Re: Reposting from David Jones on Moqui

[Adrian Crum]

Switching OFBiz to a different framework has been discussed in the past, 
and I brought up the subject again in a recent thread on the dev mailing 
list. At this time, there are some PMC members who are opposed to the 
idea, so I don't see any hope for switching to Moqui in this project.

If anyone is interested in building applications on Moqui, then they 
should do it in the Moqui community.

I don't agree that OFBiz is a sinking ship. There are a number of 
service providers who are heavily invested in this project, and they are 
not going to throw that all away for a new one.

Adrian Crum
Sandglass Software
www.sandglass-software.com

On 3/7/2014 5:58 AM, Al Byers wrote:
In light of the current discussion about the future of OFBiz I thought it
would be worth revisiting one of my finer moments by reposting this from
David Jones. In today's world 12 technology years is a lifetime. It is not
reasonable to think that we can keep scaling the original framework.
Another thing to think about is that in order to attract a new group of
users sometimes it takes a big new idea and Moqui fills that role. We
should all keep in mind that OFBiz succeeded secondly because of all the
hard work put in by the community, but firstly because of the brilliant
architecture and foundation provided by David (and Andrew). When that same
mind applies it to fixing most of OFBiz's problems we should be talking
about how to transition - end of discussion, IMHO.

I believe that it has been since David wrote this email that he has
integrated Elasticsearch and Drools into Moqui. Those are the kinds of
exciting technologies that OFBiz needs. And it is high time that we stopped
programming in Java. I have loved the seemless transition to Groovy and I
like having only one scripting language. The future is going to demand much
more flexibility than OFBiz can deliver. Moqui, with its small core size
and its application of FreeMarker to the frontend provides that flexibility.

I don't know exactly how the transition to Moqui should take place, but
that is where the discussion should take place. Anything else, and we are
just discussing how slowly the ship will be sinking. I think David should
remain in control of the core (as I think he would insist), but there are a
lot of options for porting the mantle and crust portions. I think we should
think outside the box and look to something like KickStarter to get it done
quickly.

-Al

On Wed, May 22, 2013 at 2:26 PM, David E. Jones <d...@me.com> wrote:


OFBiz has recently turned 12 years old. At the time it was written many
more modern libraries either didn't exist or were not usable, including:

- Groovy
- ehcache
- Quartz Scheduler
- Atomikos
- JackRabbit (and JCR in general)
- Shiro
- Camel
- JSON-RPC, REST, JSON in general
- ElasticSearch (and to some extent even Lucene)
- Document and other NoSQL databases (of which ElasticSearch is sort of,
but I mean CouchDB, MongoDB, Hadoop and derivatives, etc)

Some of these are used, or with some customization usable, in OFBiz. Many
of them overlap a lot with parts of the OFBiz framework, and unlike
JPA/Hibernate sorts of things, do a better job than what is in OFBiz.

Some big ones are caching, job scheduling, content management, and even
searching. The OFBiz ProductSearch stuff works well enough (though not
great) for smaller sets of products, but doesn't compare in flexibility,
scalability, and speed to ElasticSearch and some other Lucene-based
alternatives. With some simple framework extensions (like the DataDocument,
DataFeed, and DataSearch features of Moqui) implementing excellent search
for products would be easy, as would search for any other part of the
system... and all combined in a single system-wide search or segmented as
desired.

Another big one, that has been most painful for me in dealing with OFBiz,
is the lack of consistent scripting and expressions. Once you get used to
the elegance of Groovy dealing with BSH and JUEL is downright painful...
and for me anyway requires a number of misses before I finally get it
working. The ${groovy:...} work-around is there, but quirky, and the
resulting object is unreliable as in some OFBiz XML files it results in a
String while in others it results in the actual Object the expression
evaluates to.

Even if it is self-serving, I agree that OFBiz was brilliant in its day,
but it needs FAR more modernization than is currently happening or that is
likely to happen. The new feature velocity in the framework is so slow
(mostly because of the architecture and existing code, partly because of
collaboration breakdown reasons), that it can't keep up with alternatives.

So yes, OFBiz is great, but it exists in a world that is progressing far
faster than it can. My reason for starting fresh was just that simple:
development velocity.

On top of that OFBiz uses certain approaches that are difficult to deploy
and maintain. Try dropping all of OFBiz into a single war file for easy
upload deployment on the dozens of modern cloud/PaaS services. Try adding
plug-ins that require a proper init/destroy lifecycle instead of relying on
static initialization and no proper tear down. Try finding framework
functionality in thousands of static methods spread across dozens (or
hundreds?) of classes. I know these weaknesses of OFBiz well... they are my
mistakes. Correcting them is another matter... and one I didn't find
possible in the context of the project with the limited time I have
available. It was faster and easier to start fresh.

When I started OFBiz I was 23 years old and had about 2 years of
experience in ERP systems. I think it's great that there is enough interest
to keep the project alive and at whatever pace keep it progressing both
technically and for support of business activities. Still, something must
be done for it to remain competitive with open source and commercial
alternatives if it is to compete... including with what I've been calling
the "Next Generation" of OFBiz, ie Moqui Framework, Mantle Business
Artifacts, and the various projects and products built on them.

As good as it is, there is lots of room for improvement and others are
doing just that. I don't think Al was implying that "OFBiz is no longer
brilliant", maybe some are overly sensitive to that. The fact is that OFBiz
is what it is, and without major improvements alternatives exceed it in so
many ways. It doesn't make OFBiz less brilliant, but in a sky with other
bright stars its brilliance is only relevant in context.

OFBiz has lots of momentum, and pretty good marketplace around it, and a
lot of people are making good money doing work based on it (including me).
Still, I tire frequently of explaining that so many things are known issues
with the project and not easy to correct, but are corrected in the "Next
Generation", ie Moqui/Mantle. Usually the fix is a hack and workaround that
can't be committed because it breaks other things, just things they don't
intend to use (this still has consequences for bigger projects... things
all seem to come back around).

So, it is what it is. I understand the motivation to paint OFBiz the best
possible for marketing purposes and such... I personally did that for years
in spite of known flaws. Eventually that only goes so far... OFBiz versus
other open source alternatives has its pluses and minuses, and most in the
community are very aware of those minuses. This causes many to drool over
cleaner, newer solutions like Magento, even if it is based on a totally
different underlying technology and one that doesn't scale as well or
interact in enterprise environments as well.

Sooner or later reality catches up... best to stay ahead of it or at least
have long-term plans and alternatives to fall-back on.

-David


On May 20, 2013, at 11:10 AM, Adrian Crum <
adrian.c...@sandglass-software.com> wrote:

A quick clarification on this.

"OFBiz was brilliant when David created it over ten years ago, but..."
implies OFBiz is no longer brilliant. OFBiz continues to be just as
brilliant, with a talented team of developers keeping it current with
current technology.

-Adrian

On 5/20/2013 4:04 PM, Al Byers wrote:
Hi Carlos,

I am just starting to look around for OFBiz work and was intrigued to
see your email there this morning. I have been working with OFBiz for over
10 years now and am interested in what you have going.

But I must ask if you have considered Moqui (moqui.org <
http://moqui.org>) - David Jones's successor to OFBiz? I was just at a
small conference with David and the folks at HowWax Media and based on
David's comments and what I know about Moqui from my past year of working
with it, if you are starting anew, and especially if you are not using the
current e-commerce features of OFBiz, then you would be well served to look
at Moqui. OFBiz was brilliant when David created it over ten years ago, but
technology has made great advances in that time and if you have the freedom
to do so, it makes sense to start with the latest base.

I have attached David's introduction to Moqui PDF, which I don't think
is readily available off the moqui.org <http://moqui.org> website.

I hope to hear from you soon.

Al Byers
801-400-5111


On Mon, May 20, 2013 at 7:19 AM, Carlos Cruz <car...@nbtbizcapital.com<mailto:
car...@nbtbizcapital.com>> wrote:

    Hi;

    I'm looking for a Java programmer that is familiar with OFBiz.
    Particularly with OFBiz Web Services and OFBiz Entity Engine.

    I'm interested in hosting OFBiz for some very specific industries
    and I want to develop some very specific interfaces.

    This is a long term project, I could be flexible with the hours.

    If you're interested email me for more details.

    Also feel free to forward this email to someone you think might be
    interested.

    Thanks!!

    Carlos

    logo-for-social-media-sites-email_signature

    CruzControl Radius

    Your Success Is Our Service

    www.ccradius.com <http://www.ccradius.com>

    email:car...@ccradius.com <mailto:email%3acar...@ccradius.com>

    1-877-285-5499 <tel:1-877-285-5499>