nope really? yep how's it going d00d? ok flat out.. you? *** jearl has joined #gnuenterprise ok...ready for this wedding thing to be over with :) heh :) it can't be long now... *** jearl has quit IRC *** jearl has joined #gnuenterprise nope, not too long *** jearl has left #gnuenterprise *** dsmith has quit IRC *** johannesV has joined #gnuenterprise *** sjc has joined #gnuenterprise *** kilo has joined #gnuenterprise *** havoc has quit IRC *** kilo has quit IRC another sample: http://www.gnuenterprise.org/~johannes/sample.png (it's great to make use of the themes in forms :) hi johannesV hi SachaS johannesV any news on the list property ? no, not really, i've talked a bit with reinhard about this but we thought this would happen after the 0.1.0 release anyway i'm starting some work on 'appserver-generated forms' which means another xml-file (GNUe Layout Definitions [gld]) this file holds information about tab-order, translation, page-info and the like it get's integrated into the gnue_* classes and then a 'calculated field' called 'gnue_form' is added to 'gnue_class' class, returning a complete xml-code usable by forms if we then introduce another url-type like 'appserver://host/forms/classname' we can slightly change the forms-client to read a 'generated' forms-definition from appserver the same applys to something like 'appserver://host/reports/classname' which would give access to a generated report for that class does this go into 0.1.0 ? or also past 0.1.0 i think kilo was keen on having the list property for packages i'll start with that now, but i think it goes into 0.1 (which will happen in about two weeks, anyway) lots to do :) well, we just move on ... :) please dont forget about the list property :) no we don't, since this is a quite importaint one yeah i would think so. specially for someone who wants to write a gnue application and the .gcd file *** lekma has joined #gnuenterprise *** lekma has quit IRC *** lekma has joined #gnuenterprise *** dimas has joined #gnuenterprise *** kilo has joined #gnuenterprise morning hi kilo hi johannesV ok, i've just found out how to fix mouse-selections in gtk2 :() oh, that is great news *** lekma has quit IRC *** btami has joined #gnuenterprise johannesV: aftre you removed monospaced font support from gtk2 ui look at: http://www.gnuenterprise.org/~btami/gtk2-wx.png i think until forms has no layout based namspace only char based we need mono fonts just look at "Ember", "Becenév" labels i right aligned them next to corresponding entries it's OK in wx, but look ugly in gtk2 now... *** lekma has joined #gnuenterprise hi everybody hi lekma hi btami what is the gnue- script corresponding to gcdcvs? is it gcd2sql? yep hello all btami: do you working with .dbf files under linux? *** dimas has quit IRC *** dimas has joined #gnuenterprise btami: this is a change made by reinhard, which has fixed the size- and position of box-widgets right-alignment doesn't work in gtk2 alignment would be fixed automatically when layout-management works so i didn't put any effort into that; problem seems to be in gtk2 if a gtk.Label has a size_request () given, the positioning in it's fixed-container won't work anymore *** holycow has quit IRC i'll fix mouse-selection first and check the lables again, maybe i'll find something btami: can i have a copy of that gfd (from screen-dump); i'd like to check the coordinates dimas: no, i don't btami: would you please re-check your form after svn-update ? * btami svn up btami is very very anxious... about what ? http://www.gnuenterprise.org/~btami/ugyfel.gfd and http://www.gnuenterprise.org/~btami/ugyfel.gdb btami your .gdb link crashed firebird :( johannesV: i like to see fixedWidthFont/faceName config option support in gtk2 ui too, just for consistency silly question is someone testing python and gtk2 on windoze? SachaS: maybe http://www.gnuenterprise.org/~btami/ugyfel.gsd will be useful SachaS: yes, me it works well btami: gtk2 on windoze? yes cool. and there is gtk2 on mac as well, right? python and gtk2 on mac? don't know, but i think yes unfortunately we have no macs around here :( ok. is gtk2 on windoze realised through cygwin or native? native http://www.pcpm.ucl.ac.be/~gustin/win32_ports/pygtk.html btami: have you wrote this gsd manually or is it generated ? seems generated with designer... is gdb not working for you too? hi hi (-802, 'fetch.isc_dsql_fetch: arithmetic exception, numeric overflow, or string truncation. Cannot transliterate character between character sets. ') this error appears after login-handler but before the form appears use encoding=iso8859-2 in connections.conf and for textEncoding in gnue.conf too btami: the boxes in your gfd-file are overlapping the first one starts in row 2 and has 6 rows (-> from 2 to 7) the next box starts in row 7 and goes down to 12 and the third start in row 12 correct starting rows would be 2, 8, 14 i've made a simple gfd for checking boxes ... i'll upload in a sec ok johannesV: do you see anything that can cause problems if i call a .gcd module 'GNUE'? http://www.gnuenterprise.org/~johannes/box.gfd try this box.gfd with gtk2 and with wx :) so you'll see the difference kilo: in principle there's no problem in doing this, although i cannot see which advantages this will have (other than confusing people :) providing a commonly-useable typedef module in packages what kind of typedef ? just don't forget, my original complaint was about labels, not boxes :) btami: i've fixed that label-alignment already please svn-up and re-check ok or have i missed something? btami: using a fixed-width font won't solve that label-stuff like amount, measure, quantity... johannesV: about boxes, you find a bug in wx ui :) i wouldn't do that in gnue-module since this is somewhat reserverd for appserver why don't you introduce another module btami: yeah, i think so :) johannesV: about labels, in Char based forms only mono fonts can solve the label problem IMO why ? have you tried your form again ? yes *** reinhard has joined #gnuenterprise ah, now i see you haven't used the alignment-attribute of the label so there's the problem gotcha i would then prefer the following johannesV: so that one could use a GNUE_Amount as a property type this would work since the width=10 gives a max.widht measured in widget-width from which the really needed space gets substracted so just give all the right-aligned labes the same width and an apropriate alignment this way it doesn't matter wether you have a fixed-width font or not ! johannesV: cool, i'v never used "alignment" before i'v just forgot it s/it/it's exist :) ah, now i've got your form working for me (i'm using postgres now:) btw. you mentioned a new xml format for layout based forms (.gld) isn't namespace for differenk kind of forms, like Char, Layoutx, Foo ? like SimpleTabulation/Universal in reports we thought about another file-type for populating new system-tables with such information would contain the tab-order, the page a field property would be displayed on and a translated name for a given language since this information is all language-dependant we wouldn't store it in a gcd-file directly, but would add as lot of files needed (one per language) each property (appserver) could then define a 'label' in a given locale, a positin in the tab-order and so on (the latter could be helpful if a locale requires some other order for given fields (like zip)) in 'de' we would have country-code, zip in 'en' one would like to have zip, country-code and so on if no such 'gld' file is available, 'meaningfull' defaults will be used from existing 'property' data i think i don't understand the goal and the concept of gld the goal is to provide all information, so appserver can generate forms on the fly (based on classes) every class will have a calculated field called 'gnue_form' which holds a complete xml-code for piping into gnue-forms :) but to achieve this, appserver needs to know a bit more about the properties of it's classes why not just store gfd in the appserver db because i would have to write all gfd's manually -- even for each language i can imaging there would be 90% of all forms generated automatically, and about 10 % manually written the main reason is that we want those gfds to be dynamic depending on the installed modules on the permissions of the user and on the language exactly ! i don't think an aouto generated form can be good-looking enough to use why not we have all information to create widgets of all kind btami: just let us try ;-) checkboxes, dropdowns ... right if we fail, we haven't lost anything btami: it's a pity the very old mail archives aren't available any more but list properties would suffer time delay... 8-)) i've found a very old mail from jamest describing something very close to what we are doing now from 1999 when you produce a gfd, you use designer to put entries into position you like *** kilo has quit IRC you, the human decide what entry you like in the right side on the screen etc-etc. if you put these info into a new .gld why not use gfd with a new namespace instead of Char to describe the layout don't mix the gld with layout-management ah, so this is the poit point the gld just provides a means of a layout to create a gfd for appserver we don't want to touch forms at all, but to make appserver be able to create forms based on the hints given in a gld this is just for appserver so if you design a form, do it -- it would have nothing to do (well at least very little) with gld but not these hints are a partial replications of the info from a gfd? you would provide a gld (one per appserver-module) like you provide a gcd one gcd could have zero or more gld's (one per language) does a developer still need to write a .gfd if appserver has a .gld ? SachaS: no only if the form is really complicated or in other words the customer likes a special form more than the generated i think in most cases he/she would be fine with the dynamic form so the real work for defining gfd's will be very little (or at least it should) cant remember reinhard said he had like 350 forms. do you think those forms would look as good as they are if an application would create them? but not these hints are a partial replications of the info from a gfd? if yes, designer have to support for .gld next to .gfd btami: well, we have to think about this if we know it's working :) we have to think about a tool for creating gcd's anyway ... ok :) if not all guys write them in vi like me :) we thought that the dynamic-form stuff will save us from creating a lot of forms when we start writing apps * btami is using scite so that's why we'll do this before list-props ok. so a way to get forms quickly ... so you can test forms and your application. where the fine-tuned layout does not matter at all a form where it does not matter that the zip code and country come before street and name for example... well, we will see if 'fine-tuning' can be handled by appserver :)) SachaS: no, exactly these information is kept in a gld so the resulting form isn't just trash but also usefull ok. johannesV: is gld stores positional data (x/y) too ? something that gets you quickly a form like developers forms btami: no positional data will be calculated or (later on) handled by the layout-manager it will has tab-order, page (for notebook-widgets) and translation i think i get it. forms so you can continue with application development. if those forms are good enough for production (for real users) then good, otherwise you hand craft (or use designer) for forms. (at least in the beginning) SachaS: great, you got the point ok. cool. if i think of forms i think of those perfect forms for end users. but these would in the first step rather be application developer user forms ... hmm, then it will never use fe. boxes to group dtata, and "extra" labels etc btami: of what kind of boxes can you think ? ok is it in svn yet? just kidding. maybe we find a rule of thumb to build them box widgets SachaS: maybe at end of this week btami: think in terms of appserver-classes when would you need a box-widget ? dont forget forms are what users see. so if they look crap, they think the application is crap ... but i think i get the point hope so :) johannesV: i think i understand the concept, but why not simle use designer wizard instead is you want speed prototypes btami: we *really* want it dynamic btami: appserver generated forms are super-light-fast speed forms prototypes :) think of the customer form there will be a base customer module with address and such there might be a shipment module with some extra data for the customer table there will be most probably a european accounting module that adds the VAT-ID to the customer table there might be a birthday tracking module that adds the birthday to the customer table (all this being silly examples of course) now will each of these modules have its own customer form? we want appserver to go through all installed modules look who's having fields in the customer table and dynamically build a form with all those fields that's the idea reinhard: i understand this, just i think creating forms wint any AI will never can get the quality of a human created form using some designer tool except wery simple forms *** jamest has joined #gnuenterprise but i'm not against your ideas, yust arguing a bit, to understand what you want to do :) the time will tell if it will be useful or not... yes i have absolutely no other idea on how to handle this other than creating 6 forms customer form with base and shipment data customer form with base and vat-id data customer form with base, shipment and vat-id data etc. not to start when we want permissions i.e. somebody may not see the total due amount of the customer others want to see it so will this field be still visible and just show 0 (Which might be quite confusing) yes, i see your poins or will we have all forms in a with-due-amount and a without-due-amount version ok for some forms, we will have no other *chance* than living with the generated version reinhard: thx for sharing your thoughts, bbl *** btami has quit IRC *** johannesV_ has joined #gnuenterprise maybe we could add something like 'context' to the gld, where all elements with the same context get packed into a box *** kilo has joined #gnuenterprise *** kilo has joined #gnuenterprise *** johannesV has quit IRC *** bluesbaron_ has joined #gnuenterprise *** deke has quit IRC *** deke has joined #gnuenterprise yo *** bluesbaron_ has left #gnuenterprise *** jcater has joined #gnuenterprise hi jcater morning whaddup d00dz *** apropos has joined #gnuenterprise has there been any talk of some sort of BPM tool being part of the gnue tools suite? what is BPM ? business process management - process modelling big purple monster ah, thanks chillywilly: lol chillywilly: probably not too far off :) http://www.savvion.com/products/bpm_server.php sort of, kind of like that *** btami has joined #gnuenterprise ah well hi all I think I shall go home & sleep.. night all :) reinhard: i was thinking a bit while driving home, and i realized i was using something very similar to gld concept in my good'ol dos app i have a meta database like dbf_tables,dbf_fields etc. and in dbf_fields i have fields for labels and order for presenting grid/browse views dynamically (labels used for column headers) gld? :) jcater, please read the log GNUe Label Definitino erm, Definition is this for i18n of forms/reports/etc ? btami: but I'm lazy jcater: it's for appserver so appserver can generate gfd's on the fly interesting approach i'm curious how this gfd will look like after i'm finished :) jcater: usually i say "interesting approach" if i want to express "you're crazy" in a polite way :) apropos: I am interested in BPM :) reinhard: interesting thought * jcater ducks reinhard: not really... I'm not involved in appserver enough to have an opinion either way if this is what you guys think you need, I'm just glad you're reusing forms :) apropos: would be cool if gnue supports that ... SachaS: what aren't you interested in? *** holycow has joined #gnuenterprise chillywilly: windows operating system you like da winblows just admit it better'n linsux, boyo chillywilly: he didnt say he didnt like it.. he just said he wasnt interested in it SachaS: I just wonder if having some sort of process modelling tool would open up development more to domain experts there are up sides adn downsides to that I can see the up side, but what would the down side(s) be? derek: you're the last person who should try and get technical ;) anyhooters... apropos if you look at gnue appserver, then you will find business classes. business class have procedures. apropos so you can write business logic with those procedures. perl is the devil actually you can put business logic into a python library and "just" import that python library in any of those procedures. but you have business classes with procedures. it was birthed in the fires of hell it is the spawn of the Dark Lord Satan so far so good i do not have the experience of a business application writer but i think at one point your business logic will be distributed/burried/stored away in those business class procedures so only a developer will know what the hell is going on and that my friend is what we call "job security", j/k ;) :) this perl code I am reading is definitely written with this principle in mind :((( at that stage bpm will become interesting. i am still strugling with the following scenario: apropos the downside is domain experts means they are not computer people lets assume there exists a gnue-accounting so some things become bad decisions because they do not utilize computing properly for example and lets assume there exists a gnue-accounts-receivable module if i am hotel management expert but dont know dick about computers and write a hotel management system but completely ignore the fact there is this thing called the internet and that perhaps i could use that to aid in reservation booking i miss a pretty large opportunity then I am not sure if there needs to be a sales module i.e. domain experts lack technology as an expertise there is an inventory module as well therefore they generally are not going to see opportunities to utilize technology to better what they do derek: but shouldn't there be some way for the subject matter experts and the developers to work together? How many programmers can do depreciation schedules? yes so there is business process which will model a sale for example my point is BPM is pushing for the "no technology experitise required" that business process will use business classes from various gnue packages and my point was there are ups and downs to that :) derek: how about a collaborative BPM design system, kinda like bugzilla for accountants? :-) the upside is it allows business people to get what they want quicckly without fighting an engineers :) that business process will use business classes and specially their procedures from various gnue packages the downside is if they push the engineers out they do themselves an innovative disservice :) i.e. BPM is not a panacea to the problem so there might be a business process packages (i dont know if that would be the right way to do it). silly bean counters... the problem is engineers dont listen to their customers enough :) that business process package could have one class called business process execution engine bpm doesn't solve any problems for computer engineers, but it would solve a lot for business users... which is the point of an ERP system isn't it? and that business class could execute business processes as described before and run the described business process and calling various business classes (objects at run time) and their procedures. I had a convo the other day with a phb and he was like "we can't implement so and so's vision of a web base system" then he goes on "but we need something where all the branches applications come back and talk to some server here" . * derek runs off to a meeting well... I have no problem with modelling business procedures in python - I love python - but how do you get the needs of the business from the domain experts then? for instance, SachaS' sales process. There are a *lot* of different types of sale... from simple POS transactions to long drawn out contract negotiations to open purchase orders. apropos: there are some open source business process engines out there, including even if rudimentary guis to model the processes. apropos ok so there are lots of different ones. so how do you handle that? do you model each one? that's what I'm getting at. Harness the power of the FOSS development model and create some collaborative method. if there were a simple way for an accountant to describe (and maybe even model a bit) obscure processes, then a programmer could translate that. true. so you say model those lots of sale business processes and share the models or 100 accountants make them do UML models ;) hahaha yeah, just an idea I had in class the other night that'll happen.... there's no way in hell they would get it :) yeah, UML wouldn't work. I haven't figured out how it would work. *** btami_kitchen has quit IRC i support your idea apropos. lots is not infinitely if you could get a large enough population of accountants, operations managers, MBAs etc to model business processes then you could boil it down to some OO hierarchy. theoretically, at least. :-/ apropos what do you think, how many sales business processes are out there? how about use cases or "scenarios" but those still need to be transalated into software artifacts and they are not very useful for consumption by some software engine 100? 1000? more? SachaS: there are effectively an infinite number, but you should be able to come up with 5-10 models that cover 99% of situations. yeah and each sale business process interacts with an erp system (like gnue) I hate perl right, and if it is modelled in the ERP system then the actual tasks done by people could be made more efficient with help from the ERP system (which again is the point) chillywilly: I hate perl too * chillywilly kicks this perl code and if a sale business process needs "book customer(x) - stock(y) 100" in whatever state of the business process it can access the gnue-accounting package and run that action if there were a few "stock" sales process classes available, and some simple way to extend them (RAD/GUI?) then the system could be customized to each individual situation. with a few = 30 default gnue sales business processes, a company could chose which one it supports and just use that one. yes ok have to leave unfortunately :) later, nice chatting with you chillywilly: if there were 1) some intermediate format for business modelling that could be displayed and modified (and revision tracked) on a website and 2) automatically or semi-automatically translated into a business object then you could do collaborative development of the models. apropos have a look at: http://www.newsforge.com/software/03/09/08/162259.shtml?tid=11 isn't there some XML markup? Business Process Language....something like that there are several business process modelling "languages", I don't know a whole lot about them (yet) BPML SachaDinner: thanks, that is very cool. Some kind of data integration tool. http://www.bpmi.org/ http://tmitwww.tm.tue.nl/research/patterns/ if there were a simple way to model these patterns along with any data attributes, that might be usable apparently UML isn't the way to go either http://tmitwww.tm.tue.nl/research/patterns/standards.htm *** sjc has quit IRC l8r all *** reinhard has quit IRC *** btami_kitchen has joined #gnuenterprise *** johannesV_ has quit IRC *** holycow has quit IRC *** johannesV has joined #gnuenterprise *** dimas has quit IRC *** btami has quit IRC *** johannesV has quit IRC so is GNUe pure Python? pure as the driven snow suh weet can there be multiple appservers for a single db? don't know *** btami has joined #gnuenterprise apropos your welcomed apropos I can recommend the book Workflow Management by Aalst and van Hee Aalst is the professor at the university of that workflow pattern site. ah, ok thanks and the www.jbpm.org is pretty cool too. I think they are also based on studies from van der Aalst van der Aalst based his workflow management theory on petri nets I would think we could use a business process execution engine with gnue appserver today currently they would be two applications as opposed to having gnue appserver providing such a service reason: gnue appserver provides a xml-rpc interface to access the business classes. so i would think it would be possible already today. which is good news :) it is actually a matter of testing.... when you say "business process execution engine", what does that mean? so we would have the business execution engine executing one of those 30 sales business processes and kicking gnue-appserver to do certain activites ok what I mean is the follwoing what you see on the website of http://www.savvion.com/products/bpm_server.php ok is a business process modeller so the business analyst can model a business process basically this would be a tool where an analyst (or subject matter expert) models a business procedure - some done by people and some done by a computer, right? using if then else, join, forks etc correct modelled by a subject expert just the diagram the raw business model so the computer would have triggers that drive portions of the workflow (scripts even) and then they go back to a waiting state (waiting on a person) or a finished state. hold on so you get that business model that business diagram yes and the tool helps you to make sure that the model follows some rules... the idea is that you save the model as lets say an xml file and it would have to describe what data goes where as well, I think yes. ok, I'm with you the tricky part in the modelling process would be to know what systems are available. like that you have a accounting system that allows you to do a, b, and c. but thats a detail. with gnue appserver introspection something could be possible ... anyway I would assume that the modelling tool would be able to discover the API, yes. you get a sales business process yes something like that *** lekma has quit IRC in the www.jbpm.org you have to code that ... with java ... you can connect to a system. then you have that file and then you have a business process execution engine which can read the file and execute it so you can install as many business process files as you want each will be some sort of a business process instance executed by the business process execution engine well, I don't know if a full-fledged GUI tool is really necessary. It could look like a treeview, even. you can do it with vi, just write a business process file.... well, or a web page was what I was thinking but business analyst or subject experts probably prefer a gui. what would be really cool were if it were a two-way tool... could update a live codebase. java applet or something. they prefer, but as long as they aren't coding it should be ok yeah i think so too so if in the business process it says connect to gnue apperserver and call procedure x of business class y, the execution engine could do that because appserver provides an xml-rpc interface. it is realy a matter of trying it out.... essentially, but the key that I saw was the ability to collapse the BPMs into workflows that consist of triggers->events there might be an if then else in the business process checking on the result value of the business class procedure. it could do a lot, maybe I need to read that book (workflow patterns) *** lekma has joined #gnuenterprise there is an introduction chapter that gets you the terminology right (i forgot them already ;) so people should understand when you talk about trigger / event / action / activity etc... maybe I just need to read up on bpml anyway. i would be interested in having this functionality somehow in gnue. and i would think it is already possible combining two tools (business process application with gnue appserver) the crux is there is not one way to do what we are talking about: *** sjc has joined #gnuenterprise petri nets, pi calculus, business process markup language (bpml), web service business process exeuction language (ws-bpel), ebXML business process (ebBP), web service composite application framework (ws-caf), web and more yeah some are used in different ways maybe just doing it python is the right way. lol it does not matter in which programming language the engine is written in :) I think *that* is a religious issue. :) yeah i know. java might not be ok even if the tool is gpl etc.... lots of free java implimentations so how is the business logic for the erp side being designed? is there a list or something? currently if you are writing a n-tier gnue application you have to write a gnue class definition file (.gcd). there you define a module, its business classes, attributes and methods of classes. actually they are called procedures not methods. so the business logic is in the business class procedures. I found a gnue-dev list, I see a post by you on b2b stuff. Are you big into b2b? then you add a couple of user forms definition files (.gfd) and you are done :) dont get me started :) heh where are you from? that's what I do for a living tulsa, ok, us i am from liechtenstein ok too far we have to discuss it here then ;) heh discussion over espresso would be nice :-) i almost finished a masters in computer science looking at a topic of ebXML. ebXML stands for electronic business and is a framework for business to business. I'm almost finished with an MBA in technology management ebXML stands for electronic business XML. cool. I use EDI every day cool. i read about EDI but never used it. ebxml is a replacement for the ANSI X.12 EDI standard, isn't it? some call it the next generation of EDI .... some :) but EDI will be there for another long time i would think the web site is a bit slow but new b2b solutions might look at ebXML which website? yes, Wal-Mart (largest retailer here in US) uses it extensively (ebxml.org) oh ok. so what was your topic in technology management? most large retailers in US use ANSI X.12 EDI standard we don't have one for the MBA program (I assume you mean a master's thesis) yes. i did it in australia. there is a masters by coursework (u take units and exams) or masters by research... if I did it would probably be on workflow ok. today I found out that my research project is listed on ebxml.org! excellent... if I can ever get there I'll check it out. looks like there's an openebxml (and an #openebxml chan on this network) thats me :) *** btami has quit IRC anybody running gentoo around and willing to test common and appserver ebuilds i've put together? apropos have you looked closer at a workflow management system? *** jamest has left #gnuenterprise they are available at http://www.orwex.net/gnue-dev/dev-gnue_overlay_20040728.tar.bz2 *** whizman has joined #gnuenterprise Greetings all - last week (2004-07-20) I dropped in asking about format masks ... I would like to follow up to that discussion as soon as convenient. night all *** lekma has left #gnuenterprise lekma: will try to test them in the weekend *** kilo has quit IRC night all (will try coming back later) *** whizman has left #gnuenterprise *** whizman has joined #gnuenterprise *** sjc has quit IRC *** whizman has left #gnuenterprise *** holycow has joined #gnuenterprise *** jcater has quit IRC *** holycow has quit IRC *** jcater has joined #gnuenterprise *** crazymike has joined #gnuenterprise *** crazymike has quit IRC *** havoc has joined #gnuenterprise *** gsoti_away has joined #gnuenterprise *** jcater has quit IRC