Dev - Marketing Ofbiz move to apache

[cjhowe]

Can we start brainstoming some marketing ideas
coordinated with with OFBiz's exit from the incubator
and into the ASF?  Some of them might require a
community effort to get noticed.

1. Submit a story for digg.com and have a community
push to get it to the front page...and beyond
2. Release of blogs, reviews, editorials simultaneous
to the acceptance (might want to exchange writeups for
fact checking, grammar, etc beforehand.)
3.  Any other ideas?

 
_______________________________________________
Dev mailing list
[hidden email]
http://lists.ofbiz.org/mailman/listinfo/dev

[J Aaron Farr]

I'll reply to this more tomorrow (it's late here right now), but a
good idea would be to work with Apache's PR team.  There is some
sensitivity to the type of marketing that goes on with Incubated
projects since they aren't yet full-fledged Apache projects.  There's
been some issues with some companies in the past in this respect.
However, we do want to get the word out, just as long as we keep the
ASF in the loop.

I'm not sure if the Apache PR mailing list is publicly available, so
let me check on that first.

On 3/1/06, Chris Howe <[hidden email]> wrote:

> Can we start brainstoming some marketing ideas
> coordinated with with OFBiz's exit from the incubator
> and into the ASF?  Some of them might require a
> community effort to get noticed.
>
> 1. Submit a story for digg.com and have a community
> push to get it to the front page...and beyond
> 2. Release of blogs, reviews, editorials simultaneous
> to the acceptance (might want to exchange writeups for
> fact checking, grammar, etc beforehand.)
> 3.  Any other ideas?
>
>
> _______________________________________________
> Dev mailing list
> [hidden email]
> http://lists.ofbiz.org/mailman/listinfo/dev
>

--
  jaaron
 
_______________________________________________
Dev mailing list
[hidden email]
http://lists.ofbiz.org/mailman/listinfo/dev

[cjhowe]

What I was thinking was a coordinated effort for if
and when it becomes a full-fledged Apache project.

--- J Aaron Farr <[hidden email]> wrote:

> I'll reply to this more tomorrow (it's late here
> right now), but a
> good idea would be to work with Apache's PR team.
> There is some
> sensitivity to the type of marketing that goes on
> with Incubated
> projects since they aren't yet full-fledged Apache
> projects.  There's
> been some issues with some companies in the past in
> this respect.
> However, we do want to get the word out, just as
> long as we keep the
> ASF in the loop.
>
> I'm not sure if the Apache PR mailing list is
> publicly available, so
> let me check on that first.
>
> On 3/1/06, Chris Howe <[hidden email]> wrote:
> > Can we start brainstoming some marketing ideas
> > coordinated with with OFBiz's exit from the
> incubator
> > and into the ASF?  Some of them might require a
> > community effort to get noticed.
> >
> > 1. Submit a story for digg.com and have a
> community
> > push to get it to the front page...and beyond
> > 2. Release of blogs, reviews, editorials
> simultaneous
> > to the acceptance (might want to exchange writeups
> for
> > fact checking, grammar, etc beforehand.)
> > 3.  Any other ideas?
> >
> >
> > _______________________________________________
> > Dev mailing list
> > [hidden email]
> > http://lists.ofbiz.org/mailman/listinfo/dev
> >
>
>
> --
>   jaaron
>  
> _______________________________________________
> Dev mailing list
> [hidden email]
> http://lists.ofbiz.org/mailman/listinfo/dev
>

 
_______________________________________________
Dev mailing list
[hidden email]
http://lists.ofbiz.org/mailman/listinfo/dev

[davidnwelton]

> What I was thinking was a coordinated effort for if
> and when it becomes a full-fledged Apache project.

Let's make that happen first, but yes, there is a PR group at Apache
(including one person who is actually a PR person, rather than a
programmer interested in PR:-), and we can certainly talk with them.
I think David Jones is a bit worried about not being too visible too
early, but I don't think it would be mistaken to make a little bit of
noise if there is a release of some kind ready at that point.

As always, though, your call - and I'll help out as I'm able.

--
David N. Welton
 - http://www.dedasys.com/davidw/

Linux, Open Source Consulting
 - http://www.dedasys.com/
 
_______________________________________________
Dev mailing list
[hidden email]
http://lists.ofbiz.org/mailman/listinfo/dev

[David E. Jones]

I don't mind publicity, but I do mind setting expectations that are  
eventually not met or that are delayed... in this project as well as  
in general business... nothing eats at me more.

With the Apache effort there is really a LOT to do. The worst thing  
is that there are lots of little things because it turns out that the  
few big things are rather complex... In other words it could really  
take a long time, hopefully not too long of course, but on the order  
of various months and not various weeks. It could happen that fast,  
but I don't see it.

So yeah, once we're ready to leap I'd like to put together all the  
materials we can to not only advertise what exists in OFBiz, but also  
put together materials to help people "get started" with it,  
especially get an idea of what the framework is like and what it is  
like to work with it. The point would be to resolve the technical  
concerns in addition to the business functionality, and perhaps even  
help push the technical/framework features as a primary benefit of  
the project to a crowd that is often very strictly into OO stuff and  
such...

Anyway, it's probably fine to start this conversation, but everyone  
should keep in mind that this may be a ways off...

-David


On Mar 1, 2006, at 12:37 AM, David Welton wrote:

>> What I was thinking was a coordinated effort for if
>> and when it becomes a full-fledged Apache project.
>
> Let's make that happen first, but yes, there is a PR group at Apache
> (including one person who is actually a PR person, rather than a
> programmer interested in PR:-), and we can certainly talk with them.
> I think David Jones is a bit worried about not being too visible too
> early, but I don't think it would be mistaken to make a little bit of
> noise if there is a release of some kind ready at that point.
>
> As always, though, your call - and I'll help out as I'm able.
>
> --
> David N. Welton
>  - http://www.dedasys.com/davidw/
>
> Linux, Open Source Consulting
>  - http://www.dedasys.com/
>
> _______________________________________________
> Dev mailing list
> [hidden email]
> http://lists.ofbiz.org/mailman/listinfo/dev

 
_______________________________________________
Dev mailing list
[hidden email]
http://lists.ofbiz.org/mailman/listinfo/dev

[cjhowe]

All of this (copy, cordination, tutorials) will take
about as much time to put together WELL as getting the
project through the incubator process (especially
since this would all be volunteer work).  By reading
through the list, there are probably a lot more
business people here interested in programming as
opposed to programmers interested in business.
Business people should really be the target of the
transition (especially for you consulting people/
companies).  

I'm willing to contribute any of my skills and free
time (relative word) to the marketing effort because
the success of the consulting companies is what is
going to drive the growth of this project over the
months following entering the
ASF<obligitory_tagline>"if it gets
accepted"</obligatory_tagline>. Their success means I
get a better product, more functionality, etc.  

I'm sure many of you know the reaction of decision
makers when they learn that the Apache web server
hosts 75%+ of the websites on the internet.  It was
essentially the deciding factor on our company
choosing open source software in the first place.  The
future association of OFBiz with the ASF is a really
unique opportunity that needs to be fully capitalized.
 And again, to do that WELL, we need to plan what
needs to be done by when and that will take time.

--- "David E. Jones" <[hidden email]> wrote:

>
> I don't mind publicity, but I do mind setting
> expectations that are  
> eventually not met or that are delayed... in this
> project as well as  
> in general business... nothing eats at me more.
>
> With the Apache effort there is really a LOT to do.
> The worst thing  
> is that there are lots of little things because it
> turns out that the  
> few big things are rather complex... In other words
> it could really  
> take a long time, hopefully not too long of course,
> but on the order  
> of various months and not various weeks. It could
> happen that fast,  
> but I don't see it.
>
> So yeah, once we're ready to leap I'd like to put
> together all the  
> materials we can to not only advertise what exists
> in OFBiz, but also  
> put together materials to help people "get started"
> with it,  
> especially get an idea of what the framework is like
> and what it is  
> like to work with it. The point would be to resolve
> the technical  
> concerns in addition to the business functionality,
> and perhaps even  
> help push the technical/framework features as a
> primary benefit of  
> the project to a crowd that is often very strictly
> into OO stuff and  
> such...
>
> Anyway, it's probably fine to start this
> conversation, but everyone  
> should keep in mind that this may be a ways off...
>
> -David
>
>
> On Mar 1, 2006, at 12:37 AM, David Welton wrote:
>
> >> What I was thinking was a coordinated effort for
> if
> >> and when it becomes a full-fledged Apache
> project.
> >
> > Let's make that happen first, but yes, there is a
> PR group at Apache
> > (including one person who is actually a PR person,
> rather than a
> > programmer interested in PR:-), and we can
> certainly talk with them.
> > I think David Jones is a bit worried about not
> being too visible too
> > early, but I don't think it would be mistaken to
> make a little bit of
> > noise if there is a release of some kind ready at
> that point.
> >
> > As always, though, your call - and I'll help out
> as I'm able.
> >
> > --
> > David N. Welton
> >  - http://www.dedasys.com/davidw/
> >
> > Linux, Open Source Consulting
> >  - http://www.dedasys.com/
> >
> > _______________________________________________
> > Dev mailing list
> > [hidden email]
> > http://lists.ofbiz.org/mailman/listinfo/dev
>
> >  
> _______________________________________________
> Dev mailing list
> [hidden email]
> http://lists.ofbiz.org/mailman/listinfo/dev

 
_______________________________________________
Dev mailing list
[hidden email]
http://lists.ofbiz.org/mailman/listinfo/dev

[davidnwelton]

On 3/1/06, Chris Howe <[hidden email]> wrote:
> All of this (copy, cordination, tutorials) will take
> about as much time to put together WELL as getting the
> project through the incubator process

If you'd like, I can try and put you and other interested people in
contact with the PRC (public relations committee) - although I'm not
sure they accept subscribers who aren't ASF members, you could at
least begin a conversation.

Let me know, on the list or in private.  Any decisions will of course
have to be cleared with what will be the PRC for the OFBiz project
(David, Andy, et al.).

--
David N. Welton
 - http://www.dedasys.com/davidw/

Linux, Open Source Consulting
 - http://www.dedasys.com/
 
_______________________________________________
Dev mailing list
[hidden email]
http://lists.ofbiz.org/mailman/listinfo/dev

[davidnwelton]

> like to work with it. The point would be to resolve the technical
> concerns in addition to the business functionality, and perhaps even
> help push the technical/framework features as a primary benefit of
> the project to a crowd that is often very strictly into OO stuff and
> such...

Speaking for myself, what drew me to OFBiz was really its capabilities
("this does most of what we need, out of the box!") and community
("cool, these guys understand the problems I'm trying to solve!"),
rather than the underlying technology, which I came to appreciate when
looking a little bit more in depth.

My guess is that that's always going to be the up front selling point
- the Java world is full of frameworks - this one's just lucky to be
associated with a potentially 'killer app'.

> Anyway, it's probably fine to start this conversation, but everyone
> should keep in mind that this may be a ways off...

Right.  By the way, I opened up an issue for SVN migration:

http://issues.apache.org/jira/browse/INFRA-741

--
David N. Welton
 - http://www.dedasys.com/davidw/

Linux, Open Source Consulting
 - http://www.dedasys.com/
 
_______________________________________________
Dev mailing list
[hidden email]
http://lists.ofbiz.org/mailman/listinfo/dev

[David E. Jones]

Chris,

I think you're right that a lot of this could be done in parallel  
with the IP clearance, code changes, infrastructure stuff and so on  
that make up the pre-requisites of the ASF move.

Unless someone knows of a better tool or a better way of doing  
this... perhaps a series of wiki pages would be a good way to  
organize and gather contributions in this area.

I guess the first step is to establish some priorities or some types  
of content we'd like to have for marketing materials. There are some  
things along these lines in the wiki, and on the ofbiz.org web site  
there are a few things that it would be nice to update and flesh out,  
like:

http://www.ofbiz.org/general-overview.html
http://www.ofbiz.org/ofbiz-for-me.html
http://www.ofbiz.org/feature-list.html
https://ofbiz.dev.java.net/servlets/ProjectDocumentList?folderID=1193

These and other things are in the "What is OFBiz?" section in the top  
of the left side-bar and in the first main column box.

Anyway, there are a number of messages there we are trying to  
communicate to prospective (and I guess existing) users. Additional  
messages could certainly be added, and hopefully these will be a good  
starting point for more refined marketing collateral including  
brochures, press releases, and so on, and perhaps even various sets  
of these targeted at different audiences.

-David


On Mar 1, 2006, at 1:34 AM, Chris Howe wrote:

> All of this (copy, cordination, tutorials) will take
> about as much time to put together WELL as getting the
> project through the incubator process (especially
> since this would all be volunteer work).  By reading
> through the list, there are probably a lot more
> business people here interested in programming as
> opposed to programmers interested in business.
> Business people should really be the target of the
> transition (especially for you consulting people/
> companies).
>
> I'm willing to contribute any of my skills and free
> time (relative word) to the marketing effort because
> the success of the consulting companies is what is
> going to drive the growth of this project over the
> months following entering the
> ASF<obligitory_tagline>"if it gets
> accepted"</obligatory_tagline>. Their success means I
> get a better product, more functionality, etc.
>
> I'm sure many of you know the reaction of decision
> makers when they learn that the Apache web server
> hosts 75%+ of the websites on the internet.  It was
> essentially the deciding factor on our company
> choosing open source software in the first place.  The
> future association of OFBiz with the ASF is a really
> unique opportunity that needs to be fully capitalized.
>  And again, to do that WELL, we need to plan what
> needs to be done by when and that will take time.
>
> --- "David E. Jones" <[hidden email]> wrote:
>
>>
>> I don't mind publicity, but I do mind setting
>> expectations that are
>> eventually not met or that are delayed... in this
>> project as well as
>> in general business... nothing eats at me more.
>>
>> With the Apache effort there is really a LOT to do.
>> The worst thing
>> is that there are lots of little things because it
>> turns out that the
>> few big things are rather complex... In other words
>> it could really
>> take a long time, hopefully not too long of course,
>> but on the order
>> of various months and not various weeks. It could
>> happen that fast,
>> but I don't see it.
>>
>> So yeah, once we're ready to leap I'd like to put
>> together all the
>> materials we can to not only advertise what exists
>> in OFBiz, but also
>> put together materials to help people "get started"
>> with it,
>> especially get an idea of what the framework is like
>> and what it is
>> like to work with it. The point would be to resolve
>> the technical
>> concerns in addition to the business functionality,
>> and perhaps even
>> help push the technical/framework features as a
>> primary benefit of
>> the project to a crowd that is often very strictly
>> into OO stuff and
>> such...
>>
>> Anyway, it's probably fine to start this
>> conversation, but everyone
>> should keep in mind that this may be a ways off...
>>
>> -David
>>
>>
>> On Mar 1, 2006, at 12:37 AM, David Welton wrote:
>>
>>>> What I was thinking was a coordinated effort for
>> if
>>>> and when it becomes a full-fledged Apache
>> project.
>>>
>>> Let's make that happen first, but yes, there is a
>> PR group at Apache
>>> (including one person who is actually a PR person,
>> rather than a
>>> programmer interested in PR:-), and we can
>> certainly talk with them.
>>> I think David Jones is a bit worried about not
>> being too visible too
>>> early, but I don't think it would be mistaken to
>> make a little bit of
>>> noise if there is a release of some kind ready at
>> that point.
>>>
>>> As always, though, your call - and I'll help out
>> as I'm able.
>>>
>>> --
>>> David N. Welton
>>>  - http://www.dedasys.com/davidw/
>>>
>>> Linux, Open Source Consulting
>>>  - http://www.dedasys.com/
>>>
>>> _______________________________________________
>>> Dev mailing list
>>> [hidden email]
>>> http://lists.ofbiz.org/mailman/listinfo/dev
>>
>>>
>> _______________________________________________
>> Dev mailing list
>> [hidden email]
>> http://lists.ofbiz.org/mailman/listinfo/dev
>
>
> _______________________________________________
> Dev mailing list
> [hidden email]
> http://lists.ofbiz.org/mailman/listinfo/dev

 
_______________________________________________
Dev mailing list
[hidden email]
http://lists.ofbiz.org/mailman/listinfo/dev

[BJ Freeman]

what is the purpose of such an advertising?
I think that getting an install base in a certain area, would also be
important.
The clients I have lined up, don't really care if it is apache, Just
that it helps them in Business.
As the installs are completed, the business people go back to their
forums and express their positive experience. Best advertising in the
world, for the buck.
Just a different perspective.

Chris Howe sent the following on 2/28/06 10:49 PM:

> Can we start brainstoming some marketing ideas
> coordinated with with OFBiz's exit from the incubator
> and into the ASF?  Some of them might require a
> community effort to get noticed.
>
> 1. Submit a story for digg.com and have a community
> push to get it to the front page...and beyond
> 2. Release of blogs, reviews, editorials simultaneous
> to the acceptance (might want to exchange writeups for
> fact checking, grammar, etc beforehand.)
> 3.  Any other ideas?
>
>  
> _______________________________________________
> Dev mailing list
> [hidden email]
> http://lists.ofbiz.org/mailman/listinfo/dev
>

 
_______________________________________________
Dev mailing list
[hidden email]
http://lists.ofbiz.org/mailman/listinfo/dev

[David E. Jones]

Speaking of which... If anyone has any testimonials, case studies, or  
live site listings that they could share we'd love to get the  
ofbiz.org Home and Marketing pages updated and expanded a bit.

-David


On Mar 1, 2006, at 3:08 AM, BJ Freeman wrote:

> what is the purpose of such an advertising?
> I think that getting an install base in a certain area, would also be
> important.
> The clients I have lined up, don't really care if it is apache, Just
> that it helps them in Business.
> As the installs are completed, the business people go back to their
> forums and express their positive experience. Best advertising in the
> world, for the buck.
> Just a different perspective.
>
> Chris Howe sent the following on 2/28/06 10:49 PM:
>> Can we start brainstoming some marketing ideas
>> coordinated with with OFBiz's exit from the incubator
>> and into the ASF?  Some of them might require a
>> community effort to get noticed.
>>
>> 1. Submit a story for digg.com and have a community
>> push to get it to the front page...and beyond
>> 2. Release of blogs, reviews, editorials simultaneous
>> to the acceptance (might want to exchange writeups for
>> fact checking, grammar, etc beforehand.)
>> 3.  Any other ideas?
>>
>>
>> _______________________________________________
>> Dev mailing list
>> [hidden email]
>> http://lists.ofbiz.org/mailman/listinfo/dev
>>
>
> _______________________________________________
> Dev mailing list
> [hidden email]
> http://lists.ofbiz.org/mailman/listinfo/dev

 
_______________________________________________
Dev mailing list
[hidden email]
http://lists.ofbiz.org/mailman/listinfo/dev

[Andrew Sykes]

BJ,

I think the more holistic the approach to marketing the better, so the
purpose of any particular action needs to be considered as a constituent
of a marketing campaign.

You don't have to look far to find marketing gurus who tell you anything
between 5 and 20 exposures are required to elicit adoption of a product.

As for not caring about Apache, I must say I have a number of clients
who are VERY excited about Apache incubation. They see it as vindication
for their relatively high-risk early OFBiz adoption.

Andrew

On Wed, 2006-03-01 at 02:08 -0800, BJ Freeman wrote:

> what is the purpose of such an advertising?
> I think that getting an install base in a certain area, would also be
> important.
> The clients I have lined up, don't really care if it is apache, Just
> that it helps them in Business.
> As the installs are completed, the business people go back to their
> forums and express their positive experience. Best advertising in the
> world, for the buck.
> Just a different perspective.
>
> Chris Howe sent the following on 2/28/06 10:49 PM:
> > Can we start brainstoming some marketing ideas
> > coordinated with with OFBiz's exit from the incubator
> > and into the ASF?  Some of them might require a
> > community effort to get noticed.
> >
> > 1. Submit a story for digg.com and have a community
> > push to get it to the front page...and beyond
> > 2. Release of blogs, reviews, editorials simultaneous
> > to the acceptance (might want to exchange writeups for
> > fact checking, grammar, etc beforehand.)
> > 3.  Any other ideas?
> >
> >  
> > _______________________________________________
> > Dev mailing list
> > [hidden email]
> > http://lists.ofbiz.org/mailman/listinfo/dev
> >
>  
> _______________________________________________
> Dev mailing list
> [hidden email]
> http://lists.ofbiz.org/mailman/listinfo/dev

--
Kind Regards
Andrew Sykes <[hidden email]>
Sykes Development Ltd
http://www.sykesdevelopment.com

 
_______________________________________________
Dev mailing list
[hidden email]
http://lists.ofbiz.org/mailman/listinfo/dev

[Jacopo Cappellato]

Yes,

testimonials are important too.

In order to facilitate the collection of testimonials, what do you think
of preparing a simple standard form with predefined questions for
clients? The form could then be published in the OFBiz site (and maybe
also implemented with the OFBiz survey stuff).

The idea is to gather rather simple written testimonials, that can be
split up into three section (more or less three sentences):

1) information about the client (market of interest etc...)
2) how they are using OFBiz (advantages etc...)
3) information about the service provider (if any) that helped in the
implementation (quality of service etc...)

Then, 1&2 could be published in the OFBiz site; 1&3 or 1&2&3 could be
published in the service provider's site

Jacopo

David E. Jones wrote:
>
> Speaking of which... If anyone has any testimonials, case studies, or
> live site listings that they could share we'd love to get the ofbiz.org
> Home and Marketing pages updated and expanded a bit.
>
> -David
>

 
_______________________________________________
Dev mailing list
[hidden email]
http://lists.ofbiz.org/mailman/listinfo/dev

[Leon Torres-2]

Well, I'm one of those OO people. I wrote a simple framework once for
server-side application development in Java that was highly OO, and it
made things more organized, easier to change, and faster to refactor
than the Java code I inherited, which was written C style with almost
no objects. A thoughtful use of OO can make life easier.

I recommend we think about a more OO aproach to defining services than
the static methods we use (hello Procedural style). There are common
patterns in service design that can be extracted into a tidy class and
interface system. It's better to start now so that we can have something
to present later.

Also, I seriously recommend fully utilizing the flexibility of Log4r.
We need to have one logger per group of related services, that way we
can control the logging with much more granularity. I'm tired of seeing
10k line dumps from one entity error in a storeAll method. It would
facilitate debugging, by say, redirecting all the output from one logger
to a specific console so you can see the traces just for those services.
This would be best implemented with the OO service system I mentioned, so
no work has to be done to set them up.

These two things would make OFBiz feel much more Apache like, at least to me.
:-)

- Leon

David E. Jones wrote:

>
> I don't mind publicity, but I do mind setting expectations that are  
> eventually not met or that are delayed... in this project as well as  in
> general business... nothing eats at me more.
>
> With the Apache effort there is really a LOT to do. The worst thing  is
> that there are lots of little things because it turns out that the  few
> big things are rather complex... In other words it could really  take a
> long time, hopefully not too long of course, but on the order  of
> various months and not various weeks. It could happen that fast,  but I
> don't see it.
>
> So yeah, once we're ready to leap I'd like to put together all the  
> materials we can to not only advertise what exists in OFBiz, but also  
> put together materials to help people "get started" with it,  especially
> get an idea of what the framework is like and what it is  like to work
> with it. The point would be to resolve the technical  concerns in
> addition to the business functionality, and perhaps even  help push the
> technical/framework features as a primary benefit of  the project to a
> crowd that is often very strictly into OO stuff and  such...
>
> Anyway, it's probably fine to start this conversation, but everyone  
> should keep in mind that this may be a ways off...
>
> -David
>
>
> On Mar 1, 2006, at 12:37 AM, David Welton wrote:
>
>>> What I was thinking was a coordinated effort for if
>>> and when it becomes a full-fledged Apache project.
>>
>>
>> Let's make that happen first, but yes, there is a PR group at Apache
>> (including one person who is actually a PR person, rather than a
>> programmer interested in PR:-), and we can certainly talk with them.
>> I think David Jones is a bit worried about not being too visible too
>> early, but I don't think it would be mistaken to make a little bit of
>> noise if there is a release of some kind ready at that point.
>>
>> As always, though, your call - and I'll help out as I'm able.
>>
>> --
>> David N. Welton
>>  - http://www.dedasys.com/davidw/
>>
>> Linux, Open Source Consulting
>>  - http://www.dedasys.com/
>>
>> _______________________________________________
>> Dev mailing list
>> [hidden email]
>> http://lists.ofbiz.org/mailman/listinfo/dev
>
>
>
> ------------------------------------------------------------------------
>
>  
> _______________________________________________
> Dev mailing list
> [hidden email]
> http://lists.ofbiz.org/mailman/listinfo/dev

 
_______________________________________________
Dev mailing list
[hidden email]
http://lists.ofbiz.org/mailman/listinfo/dev

[David E. Jones]

Leon,

There has been a lot of discussion about this over the years, and  
that can be looked up in the mail archives, and there are even some  
places in the documents on the ofbiz.org site that mention things  
along these lines.

Pretty much everything in OFBiz started out being written in a more  
object oriented way, and even with code generation it became clear  
that the sheet volume of code would be so difficult to maintain that  
we could never compete with any existing ERP, CRM, or ecommerce  
packages. Even now there aren't enough people involved to compete in  
that way.

Given your comments I don't think you understand just how much of a  
difference there is in code volume and general effort between the  
best practice recommended tools in OFBiz and more traditional  
enterprise Java tools.

This isn't exactly a secret in the enterprise software world. Systems  
that have been written using OO techniques have largely failed, not  
in that none of them ever got off the ground, just that they simply  
could not compete with the more efficient tools that are used in  
traditional enterprise systems, especially those intended for larger  
businesses. For smaller businesses sometimes OO oriented packages  
make it better because they are customized less and have less and  
more generic needs.

Even now in OFBiz there is a LOT of stuff that was written in an  
attempt to follow OO patterns and I'd really like to see it ripped  
out... Even services written in Java methods that are only moving  
data around should be changed to simple-methods. I don't say this  
because I want to sell the simple-method tool, I say it because the  
Java code is usually a _real pain_ to maintain. A pretty good  
majority of the issues we have run into in the past have been in such  
code. It is simply more error prone because it is more complicated  
and requires more lines of code than alternatives.

The Service Oriented Architecture driven by the Service Engine may  
seem somewhat similar to procedural programming, but don't worry, it  
is simply not the case. This is a very important part of the OFBiz  
architecture and would require tens of thousands of man hours to  
change things to be done differently, not counting the continual  
additional man hours that would be required going forward to maintain  
such code...

Could the tools use some improvements? Yes, of course. Over the years  
they have seen many thousands of improvements. Logging is an area  
that could still be improved, though I guess it hasn't bothered  
anyone (including me) enough to do something about it yet...

There are also ways of doing things that are REALLY not recommended,  
like the storeAll method. That was only created early on because  
before transaction management was added it was the only way to group  
operations in a single ACID block (ie using the implicit transaction  
in a JDBC connection instead of explicit transactions using JTA).  
Clearly not a good way to do things, and should be eliminated  
everywhere, but hasn't been. This is especially tempting in Java code  
because with all of the exception handling and other very non-
business level stuff there is a lot of work using more recommended  
methods.

Okay, I guess I'm going into detail anyway, might as well finish  
up... The biggest problem with OO code and design patterns is that  
_nothing_ in the world of business software is actually object  
oriented.... Unless you choose to use a OO database (which are a real  
PITA in my opinion...) you would need an object-relational mapping  
layer. This usually involves a TON of code that is often generated  
because there is SO much redundancy in it and it is SO verbose. With  
OFBiz, we simple don't have to deal with that because we preserve the  
relational structures from bottom to top.

And yes, I've heard every response in the book to this notion and the  
OO principles it violates. No compile time data protection: this is  
dumb because you _never_ have that on lower levels anyway. No  
abstraction to avoid requiring changes in higher layers when you  
change the data model: this is bogus because how many db changes are  
not done to facilitate some user or system interaction?

The same thing is true on the UI layer. There are so many attempts to  
make web based applications object oriented, but it doesn't work, so  
there is always a lot of mapping and often a lot of compromises that  
limit flexibility. In other words: more code with less results and  
less efficiency.

With OFBiz the tools work together so that you can derive artifacts  
from lower level artifacts so it actually often requires very many  
fewer changes when adding fields or general new functionality. The  
insulation between layers with OO stuff just back-fires in a big way  
because of the increased complexity and size when you are mapping  
between very different things to very different things instead of  
following more natural patterns.

For some things I'm certainly in favor of OO techniques and tools. In  
fact OFBiz and the software world in general would not be where it is  
today with out them. But applying these things to business data  
structures, user interfaces, and even business logic, just causes  
serious code bloat and kills productivity.

If someone wanted to try to re-implement things in OFBiz this way I  
say go for it. If you succeed and can reproduce the OFBiz artifacts  
with more efficient artifacts using OO techniques I'll be very  
impressed. This has never happened, and from past experience with so  
many other alternatives to solve these same problems I don't see it  
_ever_ happening. The OFBiz tools and conventions certainly still  
have room for improvements, but moving away from a service and entity  
oriented architecture to something more object oriented isn't going  
to help us.

-David


On Mar 1, 2006, at 7:21 PM, Leon Torres wrote:

> Well, I'm one of those OO people. I wrote a simple framework once for
> server-side application development in Java that was highly OO, and it
> made things more organized, easier to change, and faster to refactor
> than the Java code I inherited, which was written C style with almost
> no objects. A thoughtful use of OO can make life easier.
>
> I recommend we think about a more OO aproach to defining services than
> the static methods we use (hello Procedural style). There are common
> patterns in service design that can be extracted into a tidy class and
> interface system. It's better to start now so that we can have  
> something
> to present later.
>
> Also, I seriously recommend fully utilizing the flexibility of Log4r.
> We need to have one logger per group of related services, that way we
> can control the logging with much more granularity. I'm tired of  
> seeing
> 10k line dumps from one entity error in a storeAll method. It would
> facilitate debugging, by say, redirecting all the output from one  
> logger
> to a specific console so you can see the traces just for those  
> services.
> This would be best implemented with the OO service system I  
> mentioned, so
> no work has to be done to set them up.
>
> These two things would make OFBiz feel much more Apache like, at  
> least to me.
> :-)
>
> - Leon
>
> David E. Jones wrote:
>>
>> I don't mind publicity, but I do mind setting expectations that are
>> eventually not met or that are delayed... in this project as well  
>> as  in
>> general business... nothing eats at me more.
>>
>> With the Apache effort there is really a LOT to do. The worst  
>> thing  is
>> that there are lots of little things because it turns out that  
>> the  few
>> big things are rather complex... In other words it could really  
>> take a
>> long time, hopefully not too long of course, but on the order  of
>> various months and not various weeks. It could happen that fast,  
>> but I
>> don't see it.
>>
>> So yeah, once we're ready to leap I'd like to put together all the
>> materials we can to not only advertise what exists in OFBiz, but also
>> put together materials to help people "get started" with it,  
>> especially
>> get an idea of what the framework is like and what it is  like to  
>> work
>> with it. The point would be to resolve the technical  concerns in
>> addition to the business functionality, and perhaps even  help  
>> push the
>> technical/framework features as a primary benefit of  the project  
>> to a
>> crowd that is often very strictly into OO stuff and  such...
>>
>> Anyway, it's probably fine to start this conversation, but everyone
>> should keep in mind that this may be a ways off...
>>
>> -David
>>
>>
>> On Mar 1, 2006, at 12:37 AM, David Welton wrote:
>>
>>>> What I was thinking was a coordinated effort for if
>>>> and when it becomes a full-fledged Apache project.
>>>
>>>
>>> Let's make that happen first, but yes, there is a PR group at Apache
>>> (including one person who is actually a PR person, rather than a
>>> programmer interested in PR:-), and we can certainly talk with them.
>>> I think David Jones is a bit worried about not being too visible too
>>> early, but I don't think it would be mistaken to make a little  
>>> bit of
>>> noise if there is a release of some kind ready at that point.
>>>
>>> As always, though, your call - and I'll help out as I'm able.
>>>
>>> --
>>> David N. Welton
>>>  - http://www.dedasys.com/davidw/
>>>
>>> Linux, Open Source Consulting
>>>  - http://www.dedasys.com/
>>>
>>> _______________________________________________
>>> Dev mailing list
>>> [hidden email]
>>> http://lists.ofbiz.org/mailman/listinfo/dev
>>
>>
>>
>> ---------------------------------------------------------------------
>> ---
>>
>>
>> _______________________________________________
>> Dev mailing list
>> [hidden email]
>> http://lists.ofbiz.org/mailman/listinfo/dev
>
> _______________________________________________
> Dev mailing list
> [hidden email]
> http://lists.ofbiz.org/mailman/listinfo/dev

 
_______________________________________________
Dev mailing list
[hidden email]
http://lists.ofbiz.org/mailman/listinfo/dev

[Leon Torres-2]

David,

I understand your concerns. In fact, I agree with them. However, I wasn't
talking about the superstructure of the framework, the SOA itself. I was talking
about how a subset of services can be better organized. It's a much lower level
topic and has _nothing_ to do with rewriting OFBiz. It was an error on my part
to profess some kind of OO zealotry when I'm nearly the complete opposite.

This code example should clarify my meaning. In controller.xml:

     <event type="ofbiz-soa-service" invoke="ServiceClass"/>

"ofbiz-soa-service" is an event handler that delegates to a self-contained
mini-framework. This mini-framework would allow each service to be set up as a
separate AbstractServiceClass with its own logger, startup/shutdown proceedure,
allow hotplugging single services without restarting, provide interfaces to help
write create/update methods (completely optional of course), etc. The services
are defined as usual in a servicedef file and would co-exist with the rest of
ofbiz nicely.

And the most important benefit, I believe, is that it would allow us to
implement a dependancy and upgrade system for services similar to package
management on linux distros. It would be best to do this in a separate
mini-framework. (This has been a desire of mine for awhile actually.)

Such a mini-framework could even be considered a separate project. It doesn't
even have to be Java centric. In fact, I'm hoping to prototype it in Ruby. What
better way to demonstrate this idea than to provide actual working code?

Again, the intent is not to rewrite OFBiz using some perfectionist ideal, but to
provide some added features in a self-contained system that would benefit end
users and developers. I was hoping that the addition of such would make OFBiz
even more attractive to Apache people: Hey, you could plug your own code
management system for services in!

- Leon



David E. Jones wrote:

> Leon,
>
> There has been a lot of discussion about this over the years, and  
> that can be looked up in the mail archives, and there are even some  
> places in the documents on the ofbiz.org site that mention things  
> along these lines.
>
> Pretty much everything in OFBiz started out being written in a more  
> object oriented way, and even with code generation it became clear  
> that the sheet volume of code would be so difficult to maintain that  
> we could never compete with any existing ERP, CRM, or ecommerce  
> packages. Even now there aren't enough people involved to compete in  
> that way.
>
> Given your comments I don't think you understand just how much of a  
> difference there is in code volume and general effort between the  
> best practice recommended tools in OFBiz and more traditional  
> enterprise Java tools.
>
> This isn't exactly a secret in the enterprise software world. Systems  
> that have been written using OO techniques have largely failed, not  
> in that none of them ever got off the ground, just that they simply  
> could not compete with the more efficient tools that are used in  
> traditional enterprise systems, especially those intended for larger  
> businesses. For smaller businesses sometimes OO oriented packages  
> make it better because they are customized less and have less and  
> more generic needs.
>
> Even now in OFBiz there is a LOT of stuff that was written in an  
> attempt to follow OO patterns and I'd really like to see it ripped  
> out... Even services written in Java methods that are only moving  
> data around should be changed to simple-methods. I don't say this  
> because I want to sell the simple-method tool, I say it because the  
> Java code is usually a _real pain_ to maintain. A pretty good  
> majority of the issues we have run into in the past have been in such  
> code. It is simply more error prone because it is more complicated  
> and requires more lines of code than alternatives.
>
> The Service Oriented Architecture driven by the Service Engine may  
> seem somewhat similar to procedural programming, but don't worry, it  
> is simply not the case. This is a very important part of the OFBiz  
> architecture and would require tens of thousands of man hours to  
> change things to be done differently, not counting the continual  
> additional man hours that would be required going forward to maintain  
> such code...
>
> Could the tools use some improvements? Yes, of course. Over the years  
> they have seen many thousands of improvements. Logging is an area  
> that could still be improved, though I guess it hasn't bothered  
> anyone (including me) enough to do something about it yet...
>
> There are also ways of doing things that are REALLY not recommended,  
> like the storeAll method. That was only created early on because  
> before transaction management was added it was the only way to group  
> operations in a single ACID block (ie using the implicit transaction  
> in a JDBC connection instead of explicit transactions using JTA).  
> Clearly not a good way to do things, and should be eliminated  
> everywhere, but hasn't been. This is especially tempting in Java code  
> because with all of the exception handling and other very non-
> business level stuff there is a lot of work using more recommended  
> methods.
>
> Okay, I guess I'm going into detail anyway, might as well finish  
> up... The biggest problem with OO code and design patterns is that  
> _nothing_ in the world of business software is actually object  
> oriented.... Unless you choose to use a OO database (which are a real  
> PITA in my opinion...) you would need an object-relational mapping  
> layer. This usually involves a TON of code that is often generated  
> because there is SO much redundancy in it and it is SO verbose. With  
> OFBiz, we simple don't have to deal with that because we preserve the  
> relational structures from bottom to top.
>
> And yes, I've heard every response in the book to this notion and the  
> OO principles it violates. No compile time data protection: this is  
> dumb because you _never_ have that on lower levels anyway. No  
> abstraction to avoid requiring changes in higher layers when you  
> change the data model: this is bogus because how many db changes are  
> not done to facilitate some user or system interaction?
>
> The same thing is true on the UI layer. There are so many attempts to  
> make web based applications object oriented, but it doesn't work, so  
> there is always a lot of mapping and often a lot of compromises that  
> limit flexibility. In other words: more code with less results and  
> less efficiency.
>
> With OFBiz the tools work together so that you can derive artifacts  
> from lower level artifacts so it actually often requires very many  
> fewer changes when adding fields or general new functionality. The  
> insulation between layers with OO stuff just back-fires in a big way  
> because of the increased complexity and size when you are mapping  
> between very different things to very different things instead of  
> following more natural patterns.
>
> For some things I'm certainly in favor of OO techniques and tools. In  
> fact OFBiz and the software world in general would not be where it is  
> today with out them. But applying these things to business data  
> structures, user interfaces, and even business logic, just causes  
> serious code bloat and kills productivity.
>
> If someone wanted to try to re-implement things in OFBiz this way I  
> say go for it. If you succeed and can reproduce the OFBiz artifacts  
> with more efficient artifacts using OO techniques I'll be very  
> impressed. This has never happened, and from past experience with so  
> many other alternatives to solve these same problems I don't see it  
> _ever_ happening. The OFBiz tools and conventions certainly still  
> have room for improvements, but moving away from a service and entity  
> oriented architecture to something more object oriented isn't going  
> to help us.
>
> -David
>
>
> On Mar 1, 2006, at 7:21 PM, Leon Torres wrote:
>
>
>>Well, I'm one of those OO people. I wrote a simple framework once for
>>server-side application development in Java that was highly OO, and it
>>made things more organized, easier to change, and faster to refactor
>>than the Java code I inherited, which was written C style with almost
>>no objects. A thoughtful use of OO can make life easier.
>>
>>I recommend we think about a more OO aproach to defining services than
>>the static methods we use (hello Procedural style). There are common
>>patterns in service design that can be extracted into a tidy class and
>>interface system. It's better to start now so that we can have  
>>something
>>to present later.
>>
>>Also, I seriously recommend fully utilizing the flexibility of Log4r.
>>We need to have one logger per group of related services, that way we
>>can control the logging with much more granularity. I'm tired of  
>>seeing
>>10k line dumps from one entity error in a storeAll method. It would
>>facilitate debugging, by say, redirecting all the output from one  
>>logger
>>to a specific console so you can see the traces just for those  
>>services.
>>This would be best implemented with the OO service system I  
>>mentioned, so
>>no work has to be done to set them up.
>>
>>These two things would make OFBiz feel much more Apache like, at  
>>least to me.
>>:-)
>>
>>- Leon
>>
>>David E. Jones wrote:
>>
>>>I don't mind publicity, but I do mind setting expectations that are
>>>eventually not met or that are delayed... in this project as well  
>>>as  in
>>>general business... nothing eats at me more.
>>>
>>>With the Apache effort there is really a LOT to do. The worst  
>>>thing  is
>>>that there are lots of little things because it turns out that  
>>>the  few
>>>big things are rather complex... In other words it could really  
>>>take a
>>>long time, hopefully not too long of course, but on the order  of
>>>various months and not various weeks. It could happen that fast,  
>>>but I
>>>don't see it.
>>>
>>>So yeah, once we're ready to leap I'd like to put together all the
>>>materials we can to not only advertise what exists in OFBiz, but also
>>>put together materials to help people "get started" with it,  
>>>especially
>>>get an idea of what the framework is like and what it is  like to  
>>>work
>>>with it. The point would be to resolve the technical  concerns in
>>>addition to the business functionality, and perhaps even  help  
>>>push the
>>>technical/framework features as a primary benefit of  the project  
>>>to a
>>>crowd that is often very strictly into OO stuff and  such...
>>>
>>>Anyway, it's probably fine to start this conversation, but everyone
>>>should keep in mind that this may be a ways off...
>>>
>>>-David
>>>
>>>
>>>On Mar 1, 2006, at 12:37 AM, David Welton wrote:
>>>
>>>
>>>>>What I was thinking was a coordinated effort for if
>>>>>and when it becomes a full-fledged Apache project.
>>>>
>>>>
>>>>Let's make that happen first, but yes, there is a PR group at Apache
>>>>(including one person who is actually a PR person, rather than a
>>>>programmer interested in PR:-), and we can certainly talk with them.
>>>>I think David Jones is a bit worried about not being too visible too
>>>>early, but I don't think it would be mistaken to make a little  
>>>>bit of
>>>>noise if there is a release of some kind ready at that point.
>>>>
>>>>As always, though, your call - and I'll help out as I'm able.
>>>>
>>>>--
>>>>David N. Welton
>>>> - http://www.dedasys.com/davidw/
>>>>
>>>>Linux, Open Source Consulting
>>>> - http://www.dedasys.com/
>>>>
>>>>_______________________________________________
>>>>Dev mailing list
>>>>[hidden email]
>>>>http://lists.ofbiz.org/mailman/listinfo/dev
>>>
>>>
>>>
>>>---------------------------------------------------------------------
>>>---
>>>
>>>
>>>_______________________________________________
>>>Dev mailing list
>>>[hidden email]
>>>http://lists.ofbiz.org/mailman/listinfo/dev
>>
>>_______________________________________________
>>Dev mailing list
>>[hidden email]
>>http://lists.ofbiz.org/mailman/listinfo/dev
>
>
>  
> _______________________________________________
> Dev mailing list
> [hidden email]
> http://lists.ofbiz.org/mailman/listinfo/dev
>

 
_______________________________________________
Dev mailing list
[hidden email]
http://lists.ofbiz.org/mailman/listinfo/dev