PMC Meeting Full 20070810
(9:45:28) modo (+ns ) por kubrick.freenode.net
(9:55:08) croo [n=crooney@89.19.88.38] ha entrado en la sala.
(10:02:53) jsSolutions [n=jsSoluti@ip72-197-81-229.sd.sd.cox.net] ha entrado en la sala.
(10:05:28) jsSolutions: hmmm, colin seems correct, not that much interest in PMC...
(10:08:16) croo: it's very quiet ok
(10:08:23) croo: i just went for a coffee
(10:09:52) CarlosRuiz: is strange, in general terms the projects is very quiet
(10:10:37) croo: yes the main channel is very quiet too ... friday afternoon in Europe ... many go earl from work!
(10:10:51) croo: also holiday season in many countries
(10:11:00) croo: in Europe
(10:12:31) jsSolutions: maybe just a generally bad time for a discussion
(10:12:52) CarlosRuiz: I mean more than that
(10:13:10) CarlosRuiz: council is quiet, committers are quiet
(10:15:17) trifon [n=Trifon@home-213-240-227-143.megalan.bg] ha entrado en la sala.
(10:15:46) CarlosRuiz: :-) you see, lonely
(10:16:12) trifon: yes i see, but i do no tunderstand why.
(10:16:47) croo: maybe the time is the issue ... as in GMT ... windows says GMT includes DST ... I think that catches many peopel out who see GMT as the time in London
(10:17:59) trifon: hi Colin
(10:18:08) trifon: how is your implementation?
(10:18:11) croo: hi daddy :)
(10:18:25) trifon: i expected to see many people today ...
(10:18:27) trifon: :)
(10:18:53) trifon: Victoria is taking 80% of my time
(10:18:57) croo: slow - but that's my fault... still pushing for quick solution now as there is a new prospect
(10:19:11) croo: :) I can imagine!
(10:19:46) croo: but at least you work from home right? so you don't have to leave her every morning for the day! a big plus I guess
(10:20:13) trifon: yes slow, but if user wants to have good system they must be patioent.
(10:20:32) trifon: yes i have small office room next my room but there is not difference.
(10:20:38) trifon: now i mostly at home.
(10:22:01) croo: how is the skype servcie working out for you?
(10:22:16) trifon: till now 0 USD from it.
(10:22:19) catsapiens [n=catsapie@190.81.28.118] ha entrado en la sala.
(10:22:25) USER_ [n=chatzill@ADSL-TPLUS-100-43.telecomplus.net] ha entrado en la sala.
(10:22:46) catsapiens ha salido de la sala (quit: "ChatZilla 0.9.78.1 [Firefox 2.0.0.6/2007072518]").
(10:22:48) croo: trifon, really! that's disaqointing!
(10:22:50) USER_ ahora se llama fredtsang
(10:22:53) trifon: probably because i'm online las weeks
(10:22:55) croo: trifon, really! that's disaponting!
(10:23:03) trifon: yes it is 0.00
(10:23:11) trifon: we could ask Carlos.
(10:23:20) catsapiens [n=catsapie@190.81.28.118] ha entrado en la sala.
(10:23:26) croo: is he also offering such a service?
(10:23:28) Bahman [n=Bahman@adempieretech.org] ha entrado en la sala.
(10:23:43) trifon: yes.
(10:23:51) trifon: he posted in forums.
(10:24:03) CarlosRuiz: I have sold 30 minutes
(10:24:10) trifon: o good. so you lead here.
(10:24:17) croo: there is so much to read it's hard to keep track of everything :)
(10:24:20) CarlosRuiz: but then I realized that my paypal can't receive money
(10:24:26) trifon: ooooo
(10:24:32) trifon: thi sis not good.
(10:24:33) croo: :)
(10:24:35) CarlosRuiz: so I think I lost that money , hahahahaha
(10:24:47) trifon: i can send you that money.
(10:24:50) CarlosRuiz: so, I suspended the service
(10:24:56) trifon: but your client must send them to me first.
(10:25:37) trifon: other way is http://ikobo.com/
(10:26:22) CarlosRuiz: skype prime requires a paypal account to work
(10:27:05) trifon: Colin if you notice i have increased number of development services i offer.
(10:27:31) croo: is that possible :)
(10:27:53) croo: did you signup for the sf services?
(10:28:13) CarlosRuiz: is a shame I can't get the skype prime services, there are three persons asking me for such service
(10:28:17) trifon: yes, but i can't find how to define my service :(
(10:28:46) croo: why don't you just sign up for paypal too then?
(10:28:59) croo: (that was to carlos)
(10:29:16) CarlosRuiz: yes, I have a paypal account - but for Colombia they just allow sending money, not receiving
(10:29:29) croo: oh?
(10:31:14) croo: can you not just say you are somewhere else? or why not take tridon up on his offer and use his paypal?
(10:31:43) croo: tridon .. sorry Trifon! :)
(10:32:04) CarlosRuiz: no, they know where am I from because of my IP
(10:32:15) CarlosRuiz: but maybe I could make the trick with Trifon's paypal account
(10:32:38) croo: when you set up paypal? what if I set the account up? could you then use it?
(10:33:00) CarlosRuiz: don't understand
(10:33:12) trifon: Calros you can ask the to send money to me and i will send money to you by bank transfer.
(10:33:37) trifon: sorry Carlos.
(10:33:42) CarlosRuiz: thanks Trifon for the offer, I'll think about it
(10:34:06) trifon: you are welcome.
(10:34:13) CarlosRuiz: I mean, I'll think how can we set up the skype prime to reach your paypal :-)
(10:34:19) CarlosRuiz: well, what do you propose about this meeting?
(10:34:19) CarlosRuiz: there is no quorum even from the proposed members of PMC - here 4/8
(10:34:57) croo: no ...
(10:34:58) jsSolutions: i wonder if Malaysia is off-line?
(10:35:13) croo: would explain red1 AND low
(10:35:39) jsSolutions: red1 and heng sin are both not on skype
(10:35:39) croo: strange neither of them are here
(10:36:33) CarlosRuiz: I think we could try to discuss a little more in forums about the PMC
(10:36:52) CarlosRuiz: Colin and Trifon have pointed issues about the role
(10:37:07) catsapiens: what is PMC?
(10:37:24) CarlosRuiz: Project Management Committee
(10:37:27) croo: touche :)
(10:37:45) CarlosRuiz: no, I share some points from you and Trifon
(10:38:16) catsapiens: ahh
(10:38:19) croo: no I meant ... you said maybe we need to discuss it more and catspaien asks "what is it?" ... so it looks like we DO nee dto discuss it moer
(10:38:32) CarlosRuiz: :-)
(10:38:46) jsSolutions: well, maybe not discuss it more, but start working the role
(10:38:54) jsSolutions: even if 'unofficial'
(10:39:20) jsSolutions: then it can define itself
(10:39:25) CarlosRuiz: I feel it deserves more community discussion
(10:40:34) CarlosRuiz: even I think the votation can be cleared and voted again after an agreement in the role
(10:41:15) trifon: my point is Roadmap.
(10:41:34) CarlosRuiz: yes, I'm agree with the toughts you put in wiki
(10:41:41) trifon: i do not understand how someone could write roadmap
(10:41:46) trifon: thank's
(10:41:50) CarlosRuiz: instead roadmap we have talked about a wishlist
(10:42:04) CarlosRuiz: a real roadmap - as you pointed with developer names
(10:42:10) trifon: yes. it is more clear anf not consing
(10:42:12) CarlosRuiz: just can be done with money
(10:42:25) CarlosRuiz: with developer names and dates
(10:42:43) croo: well we can continue to meet on Fridays to discuss it move it forward ... address functional items that have arise... that's sort of doing something as Joel says without acually having an official group?
(10:43:09) jsSolutions: maybe try Tuesday at the old cc meeting time?
(10:43:15) jsSolutions: it seemed that was more supported
(10:43:28) croo: and in between clarify the role by discussion in the forums
(10:43:41) croo: ok tuedays
(10:43:44) CarlosRuiz: agree, maybe friday is a bad day ;-)
(10:44:07) trifon: yes in bulgaria in Friday everyone is thinkiung only for coming weekend...
(10:44:36) jsSolutions: i think something of a roadmap is important
(10:44:43) jsSolutions: people look for it
(10:45:05) jsSolutions: so having the old document there is confusing
(10:45:15) CarlosRuiz: yes, the point is how to name it - I feel better calling it a wishlist
(10:45:20) trifon: yes people look for it, but currenlty sponsored develpoment ifs better roadmpan than the one we have.
(10:45:27) croo: i too but I fail to see how it's achievable ... we could say what if commited to in terms of sponsorship but outside of that it requires money
(10:45:43) CarlosRuiz: the wishlist with assigned dates and developer -> becomes a roadmap
(10:45:55) croo: sorry that's "what is committed to"
(10:46:12) CarlosRuiz: but without dates and developers, is just a wishlist
(10:46:18) CarlosRuiz: an north of what we want
(10:46:34) CarlosRuiz: but this can go in other way
(10:46:39) CarlosRuiz: as it happened all this time :-)
(10:46:57) CarlosRuiz: we want something, but always we receive different and more of what we dreamt
(10:47:02) croo: yes ... if we set a much longer time by have only a couple of (stable) releases max a year it might be easier to get people to commit
(10:47:37) croo: if you say for example .. sometime in the next 6 months I will do this ... then peopel might feel less pressurised and commit
(10:47:50) jsSolutions: maybe two beta and two stable on a set schedule
(10:48:00) jsSolutions: and a weekly discussion about what is in
(10:48:02) trifon: Colin i think that it is not only to the relese frequency.
(10:48:15) jsSolutions: then keep the document alive
(10:48:20) jsSolutions: so people can rally around it
(10:48:21) trifon: all develpoment require time.
(10:48:24) CarlosRuiz: I think the current to motivator is sponsored development
(10:48:35) CarlosRuiz: the current top motivator
(10:48:42) trifon: some developers think that in thier free time they can make something but it is not always the case.
(10:49:02) CarlosRuiz: but we're just learning how to manage sponsored development
(10:49:13) CarlosRuiz: there can be big improvements in such process
(10:49:24) CarlosRuiz: to ease sponsored developments reach the trunk
(10:49:44) trifon: yes with sponsored dev we have things to imporve in the porcess.
(10:50:31) CarlosRuiz: as Colin pointed in forums - we can't push anything - just try to steer
(10:50:54) croo: i would like to see sponsored development happen in public svns ... so people can watch comment (to improve) and ,aybe even help out
(10:51:27) CarlosRuiz: try - because this steering wheel is difficult to keep in any course :-)
(10:51:27) CarlosRuiz: it has own life
(10:51:34) croo: so long as the named developers and co-ordinators can control that branch
(10:51:40) jsSolutions: croo, interesting point
(10:51:44) trifon: projet which i develop are in Adempiere SVN.
(10:51:57) jsSolutions: i think trifon is unique
(10:52:11) trifon: no. Joel. it is just more easy for me.
(10:52:11) jsSolutions: most develop privately
(10:52:23) trifon: other developers can try what i do nad give me advices...
(10:52:34) trifon: this is what i get from public svn...
(10:52:35) jsSolutions: then when they think it's perfect they might contribute
(10:52:59) jsSolutions: it's an issue not just with sponsored devs
(10:53:11) CarlosRuiz: I think very few people are looking for branches or contributions folders :-(
(10:53:20) CarlosRuiz: or patches
(10:53:44) trifon: why?
(10:53:47) CarlosRuiz: just trunk have eyeballs
(10:53:52) CarlosRuiz: I don't know why
(10:53:58) trifon: ok. i got th epoint.
(10:54:01) CarlosRuiz: maybe as Colin said, there are too much to read
(10:54:10) trifon: but emails from SVN are for complte SVN.
(10:54:22) croo: probably because people don't know what's happening in the other branches!?
(10:54:23) CarlosRuiz: personally, I just read e-mails from trunk
(10:54:44) trifon: yes this is true also. we produce alot but thi sis not our fault.
(10:54:52) croo: well it's trues there is enough to read in just the trunk :)
(10:55:15) CarlosRuiz: yes, but we must think about that
(10:55:26) CarlosRuiz: I feel that's the main role of these "Committees"
(10:56:16) CarlosRuiz: trying to find improvements points - proposing policies for improvement - etc
(10:56:36) CarlosRuiz: community can do that but a very slow pace
(10:56:43) CarlosRuiz: veeeeery sloooooooooooooooow
(10:57:03) trifon: what do you mean by policies for improvement?
(10:57:15) CarlosRuiz: :-) no, maybe it's hard community by itself can arrive to policies
(10:57:33) croo: yes ... I was thinking about that as I read about Debian .. there its believed the commitees slow it down while ubuntu is fast because it has the community!
(10:57:34) CarlosRuiz: community make and follow tacit policies
(10:57:39) croo: a reversal of here
(10:57:54) croo: I guess we simlpy do not have enough people contributing yet
(10:58:24) CarlosRuiz: I think we must find balance
(10:58:32) croo: I think there is a great conflict between doing things and working ... they don't always mix well
(10:58:39) CarlosRuiz: too much control (like Debian) stops the project
(10:59:03) CarlosRuiz: low control (like here) looks messy
(10:59:13) croo: but it results in an very stable product ... a very important aspect for an ERP I think!
(10:59:29) CarlosRuiz: which one Colin?
(11:00:19) croo: which one? I mean Debian is very stable and that's a good thing for an ERP
(11:00:23) CarlosRuiz: it's a hard decision - if you put too much control the trunk evolves slowly - but stable and sure
(11:00:30) CarlosRuiz: but I prefer
(11:00:37) CarlosRuiz: if you don't put controls the code can become unmanageable
(11:00:49) CarlosRuiz: the architecture can become unmanageable
(11:01:15) CarlosRuiz: inapprehensible
(11:01:53) croo: yes well I don't have answers either ... just making some observations :)
(11:01:59) CarlosRuiz: but we always (as committers or as PMC) will receive critics because of our labour :-)
(11:02:27) trifon: but all we have client which require some functionality and which leads to modifications which we make. at the end of the day it is our task to watch out our functionaity.
(11:02:43) CarlosRuiz: yes Trifon, that's my point
(11:02:48) trifon: i thikn that this will evolve to module maintainers or smething like this.
(11:02:59) CarlosRuiz: the top motivator currently are sponsored developments, or specific project needs
(11:03:09) CarlosRuiz: we need to find a way for those contributions to get trunk
(11:03:11) trifon: yes, right.
(11:03:12) CarlosRuiz: in an ordered way
(11:03:17) trifon: yes.
(11:03:20) croo: yes
(11:03:34) trifon: what means ordered way?
(11:03:50) CarlosRuiz: i.e. following architecture
(11:04:25) trifon: thi sleads to another questoin. what is definitiona of architecture in Adempiere?
(11:04:38) CarlosRuiz: i.e. just bringing ideas
(11:04:38) CarlosRuiz: we could ask to add a column or a table we need to make a design discussion
(11:04:44) croo: but I think, as Carlos has poiunted out before, moving the foundation itself forward is something that will very unlikely come from sponsorship
(11:05:05) croo: unless we meet some popel with deep pockets :)
(11:05:13) CarlosRuiz: yes Colin, that's right true
(11:05:32) CarlosRuiz: or we need to get money in other forms to sponsor such efforts
(11:05:50) CarlosRuiz: i.e. charging for advertising in wiki / .com / .org ...
(11:06:16) CarlosRuiz: or any way to get money for the project and sponsor hard foundation tasks
(11:06:33) trifon: Carlos you can't control everythig. it is possible someone else to come with better design. it is possible that we look to narrow and do not see what other developer see.
(11:06:48) CarlosRuiz: Trifon, I'm not trying to control nothing
(11:06:51) croo: well Joel made soem suggestion about that a long while ago
(11:06:56) trifon: ok.
(11:07:00) CarlosRuiz: I'm saying bringing community wisdom to design
(11:07:26) CarlosRuiz: i.e. design brainstorm meetings
(11:07:34) CarlosRuiz: with top developers
(11:07:55) trifon: but what happens if we have different opinion and i woul dlike to experiment.
(11:08:09) CarlosRuiz: nothing can stop you from that
(11:08:39) trifon: this will seem bad from user point of view.
(11:08:46) CarlosRuiz: we can allow experimenting in branches
(11:08:51) trifon: i think that user do not like ot see fights in furms.
(11:09:00) trifon: yes, right.
(11:09:01) CarlosRuiz: it depends on the fights
(11:09:06) trifon: we allow this and now.
(11:09:14) CarlosRuiz: technical "fights" :-) are mostly welcome
(11:09:30) trifon: :)
(11:09:42) CarlosRuiz: I don't see them as fights - is a dialectic process
(11:09:56) trifon: ok. i used wrong word.
(11:10:25) croo: A while back I wondered if we could convince the developers of the JEE frameworks themselves (like SUN or Spring) to help us move the framework forward ... they could use us to point out how great their framework is and we the get the benefit of their experience. ??
(11:10:46) trifon: Colin just one note.
(11:10:56) trifon: all this project do not know what is multi dimension.
(11:11:24) trifon: they think plain. Exmaple one class whihc must be persisted.
(11:11:35) trifon: they define in source code which column must be mandatory.
(11:11:46) croo: but I think the AD is a little of hinderance to n-tier... but it's so important nobody wants to touch it!
(11:11:53) trifon: while we allow user to modify AD and make column mandatory when he wants.
(11:12:35) croo: well we define it in code too .. we just generate it ... the same could happen in any otehr scenario
(11:13:01) trifon: other example. they can product XML from java class, whiule we allow ser to specify which colums to be in the xml file.
(11:13:03) croo: but the basis of the AD is the Table & Column ... I think it should be an object!
(11:13:07) CarlosRuiz: in fact, I don't generate X_ classes for my customizations
(11:13:26) CarlosRuiz: better explained
(11:13:38) CarlosRuiz: I generate X_ classes for new tables, but not for columns added to core tables
(11:13:42) trifon: Carlos this means that you do not use your customized classes from java?
(11:14:32) CarlosRuiz: yes I use them, but I don't call directly the setter/getter methods for columns added to core tables
(11:14:32) CarlosRuiz: I use setcustomcolumn and a general getter
(11:14:44) trifon: ye si got it.
(11:14:51) CarlosRuiz: that's trying to keep my migration effort at minimum
(11:15:04) trifon: yes this is good also.
(11:15:41) CarlosRuiz: I didn't get your point of "multi-dimension"
(11:16:01) trifon: ok. let me explain.
(11:16:46) trifon: when normal developer create a class which must be persisted he specify in source code that given field will be mandatory.
(11:17:28) trifon: what happens when develpoer has new client which ned column to be non mandatory.
(11:17:35) CarlosRuiz: ah, yes, similar to the "ref_list in code" problem pointed by Teo
(11:18:25) trifon: yes.
(11:18:26) CarlosRuiz: there are things that must not be in X_ classes, or they must be looking for AD
(11:18:50) trifon: Regarding AD. i have some ideas which i would like to share with you.
(11:19:01) croo: go for it
(11:19:23) trifon: i think that in AD_Table we can add new tabl where we can describe finder methods.
(11:19:53) trifon: we can modify GenrateModel to create proper finder mathods based on this definiotins.
(11:20:11) trifon: i will find sometime next weeks to try it.
(11:21:18) CarlosRuiz: with finder you mean methods to get records from db?
(11:21:23) trifon: yes.
(11:21:31) CarlosRuiz: yes, that's a need, good
(11:21:51) croo: would help remove sql from the java code
(11:22:05) trifon: to allow developer to call bPartner.findByAD_Client_IDAndValue or something like this.
(11:22:31) CarlosRuiz: going deeper, what's the meaning of X_ classes? speed of development?
(11:22:57) croo: i think jorg misundersood MVC!
(11:23:33) trifon: X_ should be the model.
(11:23:36) croo: but extend the X classes results in a monolytic object anyway ... so the layers cannot be seperated
(11:24:10) trifon: I have anither idea, but i will keep it till i commit the code which i made
(11:24:22) trifon: a small surpsise regarding X_ xlasses.
(11:24:23) CarlosRuiz: yes, Trifon
(11:24:29) CarlosRuiz: I think a "generic" finder can be done also
(11:24:35) CarlosRuiz: like in PO.java
(11:24:37) croo: moving Trifons idea forward I felt that we could use the IoC concepts in Spring to allow us to slowly move the persistent method from jdbc to anything else we want
(11:25:04) trifon: in fact i managed to imlpement it but no time to commit it from my lapto. :(
(11:25:37) trifon: Calon hard part is mbeded SQL. in fact dynamicaly constructed SQL is the hard part.
(11:25:41) trifon: Colin.
(11:25:48) trifon: sorry i made typo mistake.
(11:26:16) trifon: we have to mix JDBC code and hibernate or other framwerok.
(11:26:29) trifon: This is the place where i do not know what iwll happen in such mix.
(11:26:46) croo: yes, but mostky they are finders at present ... your suggestion would remove that
(11:26:48) trifon: but i think that this can be done, just do not have experince.
(11:26:57) trifon: yes for finders.
(11:26:59) croo: maybe not immediately .. but it coudl be an ongoing task
(11:27:39) croo: once the construct for finders are there we could slowly (say over a year) replace one with the other
(11:28:02) trifon: yes i agre. thi swill be ongoing task...
(11:28:18) Bahman: I and kontro are working on some solution to embedded sql.
(11:28:22) trifon: but developers will move to them as they will save them lines of code..
(11:28:35) trifon: privately?
(11:28:45) trifon: what solution?
(11:28:46) croo: what's the idea Bahman?
(11:29:26) trifon: in fact problem is dynamicaly genrated SQL. If SQL was static we can move it outside of the code.
(11:29:37) Bahman: Parsing the sql and extracting tables and columns and functions and dispathcing them to appropriate persistent class (most likely not X class!)
(11:30:51) trifon: interesting. do you have something working?
(11:31:36) Bahman: Not yet...just begun...but the idea is clear.
(11:31:42) trifon: i mean prototype or just a small example
(11:31:44) trifon: ok. good.
(11:33:17) CarlosRuiz: big effort Bahman, sounds like another Convert layer even more complex than current
(11:35:01) Bahman: It's planned to replace the current layer Carlos.
(11:35:14) croo: well the current PO is holing us back I think
(11:35:38) Bahman: Agree with you Colin.
(11:35:45) croo: there are db independant persistence technologies that we cannot use ...
(11:35:57) croo: because of it
(11:36:28) croo: and i don't have the answers!!!! in case you are going to ask :) but that doesn't mean I can#t see the problem!
(11:37:24) trifon: PO is itself persistet layer. probably we could make it DB indepndent... just s stupid idea...
(11:38:05) CarlosRuiz: yes, what finally we're coping with is a persistence layer invented by JJ
(11:38:19) croo: if Bahman's (& Timo's) "appropriate persistent class" is a javabean we could use Springs IoC to inject and ACTUAL persistence class we wanted
(11:38:49) CarlosRuiz: if that's our inheritance, maybe we could have a different approach
(11:39:07) CarlosRuiz: what's wrong with current persistence class? and look ways for improving it
(11:39:17) trifon: in fact we need to replcase SQL code with DB independent code like hibernate or other framwerks have own language.
(11:39:39) CarlosRuiz: what's the advantage of Hibernate vs PO.java ?
(11:39:45) trifon: current problem is DB independence, right?
(11:39:50) croo: it works with any DB!
(11:39:59) trifon: Db independece i think.
(11:40:00) croo: i expect better performance too
(11:40:07) CarlosRuiz: yes, at which cost? can you write complex statements?
(11:40:17) trifon: no. for performace it is not 100% sure.
(11:40:48) trifon: i do not know...
(11:40:57) croo: well I imagine that the people at hibernate knwo more about something like caching than we do!
(11:41:05) croo: what do you mean carlos?
(11:41:14) trifon: yes i agree with Colin. we must ask hibenrate poele.
(11:41:46) trifon: some framewroks allow SQL code to be read from outside. iBatis is such framweork.
(11:41:50) croo: well I didn't mean we need to ack them how to improve ours but that their solution is probably better since that is their speciality
(11:42:33) trifon: yes, but they have limitations too.
(11:42:59) croo: ideally I would think it better if we coudl concentrate on the business solution and not the framework.. but of course we must keep the benefit so of the AD we have!
(11:43:07) trifon: not all DB support given feature which means that their object query lanugage can't be as rich.
(11:43:56) CarlosRuiz: yes, I'm asking what would be the added value migrating to Hibernate
(11:43:56) CarlosRuiz: 1 - db independence
(11:43:56) CarlosRuiz: 2 - speed (not sure, maybe PO.java is faster)
(11:43:56) CarlosRuiz: 3 - maintenance of persistence layer outside of the project
(11:43:56) CarlosRuiz: ... what else?
(11:43:56) croo: which is why I like Spring!!! using it's IoC we can inject different actual DAO classes without touching our bsuiness classes
(11:44:25) trifon: that's it i think.
(11:44:55) trifon: well from business point of view this is external project and we cn blame it when we have prblems :)
(11:44:59) croo: well I do not see why we cannot use the AD to generate the persistence layer ... just not X_classes whcih are extend as that servcies no purpose
(11:45:10) croo: we cannot separate the layer if it';s simlpy extended
(11:46:29) CarlosRuiz: yes, I think X_ classes can be easily replaced, that's not the problem
(11:46:58) trifon: ok. let me tell youwha ti have made.
(11:47:03) CarlosRuiz: X_ classes is just a collection of a constructor and getter/setter for each column
(11:47:13) CarlosRuiz: that can be replaced easily for any other persistence layer
(11:47:23) trifon: i modified GenerateModel. it generates interface and implementation of this interface.
(11:47:45) CarlosRuiz: the real problem I think is the DB.executeUpdate .... executing SQL statements spreaded in all app
(11:48:00) trifon: yes. reall problem is SQL.
(11:48:15) croo: indeed ...
(11:48:45) CarlosRuiz: and there are really COMPLEX SQL's there
(11:48:50) croo: well I don't see a quick solution ... but I think we need one that allows us to slowly piece by piece move
(11:49:02) CarlosRuiz: and worst - DYNAMIC COMPLEX SQLs
(11:49:36) croo: yes the Dynamic SQL is quiet useful .. I imagine many people use it
(11:49:38) croo: a lot
(11:49:50) croo: so we need a way to achieve it without SQL ...
(11:50:03) croo: didn;t victir make a suggestion along those lines before?
(11:50:06) CarlosRuiz: that's why I asked if hibernate allowed complex SQL's
(11:50:37) croo: well I think the logic is moved out of the SQl into java!
(11:51:00) CarlosRuiz: one idea!
(11:51:15) CarlosRuiz: instead of constructing Convert_Postgresql we could make a Convert_Hibernate
(11:51:18) croo: or perhaps it sjould be pushed to the DB layer itself with stored prcedures?
(11:51:50) CarlosRuiz: the Convert is working well converting complex sql from oracle to postgresql
(11:52:01) CarlosRuiz: it could work converting to hibernate standarized language
(11:52:07) croo: yes its working and a credit to all involved ...
(11:52:11) trifon: interesting idea
(11:52:46) CarlosRuiz: so, it could be 2 steps - replacing X_ classes with another persistence engine (i.e. Hibernate)
(11:52:54) CarlosRuiz: like Trifon is pointing sounds good - via interfaces
(11:53:10) CarlosRuiz: with interfaces we could have plugabble persistence engine :-)
(11:53:29) trifon: i'm not sure Calros. SQL is the stopper.
(11:53:29) CarlosRuiz: and make experiments with hibernate without dropping completely PO.java at first
(11:53:53) CarlosRuiz: yes, second step could be construct a Convert_Hiberante for SQL statements
(11:54:42) Bahman: Sorry...I have to leave...will read the log later.
(11:54:47) CarlosRuiz: thanks Bahman
(11:54:53) croo: bey Bahman
(11:54:55) croo: bye
(11:55:56) croo: don't forget Trifons idea to have findbys in the AD... we could ermove much of the embedded SQL using that idea too
(11:55:58) trifon: bye
(11:57:28) trifon: i have ot do some housework.
(11:57:51) croo: ok bye ...
(11:58:33) CarlosRuiz: ok, bye, greeting to Victoria and Dete
(11:59:35) croo: well that 2 hours ... I guess we are all talked out :)
(11:59:58) croo: have you ever implemented a project that used the Project Management functioanlity Carlos?
(12:00:28) croo: just wondered what the quality was like
(12:00:42) CarlosRuiz: no, I'm not using it
(12:00:53) CarlosRuiz: just to reference projects in accounting
(12:01:56) jsSolutions: about persistence, anyone know the status of this:
(12:01:58) jsSolutions: http://java.sun.com/javaee/overview/faq/persistence.jsp
(12:02:00) croo: ok - i have a new prospect but they would realy heavily on it for job costing. in theory it looks like it has all the functionality needed. But I don;t think many use it
(12:02:32) jsSolutions: croo, project is about as complete as the rest of costing...
(12:03:15) croo: ok thx Joel
(12:04:06) jsSolutions: i know that's not much of an answer, but I think you know what it means in Compiere context :-o
(12:04:25) croo: yes I read between the ... line :)
(12:05:18) jsSolutions: i was just looking at the sun persistence stuff, because we might be able to get them to put a team on it
(12:05:23) croo: re: JPA - I think it is a popular solution now ...
(12:05:46) croo: yes I read that email .. could be very interesting ok ...
(12:05:47) trifon: yes JPA is good.
(12:06:21) croo: another reason why I like Spring .. would allow us to plugin JPA or Hinernate :)
(12:07:45) croo: the thing is - I don't the issue is finding a better persistence but how easy it is to convert ... and can we do it piecemeal.
(12:08:05) croo: by piecemeal I do not mean badly :) but that it would allow us to move slowly forward
(12:08:20) CarlosRuiz: Glassfish has JPA already implemented - according to the page
(12:08:38) CarlosRuiz: Colin
(12:09:04) CarlosRuiz: the idea from Trifon of constructing an interface layer for X_ classes - and maybe for PO.java can help us to achieve the piecemeal
(12:10:04) croo: yes. I remeber discussing this with Victor a long time ago and it turned out to be more problematic than first thought ... I can#'t remember what the issues were right now
(12:10:18) croo: but yes I open to what ever works :)
(12:11:26) croo: re: glassfish & JPA ... I haven't looked at App Servers recently but I thought they all supported JPA now!?
(12:16:05) CarlosRuiz: ok, thanks for the informal meeting :-) gtg
(12:16:39) croo: yeah me too ..
