Hi David,
Thanks for the technical feedback on workday. What I was more
intrigued by them was not the internal architecture but the way they
took the approach for building a system from business process
perspective and able to answer when, where, how and why of a
particular transaction and have a nice integrated UI on top which
very well serve the regulatory requirement as well.
In general since I am more into supporting direct business and apply
technology for achieving business objective, the approach made a
perfect sense to me, and if you see the article the Finance comes
automatically in the end. They were not trying to build a Finance
system and then slap supporting systems around. This is what I was
trying to emphasize on ofbiz that it can do the same since it's not
trying to be Finance system and then put other parts in place.
I have seen other open source alternative like Compiere, Adempiere
(offshoot of Compiere), Postbooks, OpenBravo (the data model
borrowed from Compiere, which in turn inspired by SAP), TinyERP they
all are built around Finance and then supporting process built
around Financial Management. Indeed Oracle and SAP works the same
way although they try to cater to multiple industries and in trying
to be generic they are really very heavy.
This is for some good reason as well since the heaviest users for
ERP are mostly financial staffs (which are forced to use a system
due to regulatory and statutory requirements). The company is
compelled to invest money in these system because it's forced and
then if the same vendor offers a better business management suite
along with Finance it's an easy sale.
Recently I was reading Mcluhan theory of Laws of Media which opened
my eyes on to how innovation works. It's altogether a different way
of looking at business. I find workday was trying to do the same
taking up a different approach.
Ofbiz is still work in progress it's there and not there yet I
envision ofbiz as a single integrated development framework for
building Unified Business Applications (means you have single
framework handling your core business). Then it becomes easy to
manage such system with homogenous team which can rotate across or
homogenous infrastructure team which can rotate across. I saw the
same thing about ofbiz in my company and did some presentation of
which some parts I can share with the community later.
But you know when I wanted to sale the product the only drawback
came was the UI although it made perfect sense to me but to other
users it seems very complicated as you know sometimes first
impression is the last impression it would be nice if the OOTB
applications are made a little appealing.
Anyways this was a very long post and I can write a lot about it
since I really tried selling this concept against SAP, Oracle
Netsuite and other open source offerings. In one case I was
successful as well and its work in progress.
Thanks
With best regards,
Vikrant
-----Original Message-----
From: David E Jones [mailto:[EMAIL PROTECTED]
Sent: Thursday, January 17, 2008 2:30 AM
To: user@ofbiz.apache.org
Subject: Re: Workday
Out of curiosity I looked into Workday a little more. It sounds like
they are using a VERY object-oriented architecture, so much so that
they only have 3 database tables and map all objects to them in an
automated way.
This may seem really nice from a development perspective, but I've
been through a few of these systems and they are a bit of a nightmare.
When you get into millions of records and query on more than one
field/
column you end up joining on the same table, etc, etc.
My worst experience with this was part of a project to migrate of off
such a system, namely the now mostly defunct Blue Martini which did
persistence in a very similar way. Fortunately I've never done
significant development or support on such a system. The queries to
pull data from the database were massive and required about 3 man-
months of effort to complete the migration of product, customer, and
order information to the OFBiz database. The normal effort for that
sort of thing is generally around 2-3 weeks for complex mid-high end
systems, or for simple data (like basic product information
spreadsheets) the time required can get down to a couple of hours.
It's amazing to me how many systems still try to do things this way.
IMO OFBiz has about the right number of tables and columns in the
database, ie 758 tables and 7817 columns. The tables in general
represent a nice level of conceptual granularity and are adaptable to
many variations on the supported concepts.
When looking at a database for an enterprise system I get really
scared by 2 major trends:
1. a really small number of tables, like 3, where everything is in the
object model
2. a really large number of tables and columns, like Oracle Financials
~2 million columns, where things are excessively denormalized,
redundant and in general just chaotic
-David
On Jan 15, 2008, at 7:26 PM, <[EMAIL PROTECTED]> <[EMAIL PROTECTED]
wrote:
HI David,
Further to this would suggest following article they are good:
http://blogs.zdnet.com/BTL/?p=5983#more-5983
http://blogs.zdnet.com/Howlett/?p=128
Thanks
With best regards,
Vikrant
-----Original Message-----
From: David E Jones [mailto:[EMAIL PROTECTED]
Sent: Wednesday, January 16, 2008 3:10 AM
To: user@ofbiz.apache.org
Subject: Re: Workday
I would be really interested in hearing more from Vikrant about what
he likes about Workday. If you're listening in Vikrant, please do
write more about what parts of the Workday approach you like.
For my part I have looked into it a little bit and I think their
approach is great. More specifically what I like is the idea of
creating a software suite for a particular type of business that
basically does everything that such a business would need. Doing just
this sort of thing as a derivative work based on OFBiz has been part
of the long-term strategy from the beginning.
The accounting piece would benefit from this, but in a way that is
similar to how all other parts would benefit. That benefit is
reducing
redundant data and tying things together directly that might
otherwise
be in different systems. For accounting this is great because you
don't need to introduce so many artificial structures to take partial
data from other systems that is needed for operations or reporting,
especially ad-hoc reporting when creating a data warehouse to
consolidate data from multiple systems is not really a tenable
approach. The "cost centers" thread that was active recently is a
good
example of partially redundant data that is necessary because of
separate systems and a light integration between them.
I don't know if that is what Vikrant liked about Workday, but that is
something that struck me as I was reviewing the public material
available.
-David
On Jan 11, 2008, at 2:59 AM, Jacques Le Roux wrote:
Hi all,
I recently had an exchange with Vikrant Rathore about new accounting
efforts. I think it is worth to be reported, here it is
Vikrant : Have a look at http://www.workday.com/ I would prefer to
implement finance in their way not the traditional ERP way.
Me : What is so specific in their offer apart it's an ASP ?
Hi Jacques,
In financial component they combined 5 major parts of Finance:
1. GL Management (ofbiz is still working on to implement it)
2. Reporting (both statutory and financial management related)
3. Auditing
4. Budgeting and tracking.
5. Asset Management.
Normally none of the Financial Management system itself comes close
to it. As you have to buy separate software to do it. Besides this
the way they presented the whole solution is more business oriented
then technical jargon oriented which is what most ERP's do.
So I feel there way of doing Financial Management linked to all the
other parts of business is a good implementation. You do not need a
separate budgeting, asset management, reporting and auditing
requirements (for SOX compliance). Everything comes out of the box.
I haven't seen such a offering either by SAP business byDesign or
Netsuite and I guess ofbiz can do it since it's a single data model
but no component is there to address it.
Regards,
Vikrant
