Neo4j Hangout 2. Feb 2012 Transcript

gistfile1.txt

Hatim: yeah Peter, looking forward to it




Nigel Small: at least i have text :/




Toni Menzel: i am right on it at that moment. Hope to give stuff back in next days (better OSGi+SDN compatibility)




Nigel Small: feels like 1980s




Hatim: don't know how it is in Malmo, but winter has finally come here in Stockholm




Nigel Small: i've even tried booting windows - thats how desperate its become!




Toni Menzel: @Peter: does this Adobe Connect thing have a transcript feature ?




Toni Menzel: think some useful questions+answers could happen in this chat, lost forever if not indexed by google




Hatim: Toni, i can see chat log from before in this box 




Toni Menzel: .. which google will never index




Toni Menzel: anyway




Hatim: oh well I was thinking we should video record meetups, this is meant to be an online, sound video meetup




Hatim: (or so I thought)




ChrisR: low-tech suggestion: someone remember to post it to the mailing list?




ChrisR: copy & paste?




Peter Neubauer: yes, I can post the chat to the ML




ChrisR: (doesn't help if there's video)




Nigel Small: ascii art video? :-)




Andreas Kollegger: if somebody does kick into audio mode with screensharing, I think we can record it




Peter Neubauer: probably the best. We had video last time, do you want to do that? I felt I talked too much :)




Toni Menzel: technical question: G+ Hangout is no solution?




Toni Menzel: Ascii-Art Video.. thats good ;)




Andreas Kollegger: G+ is worth a try, I think




Peter Neubauer: well, if we are fine with no video, that would work




Andreas Kollegger: particularly without video




Andreas Kollegger: agreed PN




Peter Neubauer: so, you think we should try it?




Hatim: sorry, try what?




Toni Menzel: G+?




Peter Neubauer: yes.




Hatim: oh Ggoogle plus 




Toni Menzel: or IRC *g*




Toni Menzel: speaking of low tech




Hatim: yeah sure, it was more or less a fail at work for me




Hatim: btw, really happy that the Neo4j challange was streached with respect to timelines, just hope it gets more attention too in the more time 




Nigel Small: Peter: you getting on ok with the Geoff plugin then?




Nigel Small: i made a few minor additions this week




Peter Neubauer: yes, I have it deploying and everything.




Nigel Small: woo :-)




Peter Neubauer: right now, the biggest challenge is on where to put docs for all the plugins and components that are not in the main manual




Nigel Small: ahh ok - well mine is currently all in raw HTML




Niels: can someone tell me a little more about what GEOFF is?




Nigel Small: coded in VI like all good code




Peter Neubauer: yes, off you go Nigel, want video?




Nigel Small: video fail my end but i can type :-)




Peter Neubauer: maybe you get into presenter mode and show some GEOFF




Peter Neubauer: ahh ok




Nigel Small: basically Geoff started off as a simple file format for serialsed graph data




Nigel Small: and has evolved more recently into a fuller declarative language




Nigel Small: more or less and a "write" complement to Cypher




Niels: a data manipulation language?




Hatim: are we doing a google plus hangount?




Nigel Small: indeed




Hatim: URL?




Andreas Kollegger: it's the GEOFF/Cypher balance that I can't get straight




Andreas Kollegger: do you imagine somehow both languages intermingling




Nigel Small: Cypher = SELECT, Geoff = INSERT/UPDATE/DELETE




Andreas Kollegger: or remaining separate?




Peter Neubauer: I think this cries for a Andres/Nigel love day of talking




Nigel Small: separate i think for now but both share a common(ish) syntax




Nigel Small: geeky man-love?




Peter Neubauer: and, I would like to do the same with Mattias and Niels on supernodes and stores




Peter Neubauer: ye, always!




Nigel Small: so for example....




Nigel Small: (A) {"name": "Bert"}




Niels: would it be possible to do something like analogous to SELECT UPDATE?




Michael Hunger: Hey




Peter Neubauer: Hi Iron Man




Nigel Small: describes a node, named "A" with a single property




Michael Hunger: Can Nigel share his Screen and demo stuff?




Nigel Small: sorry chaps nothing much to demo from here right now - havent had enough time to prepare anything fancy!




Michael Hunger: but we can have a look at your docs webpage ?




Nigel Small: most of the info is on http://py2neo.org/geoff but agreed i think i need to work more with Neo guys on how it fits against Cypher




Michael Hunger: Don't hear anything, is audio working for you guys?




Hatim: i cannot hear or see 




Hatim: well i can see the chat




Michael Hunger: Now that cypher can select different parsers it would be cool to integrate cypher + geoff




Peter Neubauer: yes really




Nigel Small: how does the parser selection fit in then?




Nigel Small: (i will try to prepare a powerpoint for next time maybe?)




Peter Neubauer: no PPT please! this is a hangout ;)




Nigel Small: ok lol




Nigel Small: sorry i said something evil




Peter Neubauer: live demo is da sh..




Michael Hunger: just showing your docs and talking is more than enough




Peter Neubauer: PPT just sux in your time




Nigel Small: hehe ok point taken




Michael Hunger: do you have the server plugin ready for a live demo Peter ?




Nigel Small: some of us work in low-tech worlds!




Peter Neubauer: mmh, I can pull it up, just a moment




Michael Hunger: I'd say Nigel pulls up the docs and enables his mic and then we gonna chat about Geoff




Peter Neubauer: yes, will pull up the plugin ...




Michael Hunger: yes




Toni Menzel: y




Hatim: i hope people cannot hear me cheawing gum




Michael Hunger: chew louder :)




Andreas Kollegger: why geoff within json?




Andreas Kollegger: why not send the raw geoff?




Andreas Kollegger: er, um.




Michael Hunger: but there is not really a DSL to build up geoff? right?




Hatim: yes 




Michael Hunger: and each rule fits nicely as a entry of the json entry




Michael Hunger: so we should talk to mattias to have a sink for index content




Michael Hunger: via GraphOperations




Peter Neubauer: IndexReader r = IndexReader.open( "prdb_index");		int num = r.numDocs();for ( int i = 0; i < num; i++){	if ( ! r.isDeleted( i))	{		Document d = r.document( i);	        System.out.println( "d=" +d);	}}r.close();




Michael Hunger: start n= ... return GEOFF(n)




Michael Hunger: or GEOFF(p) or GEOFF(r)




Michael Hunger: p being path and r being relationship




Michael Hunger: application/geoff




Michael Hunger: Accept: application/geoff




Peter Neubauer: yes, true




Nigel Small: sounds like "jeff"




Hatim: so a stupid question, can geoff  be used  to input data 




Nigel Small: yes via the load routine




Nigel Small: install the plugin and pass rules to one of the load methods




Nigel Small: over REST




Nigel Small: all data from a single file will be created in a single transaction




Hatim: and can some thing be done along this line in cypher 




Hatim: cypher + geoff




Hatim: (not REST)




Nigel Small: most of the work so far has been to enable the server plugin but no reason it couldnt be used without




Michael Hunger: I have the idea of combining cypher and geoff like this




Nigel Small: http://py2neo.org/geoff/clients/java-embedded/




Michael Hunger: start n  match .... where .... with n, r, p do [geoff code using n,r,p and cypher inputs as parameters]




Nigel Small: not many docs on that yet sorry :-)




Peter Neubauer: yes but the tests are good.




Michael Hunger: with instead of return




Hatim: ok




Hatim: yeah, this sounds very interesting




Michael Hunger: and should be possible with the new parser-selector in cypher




Peter Neubauer: yes, it's a perfect project for a spike




Michael Hunger: great blog posts by max again this week 




Nigel Small: got a link?




Hatim: yeah, make me almost wana start learning ruby (more seriously)




Hatim: -> the blog posts from Max




Hatim: any thing else on the agenda here 




Hatim: ?




Hatim: not sure if we will be doing a video/audio meetup from now on but i thought recording it could be useful (provided people give permission)




Hatim: we may start recording our in life meetups in Stockholm and post them to vimeo




Hatim: or some thing




Nigel Small: does anyone visit the IRC channel very often?




Nigel Small: i pop in occasionally but its always pretty quiet




Michael Hunger: not really




Michael Hunger: too much other stuff going on :(




Hatim: Nigel Small, irc has lost alot of traction with the advent of twitter (IMO)




Michael Hunger: can't keep up with all the channels anyway




Nigel Small: twitter can be slightly less focussed! :-)




Peter Neubauer: yes. IRC is one more channel. I would liek to keep interaction focused




Michael Hunger: is there a nice irc channel notifier ?




Michael Hunger: which sends a growl message when somethings happing




Peter Neubauer: so, we could switch to either IRC or G+




Peter Neubauer: if we are not doing very often, then chat is fine. however I like the occasional screen sharing.




Nigel Small: the screen sharing is useful i think




ChrisR: I enjoyed having this session, but the time of day is bad for me (time zones :-( ) ... I couldn't pay much attention




Nigel Small: more caffeine?




Hatim: yeah, Peter may be we can do an early time slot every other meetup




ChrisR: No, work expects me to actually be producing code for them during the day :-) (it's 4PM here now)




Nigel Small: ahh ok :-)




Michael Hunger: Hmm we could also encourage other hangouts happening w/o being organized by us?




ChrisR: But don't schedule for me by any means....I'm ok on the mailing lists for now...




Michael Hunger: so that we have one per week in a different timezone ?




Peter Neubauer: yes, I think just announce a new meetup, and we will make it happen.




Michael Hunger: there are more than enough topics




Peter Neubauer: there has been interest for a later one for Aisa/AU




Michael Hunger: I would love to hear from the bio4j and structr guys in person for instance




Peter Neubauer: yes, me too. we can reach out and invite them personally, and have Marko demo crazy recommendation stuff




ChrisR: Thanks for your time guys...if I see another hangout I'll try to join. I think it's a nice idea. Bye.




Nigel Small: bye chris




Peter Neubauer: bye, need to drop out too




Peter Neubauer: nice to talk to you all, so, next time Adobe or something else?




Hatim: ok, Peter




Niels: was there anything else on the agenda today?




Nigel Small: adobe unless something else agreed on mailing lists?




Hatim: well, besides mailing list where do we discuss priorities of features




Niels: what sort of features would you suggest?




Hatim: that was one thing i am interested to know, after seeing the Doodle list of features




Hatim: Niels, no features as such, i would like to vote up or down on features 




Nigel Small: there was a poll wasnt there?




Hatim: like in Springsource JIRA, you can watch and vote for issues 




Peter Neubauer: yeah, sadly Github doesn't have it




Niels: is that an option on GitHub?




Peter Neubauer: I would like to keep it there, maybe some substitute




Niels: can't we vote up/down by adding a comment with +1/-1 ... it's a bit more time consuming to distill the numbers though




Peter Neubauer: but I really love GIThub issues otherwise, voting is just not there. maybe Niels or Hatim, could you investigate options to get  voting integrated?




Peter Neubauer: it also gives a great indication where to pitch in, and for the Neo technology engineers what to fix first




Niels: wish i could vote more than once on one particular issue :-)




Hatim: sure, will take a look, if there is none then this sounds like a small exercies with spring-social and github :)




Hatim: Niels . hehe, same here. Which issue is dear to you at the moment




Peter Neubauer: right, we could host a voting-app as part of the neo4j challenge!!




Peter Neubauer: that is a winner




Niels: densely connected node (has been for many months now




Michael Hunger: I also like youtrack :) which has all of that




Peter Neubauer: yes Niels +1 on that.




Peter Neubauer: ok I need to go




Nigel Small: bye peter




Michael Hunger: the densely connected nodes are being worked on, but right now there was other more pressing stuff Niels




Peter Neubauer: voting is a great action point. let's get that in place guys!




Michael Hunger: great idea hatim !




Niels: i understand there are more pressing issues... i just keep annoying you guys until it's fixed




Hatim: thanks, after we have voting i have to find lots of friends for the nice features as well :)




Michael Hunger: thanks Niels




Hatim: bye Peter




Michael Hunger: thats goo




Michael Hunger: good




Hatim: Nelis for the un-initianted, how do densly connected nodes help you 




Michael Hunger: so that would also be a nice thing for SDN + Spring Social




Hatim: or solve what kind of problems...never read any thing about it before




Michael Hunger: if you are brave enough Hatim :)




Niels: they don't necessarily help me, they are a nuisance when they occur,due to store layout




Michael Hunger: yes, that's true Niels




Hatim: Michael i can try :), i was fishing for ideas any way 




Michael Hunger: especially when the relationships are spread across the whole relationship-store-file




Michael Hunger: it would be even better if that app could sync with the github issues :)




Michael Hunger: back and forth




Michael Hunger: one idea is to represent voting on gh as labels (1 unimportant ... 5 very important)




Niels: is there a label per user, or would it be wiki-like, where you can vote it as 5, and i can change that in 4?




Michael Hunger: just for the viz on github, just 3-5 labels, otherwise it gets too crowded




Michael Hunger: in the neo4j app we could have arbitrary stars




Hatim: hmm, 




Michael Hunger: start issue = node:issues("id:*") match issue<-[v:VOTED]-user return issue, avg(v.stars)




Michael Hunger: start issue = node:issues("id:*") match issue<-[v:VOTED]-user return issue.id, avg(v.stars)




Niels: i see you already speak cypher natively




Nigel Small: michael is a graph-powered cyborg




Hatim: well i think there are lots of ideas regarding the voting part and how it can be analyzed etc. but the most importnat thing for the app is to be super easy 




Michael Hunger: sure :)




Hatim: for people to vote easily




Michael Hunger: cypher comes easy as nth language




Niels: any chance there will ever be a cypher version as an internal Scala DSL




Niels: ?




Michael Hunger: yep, so the UI would be interesting, perhaps utilize the ask-ken stuff max showed in his blog-posts ?




Michael Hunger: Niels I think fakod did one or adriano almeida




Niels: then i will look into that, as a Scala programmer myself, I am more interested in such an approach than parsing text




Niels: though i have written my own traversal DSL in scala




Nigel Small: well it seems easy to scrape the github issues using their API so it should be trivial to create a mini voting app: http://py2neo.org/voting/




Hatim: Micheal, i will do a small write up for the voting app




Hatim: but at the same time we should have a production voting solution , simple and easy to vote for github features, i will ask around, other wise even Doodle is ok 




Michael Hunger: yay Nigel !




Michael Hunger: OTOH, why not use the voting app for production?




Hatim: Micheal well, i meant is that we should not wait for a custom solution, 




Hatim: voting in it self is important




Hatim: but i will give it a good shot, hopefully this weekend, now i hvae some idea :)




Hatim: ok ppl 




Hatim: i have to go now, see you in next meetup




Nigel Small: ciao




Niels: i have to get going... looking forward to a voting option




Michael Hunger: great !




Michael Hunger: take care :)




Nigel Small: @michael i think i will need to do some considerable work on speeding up geoff




Nigel Small: i dont think the current implementation will scale too well




Nigel Small: when dealing with many rules in a txn