re social network and data > portability, and I've lately been wondering about design decisions > where I'm setting up personal/domestic computing APIs and feel > drawn to XMPP rather than HTTP mainly due to NAT/Firewall traversal > issues: XMPP services on a laptop can be universally addressed, > unlike HTTP services. So I wanted to ask - is there a XEP spec for > proxying HTTP over XMPP? Would this be relevant to Opera Unite > scenarios such as the following? http://xmpp.org/extensions/inbox/jingle-httpft.html describes a proposal for setting up HTTP over Jingle - by which I mean an HTTP session running on a transport negotiated using the Jingle protocol over XMPP. I think something along those lines would be the most seamless, although the firewall traversal properties of existing transports aren't exactly great - there is, of course, always IBB, and when newer transports come along, it'd simply use them. Alternately, you could write something that actually rewrote the HTTP requests into XMPP stanzas, for a more native-XMPP feel - so HTTP methods and header fields would become XML elements within stanzas. It would probably be more efficient, but I'm less than convinced it'd be worth the effort. The tricky bit is that you need do proxy the HTTP somewhere, and have a way of transforming the HTTP URI into a jid and back again. I'd guess something like http://example.com/~username/ => username at example.com would be the most obvious, but the tilde has i18n issues, so perhaps just http://example.com/username/ or maybe even have some way to allow clients to query what their external URI root looks like. Dave. -- Dave Cridland - mailto:dave at cridland.net - xmpp:dwd at dave.cridland.net - acap://acap.dave.cridland.net/byowner/user/dwd/bookmarks/ - http://dave.cridland.net/ Infotrope Polymer - ACAP, IMAP, ESMTP, and Lemonade From alexis.richardson at gmail.com Tue Jun 16 06:48:58 2009 From: alexis.richardson at gmail.com (Alexis Richardson) Date: Tue, 16 Jun 2009 12:48:58 +0100 Subject: [Social] Opera Unite: webserver in a browser In-Reply-To: <27893.1245148444.945048@puncture> References: <4A374A97.3040504@danbri.org> <27893.1245148444.945048@puncture> Message-ID: <167204d20906160448q3bd95966q767e192c173524d4@mail.gmail.com> Folks, Here is another way of serving HTTP requests from a browser and for doing various kinds of relaying: http://www.reversehttp.net/ It's HTTP specific so I am not sure if it's OT for this list. If not, apologies. Nevertheless: MBOI and comments are welcomed. I am cc'ing the author, Tony Garnock-Jones from the RabbitMQ team. Cheers, alexis On Tue, Jun 16, 2009 at 11:34 AM, Dave Cridland wrote: > On Tue Jun 16 08:32:39 2009, Dan Brickley wrote: >> >> I'm crossposting this to the XMPP Social list and the W3C SocialWeb XG >> list, since the intro in http://labs.opera.com/news/2009/06/16/ has some >> interesting motivation re social network and data portability, and I've >> lately been wondering about design decisions where I'm setting up >> personal/domestic computing APIs and feel drawn to XMPP rather than HTTP >> mainly due to NAT/Firewall traversal issues: XMPP services on a laptop can >> be universally addressed, unlike HTTP services. So I wanted to ask - is >> there a XEP spec for proxying HTTP over XMPP? Would this be relevant to >> Opera Unite scenarios such as the following? > > http://xmpp.org/extensions/inbox/jingle-httpft.html describes a proposal for > setting up HTTP over Jingle - by which I mean an HTTP session running on a > transport negotiated using the Jingle protocol over XMPP. > > I think something along those lines would be the most seamless, although the > firewall traversal properties of existing transports aren't exactly great - > there is, of course, always IBB, and when newer transports come along, it'd > simply use them. > > Alternately, you could write something that actually rewrote the HTTP > requests into XMPP stanzas, for a more native-XMPP feel - so HTTP methods > and header fields would become XML elements within stanzas. It would > probably be more efficient, but I'm less than convinced it'd be worth the > effort. > > The tricky bit is that you need do proxy the HTTP somewhere, and have a way > of transforming the HTTP URI into a jid and back again. I'd guess something > like http://example.com/~username/ => username at example.com would be the most > obvious, but the tilde has i18n issues, so perhaps just > http://example.com/username/ or maybe even have some way to allow clients to > query what their external URI root looks like. > > Dave. > -- > Dave Cridland - mailto:dave at cridland.net - xmpp:dwd at dave.cridland.net > ?- acap://acap.dave.cridland.net/byowner/user/dwd/bookmarks/ > ?- http://dave.cridland.net/ > Infotrope Polymer - ACAP, IMAP, ESMTP, and Lemonade > From dmeyer at tzi.de Wed Jun 17 13:18:06 2009 From: dmeyer at tzi.de (Dirk Meyer) Date: Wed, 17 Jun 2009 20:18:06 +0200 Subject: [Social] Opera Unite: webserver in a browser In-Reply-To: <27893.1245148444.945048@puncture> (Dave Cridland's message of "Tue\, 16 Jun 2009 11\:34\:04 +0100") References: <4A374A97.3040504@danbri.org> <27893.1245148444.945048@puncture> Message-ID: <87my86sp0h.fsf@phex.sachmittel.de> Dave Cridland wrote: > On Tue Jun 16 08:32:39 2009, Dan Brickley wrote: >> I'm crossposting this to the XMPP Social list and the W3C SocialWeb >> XG list, since the intro in http://labs.opera.com/news/2009/06/16/ >> has some interesting motivation re social network and data >> portability, and I've lately been wondering about design decisions >> where I'm setting up personal/domestic computing APIs and feel >> drawn to XMPP rather than HTTP mainly due to NAT/Firewall traversal >> issues: XMPP services on a laptop can be universally addressed, >> unlike HTTP services. So I wanted to ask - is there a XEP spec for >> proxying HTTP over XMPP? Would this be relevant to Opera Unite >> scenarios such as the following? > About file sharing, I plan to release a file server XEP proposal based on XEP-0265 at the beginning of July. The idea behind it is similar to the UPnP AVServer: browse and access file from another client. > Alternately, you could write something that actually rewrote the HTTP > requests into XMPP stanzas, for a more native-XMPP feel - so HTTP > methods and header fields would become XML elements within stanzas. > It would probably be more efficient, but I'm less than convinced it'd > be worth the effort. I'm playing with HTTP, Jingle and XMPP in my head for a long time. The question is: what do we want? Why do we want it? 1. We don't want HTTP over something, we just want to display HTML sites and stuff like that. What about And to make it faster, maybe merge XEP-0265 into it. 2. We want HTTP as it is over Jingle http://www.tzi.de/~dmeyer/jingle-http.html 3. We want to merge HTTP and XMPP somehow
bar
text/html
BASE64 Why is it usefull? Maybe http://en.wikipedia.org/wiki/CE-HTML Just my mind having strange ideas.... Dirk -- Misfortune, n.: The kind of fortune that never misses. -- Ambrose Bierce, "The Devil's Dictionary"