Heh, that's a good point. I could always set up a daily cron job to download the RSS feed, or feed my Jaiku page to WebCite or Iterasi on a regular basis. ;)
Yes I've been searching for a way how to do that. The main posts are alright since FF will keep them safe. But the comments are hard to backup. No RSS and no simple way using the API for comments.
Not familiar with WebCite and Iterasi. But a quick search tells me they are for scraping web pages? That might work.
Hopefully when Jaiku goes open source there will be some options. Wishing for an upgraded API and a way to dig through the main database.
The long threads with comments are what is the best thing with Jaiku I think. Gotta find a way to preserve these if Google decide to do something stupid.
Yeah, that's the essential idea behind WebCite and Iterasi. There's Warrick too, although they take ages to archive stuff, since their crawler is supposedly fairly thorough (although I only tried it once, and haven't examined the dump file they sent me).
@jesperlind When the relaunch happens there will be a Google Code project and people will be able to track the progress of feature requests and bug reports using a public tracker.
@vmlemon Welcome to Jaiku. Don't forget to explore the channels.
There's supposedly an option for multi-level crawling, although it doesn't seem to go very deep, and does weird stuff with JavaScript and CSS in external files...
I agree, it's always fun when they "lift the veil" on proprietary products, especially if they've been around for a long time (since there's a certain mystique of "just how does this thing actually work?" around them). I'd be interested to see how much they end up re-engineering Jaiku to fit Google App Engine.
Not sure why, but I can see them building a new "Made for Open Source" version, kind of like RealNetworks did with their Helix products.
Yes they will probably rewrite parts of the code. It will now have to be different from the current implementation I guess, since people will use it in all kinds of different ways.
Hope that Jaiku will embrace the OMB standards so that more networks can talk to each other. Rather than making up a new standard. On the other hand someone will probably implement OMB in one way or another when the Jaiku code is opened.
Hmm, I doubt it'd be easy to build, given that they're probably substantially different internally, and built with different languages, but a Laconi.ca/Jaiku hybrid (Jaikoni.ca?) might be an interesting proof-of-concept.
@jesperlind I'm actually on the OMB mailing list. I'm hopeful that if we can get the spec tidied up (there's a few things that need to be fixed) I'll be able to implement OMB support for JaikuEngine in my spare time. That should enable people here to keep in touch with people they know on various Laconica implementations. The other thing I really want to add is SUP support so that FriendFeed's Jaiku support has lower latency.
@adewale Cool. I look forward to seeing what you come up with. I'm not sure how well certain Jaiku-specific concepts/data types would map though, but I'm sure you'll find a way. (Unless you're going for the lowest common denominator).
If they accept my proposal for using Atom Entries as the message format then we don't have to settle on a lowest common denominator approach. Instead there would be a core set of features that everybody has to support but JaikuEngines would be free to add extra data to their payload i.e. thread-id or channel-id.
Systems that don't have channels or threading would be free to ignore the extra data and just treat every message as a new post.
@adewale Interestingly, some comments (in a truncated/preview form) are available in the Atom feed, although the RSS feed doesn't have them.
Interestingly, there's some proprietary/Jaiku-specific data in the RSS feed (XML namespace xmlns:jaiku="http://jaiku.com/ns") related to "timeslices" and in the <jaiku:user> tag that may be of use. It looks like they're not using standard <img> XHTML tags for the avatars and Jaiku logo.
The RSS dump at http://www.webcitation.org/5eH1LHGsk might be a useful source of data to examine, if you're looking towards adding Jaiku "enhancements" to OMB. Of course, it's all XML, so it shouldn't add any incompatibilities as far as I know.
19 comments so far
At least there's a backup in FriendFeed: http://friendfeed.com/e/176f0e92-1e46-ba36-37f3-9b31f97ae4a9/Trying-out-Jaiku-I-wonder-how-long-it-lasts/
But this comment will disappear if they pull the plug, since there is no RSS for comments.
11 months, 1 week ago by jesperlind
Heh, that's a good point. I could always set up a daily cron job to download the RSS feed, or feed my Jaiku page to WebCite or Iterasi on a regular basis. ;)
11 months, 1 week ago by vmlemon
Yes I've been searching for a way how to do that. The main posts are alright since FF will keep them safe. But the comments are hard to backup. No RSS and no simple way using the API for comments.
Not familiar with WebCite and Iterasi. But a quick search tells me they are for scraping web pages? That might work.
Hopefully when Jaiku goes open source there will be some options. Wishing for an upgraded API and a way to dig through the main database.
The long threads with comments are what is the best thing with Jaiku I think. Gotta find a way to preserve these if Google decide to do something stupid.
11 months, 1 week ago by jesperlind
Yeah, that's the essential idea behind WebCite and Iterasi. There's Warrick too, although they take ages to archive stuff, since their crawler is supposedly fairly thorough (although I only tried it once, and haven't examined the dump file they sent me).
11 months, 1 week ago by vmlemon
An example of how this Jaiku message and comments appears after being passed through WebCite: http://www.webcitation.org/5eGStCtBo
11 months, 1 week ago by vmlemon
That's cool, you saved my comments from extinction!
Seems like a very useful service. Gonna try it on some web sites I'm running. Nice with extra backup of content.
11 months, 1 week ago by jesperlind
@jesperlind When the relaunch happens there will be a Google Code project and people will be able to track the progress of feature requests and bug reports using a public tracker.
@vmlemon Welcome to Jaiku. Don't forget to explore the channels.
11 months, 1 week ago by adewale
Thanks @adewale!
11 months, 1 week ago by vmlemon
There's supposedly an option for multi-level crawling, although it doesn't seem to go very deep, and does weird stuff with JavaScript and CSS in external files...
11 months, 1 week ago by vmlemon
@adewale: Really looking forward to the release on Google Code. It's gonna be great fun to finally see the inner workings of Jaiku.
11 months, 1 week ago by jesperlind
I agree, it's always fun when they "lift the veil" on proprietary products, especially if they've been around for a long time (since there's a certain mystique of "just how does this thing actually work?" around them). I'd be interested to see how much they end up re-engineering Jaiku to fit Google App Engine.
Not sure why, but I can see them building a new "Made for Open Source" version, kind of like RealNetworks did with their Helix products.
11 months, 1 week ago by vmlemon
Yes they will probably rewrite parts of the code. It will now have to be different from the current implementation I guess, since people will use it in all kinds of different ways.
I just found been reading these which I found interesting: http://controlyourself.ca/2009/01/08/giving-openmicroblogging-the-attention-it-deserves/ http://openmicroblogging.org/
Hope that Jaiku will embrace the OMB standards so that more networks can talk to each other. Rather than making up a new standard. On the other hand someone will probably implement OMB in one way or another when the Jaiku code is opened.
11 months, 1 week ago by jesperlind
Hmm, I doubt it'd be easy to build, given that they're probably substantially different internally, and built with different languages, but a Laconi.ca/Jaiku hybrid (Jaikoni.ca?) might be an interesting proof-of-concept.
11 months, 1 week ago by vmlemon
@jesperlind I'm actually on the OMB mailing list. I'm hopeful that if we can get the spec tidied up (there's a few things that need to be fixed) I'll be able to implement OMB support for JaikuEngine in my spare time. That should enable people here to keep in touch with people they know on various Laconica implementations. The other thing I really want to add is SUP support so that FriendFeed's Jaiku support has lower latency.
11 months, 1 week ago by adewale
@adewale Cool. I look forward to seeing what you come up with. I'm not sure how well certain Jaiku-specific concepts/data types would map though, but I'm sure you'll find a way. (Unless you're going for the lowest common denominator).
11 months, 1 week ago by vmlemon
If they accept my proposal for using Atom Entries as the message format then we don't have to settle on a lowest common denominator approach. Instead there would be a core set of features that everybody has to support but JaikuEngines would be free to add extra data to their payload i.e. thread-id or channel-id.
Systems that don't have channels or threading would be free to ignore the extra data and just treat every message as a new post.
11 months, 1 week ago by adewale
@adewale Interestingly, some comments (in a truncated/preview form) are available in the Atom feed, although the RSS feed doesn't have them.
Interestingly, there's some proprietary/Jaiku-specific data in the RSS feed (XML namespace xmlns:jaiku="http://jaiku.com/ns") related to "timeslices" and in the <jaiku:user> tag that may be of use. It looks like they're not using standard <img> XHTML tags for the avatars and Jaiku logo.
11 months, 1 week ago by vmlemon
The RSS dump at http://www.webcitation.org/5eH1LHGsk might be a useful source of data to examine, if you're looking towards adding Jaiku "enhancements" to OMB. Of course, it's all XML, so it shouldn't add any incompatibilities as far as I know.
11 months, 1 week ago by vmlemon
Heh, it appears that archiving the file failed that time. https://www.iterasi.net/Viewer.aspx?RootAssetID=771926 should have a working copy.
11 months, 1 week ago by vmlemon