David, et al,
I dont use opentaps because of the licensing but Si seems to be doing a pretty good job over there. I also try to follow both the opentaps and ofbiz forums and learn as much as I can. That said, there was this post from Si the other day and I just wonder what your opinion was. Here it is; "About three years ago ofbiz used a screen templating system called "jpublish", which they've subsequently replaced with a the ofbiz screen widget. During that time, jpublish has been advancing as well and is now getting ready to a new release. Florin, who is the lead developer of jpublish now, has published these documents about it: http://code.google.com/p/jpublish/wiki/WhyJPublish http://code.google.com/p/jpublish/wiki/JPublishUserGuide I'm considering whether we should go back to jpublish for the opentaps applications. While we (meaning Open Source Strategies) are very familiar with the screen widget now and have a lot of code built around it, I have a nagging feeling that it is a very counterintuitive tool and inhibiting other developers. I'd like to get some feedback from everybody about what they think of the ofbiz screen widget and whether jpublish might be a better long-term solution? " |
Why not JSF?
I think it's about the become the "standard" technology. Regards, Torsten Scott A schrieb: > David, et al, > > I dont use opentaps because of the licensing but Si seems to be doing a > pretty good job over there. I also try to follow both the opentaps and ofbiz > forums and learn as much as I can. > > That said, there was this post from Si the other day and I just wonder what > your opinion was. Here it is; > > "About three years ago ofbiz used a screen templating system called > "jpublish", which they've subsequently replaced with a the ofbiz screen > widget. During that time, jpublish has been advancing as well and is now > getting ready to a new release. > > Florin, who is the lead developer of jpublish now, has published these > documents about it: > http://code.google.com/p/jpublish/wiki/WhyJPublish > http://code.google.com/p/jpublish/wiki/JPublishUserGuide > > I'm considering whether we should go back to jpublish for the opentaps > applications. While we (meaning Open Source Strategies) are very familiar > with the screen widget now and have a lot of code built around it, I have a > nagging feeling that it is a very counterintuitive tool and inhibiting other > developers. I'd like to get some feedback from everybody about what they > think of the ofbiz screen widget and whether jpublish might be a better > long-term solution? " > |
In reply to this post by Scott.
Scott,
In my opinion it all boils down to personal preference. OFBiz still supports JPublish, so a user can utilize it if they wish. -Adrian Scott A wrote: > David, et al, > > I dont use opentaps because of the licensing but Si seems to be doing a > pretty good job over there. I also try to follow both the opentaps and ofbiz > forums and learn as much as I can. > > That said, there was this post from Si the other day and I just wonder what > your opinion was. Here it is; > > "About three years ago ofbiz used a screen templating system called > "jpublish", which they've subsequently replaced with a the ofbiz screen > widget. During that time, jpublish has been advancing as well and is now > getting ready to a new release. > > Florin, who is the lead developer of jpublish now, has published these > documents about it: > http://code.google.com/p/jpublish/wiki/WhyJPublish > http://code.google.com/p/jpublish/wiki/JPublishUserGuide > > I'm considering whether we should go back to jpublish for the opentaps > applications. While we (meaning Open Source Strategies) are very familiar > with the screen widget now and have a lot of code built around it, I have a > nagging feeling that it is a very counterintuitive tool and inhibiting other > developers. I'd like to get some feedback from everybody about what they > think of the ofbiz screen widget and whether jpublish might be a better > long-term solution? " > |
Actually, OFBiz doesn't support JPublish any more. We have library conflicts with it so to update other libraries it had to go. The newer version may fix this. Even still, I think it is more than just a matter of preference. There are significant developer efficiency, design flexibility, and code organization problems with JPublish that were solved with the Screen Widget. Our motive with the Screen Widget was to solve those problems and improve various things... not to just get rid of JPublish (there was no library conflict at the time, that was about 2 years later). The OFBiz framework is intended to be flexible to allow using technologies like this with other tools in OFBiz, but I would be really surprised to see any even minor advantages to changing the project best practices to recommend JPublish over the Screen Widget. Si can think and say what he wants and sometimes I agree, but not in this case, I'd have to say I 100% disagree with the notion. -David Adrian Crum wrote: > Scott, > > In my opinion it all boils down to personal preference. OFBiz still > supports JPublish, so a user can utilize it if they wish. > > -Adrian > > > Scott A wrote: >> David, et al, >> >> I dont use opentaps because of the licensing but Si seems to be doing a >> pretty good job over there. I also try to follow both the opentaps and >> ofbiz >> forums and learn as much as I can. >> >> That said, there was this post from Si the other day and I just wonder >> what >> your opinion was. Here it is; >> >> "About three years ago ofbiz used a screen templating system called >> "jpublish", which they've subsequently replaced with a the ofbiz screen >> widget. During that time, jpublish has been advancing as well and is now >> getting ready to a new release. >> Florin, who is the lead developer of jpublish now, has published these >> documents about it: http://code.google.com/p/jpublish/wiki/WhyJPublish >> http://code.google.com/p/jpublish/wiki/JPublishUserGuide >> I'm considering whether we should go back to jpublish for the opentaps >> applications. While we (meaning Open Source Strategies) are very familiar >> with the screen widget now and have a lot of code built around it, I >> have a >> nagging feeling that it is a very counterintuitive tool and inhibiting >> other >> developers. I'd like to get some feedback from everybody about what they >> think of the ofbiz screen widget and whether jpublish might be a better >> long-term solution? " >> |
David E Jones wrote:
> Actually, OFBiz doesn't support JPublish any more. We have library > conflicts with it so to update other libraries it had to go. The newer > version may fix this. I wasn't aware of that. I saw the JPublishViewHandler class and I assumed it worked. Thanks for the clarification. -Adrian |
In reply to this post by David E Jones
David E Jones wrote:
> Even still, I think it is more than just a matter of preference. There > are significant developer efficiency, design flexibility, and code > organization problems with JPublish that were solved with the Screen > Widget. Our motive with the Screen Widget was to solve those problems > and improve various things... not to just get rid of JPublish (there was > no library conflict at the time, that was about 2 years later). This point may be moot now, but at the time OFBiz was using JPublish I was able to solve many of those problems on my local copy. It turned out that JPublish wasn't configured correctly in OFBiz. By the time I submitted my fixes to Jira, the decision had already been made to switch over to widgets, so the fixes never made their way into the project. So, from my perspective the change from JPublish to screen widgets was imperceptible - because the "problems" the widgets solved didn't exist on my copy. -Adrian |
Adrian Crum wrote: > David E Jones wrote: >> Even still, I think it is more than just a matter of preference. There >> are significant developer efficiency, design flexibility, and code >> organization problems with JPublish that were solved with the Screen >> Widget. Our motive with the Screen Widget was to solve those problems >> and improve various things... not to just get rid of JPublish (there >> was no library conflict at the time, that was about 2 years later). > > This point may be moot now, but at the time OFBiz was using JPublish I > was able to solve many of those problems on my local copy. It turned out > that JPublish wasn't configured correctly in OFBiz. By the time I > submitted my fixes to Jira, the decision had already been made to switch > over to widgets, so the fixes never made their way into the project. > > So, from my perspective the change from JPublish to screen widgets was > imperceptible - because the "problems" the widgets solved didn't exist > on my copy. Could you be more specific about the "problems" you were working on? My reason for asking is probably fairly clear: there were lots of issues and limitations and things that weren't terribly efficient or flexible... the list goes into the dozens of items so I'm wondering which you found solutions to. -David |
David E Jones wrote:
> Adrian Crum wrote: > >> David E Jones wrote: >> >>> Even still, I think it is more than just a matter of preference. >>> There are significant developer efficiency, design flexibility, and >>> code organization problems with JPublish that were solved with the >>> Screen Widget. Our motive with the Screen Widget was to solve those >>> problems and improve various things... not to just get rid of >>> JPublish (there was no library conflict at the time, that was about 2 >>> years later). >> >> >> This point may be moot now, but at the time OFBiz was using JPublish I >> was able to solve many of those problems on my local copy. It turned >> out that JPublish wasn't configured correctly in OFBiz. By the time I >> submitted my fixes to Jira, the decision had already been made to >> switch over to widgets, so the fixes never made their way into the >> project. >> >> So, from my perspective the change from JPublish to screen widgets was >> imperceptible - because the "problems" the widgets solved didn't exist >> on my copy. > > > Could you be more specific about the "problems" you were working on? > > My reason for asking is probably fairly clear: there were lots of issues > and limitations and things that weren't terribly efficient or > flexible... the list goes into the dozens of items so I'm wondering > which you found solutions to. > > -David Oh geez, that was years ago... I vaguely remember the main complaint being you couldn't share scripts and templates between components. There were one or two more issues I can't recall - I just remember thinking to myself that those issues could be fixed by changing a few JPublish configuration xml files. With a little research and a little effort we were able to have global scripts and global templates. In addition, components shared bits of bsh and ftl, etc. We broke down common screen elements into reusable pieces and shared them between components. I'm sure you are aware of issues that I'm not aware of. Maybe this is a good time to go over them and see if they have been fixed. Not that I'm trying to push JPublish or anything, but since the subject has come up, why not look into it again? -Adrian |
Administrator
|
In reply to this post by David E Jones
David,
It would be very interesting if you could dress a list of all these problems or maybe the more importants. Then we (OFBiz developpers and users) would have solids arguments to choose between the future JPublish version and the widgets. BTW I like to work with the widgets but I don't know much about JPublish... Thanks Jacques PS : I may help to retrieve them from the old MLs and such if some pointers still exist... De : "David E Jones" <[hidden email]> > > > Adrian Crum wrote: > > David E Jones wrote: > >> Even still, I think it is more than just a matter of preference. There > >> are significant developer efficiency, design flexibility, and code > >> organization problems with JPublish that were solved with the Screen > >> Widget. Our motive with the Screen Widget was to solve those problems > >> and improve various things... not to just get rid of JPublish (there > >> was no library conflict at the time, that was about 2 years later). > > > > This point may be moot now, but at the time OFBiz was using JPublish I > > was able to solve many of those problems on my local copy. It turned out > > that JPublish wasn't configured correctly in OFBiz. By the time I > > submitted my fixes to Jira, the decision had already been made to switch > > over to widgets, so the fixes never made their way into the project. > > > > So, from my perspective the change from JPublish to screen widgets was > > imperceptible - because the "problems" the widgets solved didn't exist > > on my copy. > > Could you be more specific about the "problems" you were working on? > > My reason for asking is probably fairly clear: there were lots of issues and limitations and things that weren't terribly > > -David > |
Really, what's the point? A basic review of the example application and the framework introduction videos demonstrates the power of these things, and anyone who is really interested in becoming educated about it can study JPublish to see how things work with it. Anyway, this is really old stuff, and I don't think there is even the most remote chance of development in OFBiz being more efficient and flexible using JPublish instead of the Screen Widget. If someone wants to make an argument for that, then that's cool. Trying to catalog every reason why the screen widget was created would, IMO, just be a waste of time and I can think of approximately 1E6 better things to do with the few hours it would take. Of course, that's my opinion and again of others feel otherwise that time investment may be very valuable for them. -David Jacques Le Roux wrote: > David, > > It would be very interesting if you could dress a list of all these problems or maybe the more importants. Then we (OFBiz > developpers and users) would have solids arguments to choose between the future JPublish version and the widgets. BTW I like to work > with the widgets but I don't know much about JPublish... > > Thanks > > Jacques > > PS : I may help to retrieve them from the old MLs and such if some pointers still exist... > > De : "David E Jones" <[hidden email]> >> >> Adrian Crum wrote: >>> David E Jones wrote: >>>> Even still, I think it is more than just a matter of preference. There >>>> are significant developer efficiency, design flexibility, and code >>>> organization problems with JPublish that were solved with the Screen >>>> Widget. Our motive with the Screen Widget was to solve those problems >>>> and improve various things... not to just get rid of JPublish (there >>>> was no library conflict at the time, that was about 2 years later). >>> This point may be moot now, but at the time OFBiz was using JPublish I >>> was able to solve many of those problems on my local copy. It turned out >>> that JPublish wasn't configured correctly in OFBiz. By the time I >>> submitted my fixes to Jira, the decision had already been made to switch >>> over to widgets, so the fixes never made their way into the project. >>> >>> So, from my perspective the change from JPublish to screen widgets was >>> imperceptible - because the "problems" the widgets solved didn't exist >>> on my copy. >> Could you be more specific about the "problems" you were working on? >> >> My reason for asking is probably fairly clear: there were lots of issues and limitations and things that weren't terribly > efficient or flexible... the list goes into the dozens of items so I'm wondering which you found solutions to. >> -David >> > |
In reply to this post by Jacques Le Roux
We should really do the other way round: if there are issues in the
current widget implementation that we should discuss them in details and find the best way to fix them. I really think it is a waste of time to discuss these subjects for the sake of discussion... there are so many options and technologies available in the World and if we start a thread for each of them then we will completely stop the development of OFBiz, i.e. the development of an open source ERP system. Jacopo Jacques Le Roux wrote: > David, > > It would be very interesting if you could dress a list of all these problems or maybe the more importants. Then we (OFBiz > developpers and users) would have solids arguments to choose between the future JPublish version and the widgets. BTW I like to work > with the widgets but I don't know much about JPublish... > > Thanks > > Jacques > > PS : I may help to retrieve them from the old MLs and such if some pointers still exist... > > De : "David E Jones" <[hidden email]> >> >> Adrian Crum wrote: >>> David E Jones wrote: >>>> Even still, I think it is more than just a matter of preference. There >>>> are significant developer efficiency, design flexibility, and code >>>> organization problems with JPublish that were solved with the Screen >>>> Widget. Our motive with the Screen Widget was to solve those problems >>>> and improve various things... not to just get rid of JPublish (there >>>> was no library conflict at the time, that was about 2 years later). >>> This point may be moot now, but at the time OFBiz was using JPublish I >>> was able to solve many of those problems on my local copy. It turned out >>> that JPublish wasn't configured correctly in OFBiz. By the time I >>> submitted my fixes to Jira, the decision had already been made to switch >>> over to widgets, so the fixes never made their way into the project. >>> >>> So, from my perspective the change from JPublish to screen widgets was >>> imperceptible - because the "problems" the widgets solved didn't exist >>> on my copy. >> Could you be more specific about the "problems" you were working on? >> >> My reason for asking is probably fairly clear: there were lots of issues and limitations and things that weren't terribly > efficient or flexible... the list goes into the dozens of items so I'm wondering which you found solutions to. >> -David >> |
Administrator
|
David, Jacopo,
I agree, thanks for comments. I will have a look at JPublish when the new version (JP4) will be there. Jacques De : "Jacopo Cappellato" <[hidden email]> > We should really do the other way round: if there are issues in the > current widget implementation that we should discuss them in details and > find the best way to fix them. > I really think it is a waste of time to discuss these subjects for the > sake of discussion... there are so many options and technologies > available in the World and if we start a thread for each of them then we > will completely stop the development of OFBiz, i.e. the development of > an open source ERP system. > > Jacopo > > > Jacques Le Roux wrote: > > David, > > > > It would be very interesting if you could dress a list of all these problems or maybe the more importants. Then we (OFBiz > > developpers and users) would have solids arguments to choose between the future JPublish version and the widgets. BTW I like to > > with the widgets but I don't know much about JPublish... > > > > Thanks > > > > Jacques > > > > PS : I may help to retrieve them from the old MLs and such if some pointers still exist... > > > > De : "David E Jones" <[hidden email]> > >> > >> Adrian Crum wrote: > >>> David E Jones wrote: > >>>> Even still, I think it is more than just a matter of preference. There > >>>> are significant developer efficiency, design flexibility, and code > >>>> organization problems with JPublish that were solved with the Screen > >>>> Widget. Our motive with the Screen Widget was to solve those problems > >>>> and improve various things... not to just get rid of JPublish (there > >>>> was no library conflict at the time, that was about 2 years later). > >>> This point may be moot now, but at the time OFBiz was using JPublish I > >>> was able to solve many of those problems on my local copy. It turned out > >>> that JPublish wasn't configured correctly in OFBiz. By the time I > >>> submitted my fixes to Jira, the decision had already been made to switch > >>> over to widgets, so the fixes never made their way into the project. > >>> > >>> So, from my perspective the change from JPublish to screen widgets was > >>> imperceptible - because the "problems" the widgets solved didn't exist > >>> on my copy. > >> Could you be more specific about the "problems" you were working on? > >> > >> My reason for asking is probably fairly clear: there were lots of issues and limitations and things that weren't terribly > > efficient or flexible... the list goes into the dozens of items so I'm wondering which you found solutions to. > >> -David > >> > |
Free forum by Nabble | Edit this page |