Fusion = Apps 13

This article by Tim Dexter of BI Publisher name and fame just tells in between the lines that all eBusiness Suite reports will be ported to Fusion. What is in fact the same as saying that Fusion is Apps 13, an evolution from R12, not a new game.

So the good old per_all_assignments_f will still exists. Maybe with a good looking UI on top of it, which passed for the first time an usability lab. Skip the last part, I'm asking too much.

SOA only matters at the back end side, not at the front end, despite the Web 2.0 vague.

My Virtual Apps

Since a few days I run APPS via a VMWARE virtual machine on my little laptop. Vision 11.5.10. And all data on an external disk.

With 1G of memory it is slow, but not too slow. Anyway, I ordered already an extra Giga RAM.

Summary till now: i love it, but it needs to be seen how handy this setup will be when i have to restore the virtual machine, a heafty 40G of unzipping.

Next step is to run Documentool on it. And publish the output. But we discovered some data corruptions in the Vision database, which needs to be worked around in Documentool.

I expected these corruptions. You see them on any project, and the Vision database is a very old, ongoing project.

Oracle Fusion Capable

Documentool is Oracle Fusion Capable.

Do you need an additional save harbor statement?

A license for one LOV?

The very excited people contact form has two fields to store the begin and end reason of a contact relationship.

These lists of values are standard empty, but when you create some Personal life events, they show up nicely in the relationship reason fields.

Too bad that both List of Values show identical reasons. A divorce is not the most straightforward reason to start a relationship. Or maybe it is, but then in the other way.

Even worse is that you need a license for OAB (Oracle Advanced Benefits) to create these personal life events. So far the integrated solution.

Forms personalizations can eventually overcome both issues.

Distributed costing, again

First of all, the input value over which you want to distribute must be Pay Value. Any other name will not do, even if the uom is Money.

And it seems that distribution over distribution seems to work (by accident, or by purpose?). When setting up the element link for distributed costing, it will tell you the element is not eligible for distributed costing since it is part of a distributed element set.

Remove it from ther element set, configure the element link, and add it back to the element set.

Desupport of 11.5.9

As you can read on Schan's excellent blog , 11.5.9 Premier Support ends by June 2008. This is almost tomorrow. There is no Extended Support offered, but Sustaining Support is there to stay.

By the way, Sustaining Support is not giving you any new updates.

So if you are on 11.5.9 with a localized payroll, you simply need these patches and fixes, due to statutory legislative changes. So Sustaining Support is not an option. So you have to upgrade to 11.5.10. By tomorrow.

my fast formula editor

I used several years the free ConText editor to edit fast formulas. Recently Paolo suggested me Notepad++, which is open source, and does a great job in syntax highlighting and brace matching.

I used it today for the whole day, and even when i have to get used to shortcuts and menu options, i would not be surprised to stick with it.

By the way, i used the sql language as syntax style.

Oracle acquires BEA

The first thought is that this is a pure middleware play.

Buy installed base and generate support revenue.

Or has it to do something with Fusion Apps. I do not think so. The architecture is defined right now, and BEA will only play a role later on in the life cycle of Fusion Apps. If any.

But the hidden gem could be Tuxedo. Oracle wants to position itself in the primary processes. Think banking processes for banks. And these transactions are handled by Tuxedo, already for years.

And there is the big money. Compared to SAP that hunts for the little cents with Business By Design.

Inovation. Nope. Smart. Yes.

Any link with HR. Nope.

HR 2.0

Every HR implementation tries to structure all kinds of employee related data within the venders shoebox. We just make it fit, but a lot of data and processes are not structured at all.

Why not load all employees within a social network, tag them with competences, positions, domains of interest, ... from the corporate HR system? And then let the employees connect to their peers, discuss issues, create communities within the enterprise, learn from collegues the other side of the world. Free flow versus structure.

From a technical point of view, it is easy. But which organisation will be ready to give this a try?

gross up rounding error

One of the outputs generated by the gross up formula is the remainder. We noticed that we missed a few eurocents in costing, due to these remainders.

First we set the "to within" input value to 1 eurocent, but gross up still was coming up with a few remainders of one cent.

And we could not decrease the value of "to within", since the unit of measure was money.

So we changed the unit of measure to Number, filled 0.009 as the "to within" tollerance, and grossup did a perfect job, leaving 0 as remainder. We still have to see if this perfect job has no performance consequences.

We opted not to create an additional element to store the remainder, and cost it.

Gross up, out of the box

This article describes the configuration steps for the grossup or gross to net functionality. The seeded fast formulas are used, but you could implement your own variations on them.

1. define an element type.

• In the Advance tab, set Gross up (which activates the iterative flag and the process seperate)
• define input values, based on the inputs of the seeded fast formula default_grossup.

• Pay Value
• Amount
• Additional amount
• Low gross
• High gross
• Remainder
• To within
• Method (defaulted to interpolation)
  
• Grossup balance

• map the input values to the inputs of the fast formula (which is in this case a 1-1) via the iterative rules button.
  


• (optional) exclude grossup balances, via the exclude balances button. Only those balances that are defined as grossup show up in the list to be excluded.

2. Link the seeded fast formula calc_grossup_pay_value to the element type created above. The output of that fast formula (payment_amount) must be linked directly to the pay value.

3. Define an element link, and happy testing.

Yes you can change the names of the input values, since a mapping is foreseen by the payroll engine, and also the fast formulas can be used as a template for your custom solution, although i would not advise that.

Grossup only works from an element entry, so you can not activate it from an indirect run result.

result name	type	input value	severity
ADDITIONAL_AMOUNT	Adjust	Additional amount	.
CHANGER	Adjust	Changer	.
GROSSUP_BALANCE	Adjust	Grossup balance	.
HIGH_GROSS	Adjust	High gross	.
LOW_GROSS	Adjust	Low gross	.
MESG	Message	_	Information
REMAINDER	Adjust	Remainder	.
STOPPER	Stop	_	.

BEE Spreadsheet Interface

We wanted to load BEE batch element lines via Excel. And we remembered somehow it was described as new standard functionality a few family packs ago.

So just search metalink and there we go. At least, that was the plan. Metalink came with a lot of results, pointing to web ADI, error messages and how to work around them, but a simple doc that describes how to activate the functionality was not found.

I guess Google set the search standards too high.

But ok, a new try the next day on some slightly different keywords, and all of a sudden "BEE Spreadsheet Interface" docs showed up within the top 40 list. So that was the magic search string.

And from that moment on we found some docs, and one hour later a first xls (sorry spreadsheet) was loaded.

Before you ask it: note 341427.1 did the trick.

apps password

today the customer asked to work out a fairly easy forms personalisation. Trouble was that we were not getting the apps password on the development instance. Some pressure, some extra decibles, and we were delighted to receive the very secret apps password.

By the way, on production the password is simply apps. Go figure out.

Documentool for Alerts

In the previous post i mentionned the Belgian summer. Sorry about that.

We just created a Documentool for Alerts. So if you are in need to document all your alerts, leave a comment over here, and i'll send you the code.

Discussions are ongoing, but we think to make it freeware, as Documentool for Flexfields.

HRMS Release 11i Family Pack K Rollup 2

The summer is almost there (here in Belgium), and a new family pack is coming. Sorry, roll up.

One of the nice new payroll core features is the possibility to handle payments to foreign bank accounts. Dev made it now possible to select organizational payment methods from an other legislation then the one of your business group.

And i have the impression this is only implemented in framework, so i wonder how the form will behave.

Oracle in the blogging community: they don't get it yet

In the eBusiness suite corner, i'm not aware of good blogs from Oracle, except the tech stack blog of shan.

With a good blog, i mean a blog where personal opinions are expressed in an open, transparent way, by key players. For that to happen, Oracle is way too closed and hierarchical. Marketing and legal are close followers of all expressions.

First thing to happen is an internal cultural shift to openess. Inviting the blog community to their events can help that transformation to happen. I'm sure that the top bloggers will not all follow the official marketing strategy and message, and will post critical questions. If Oracle internals will then reply, an open communication and exchange can start.

But this will take time, and i'm not sure if Oracle wants to open the windows when the Fusion story needs to be told. A single marketing message is then so important, and the fusion message will not be easy to communicate anyway.

applications unlimited

From a marketing point of view, again, genious. Oracle will support your solution all over the time, if you pay for it.

But the pieces below will be desupported anyway by other suppliers, so you are forced by others to upgrade.

I'm afraid to raise a SR

Disclaimer: i work sometimes directly for Oracle support, so receive money from them.

But on almost every SR i have to pleasure to work on, they drive me nuts. To bring the status of an SR asap back to the customer, they paste in the most fanstatic texts. Or ask for whatever diagnostic information. And ask first for the traces that have nothing to do with the problem itself, but take ages to generate. Or ask for versions, even when you pasted it in already 3 times.

But ok, that is the first layer of analysts you should try to go through. It takes time, but once you now their top 10 of preferred useless info, you can prepare yourself pretty well. Unless you do not have the time (P1).

It becomes more dangereous when they suggest out of context recommendations, that can make the problem worse, or can have impact on other modules. But if you explain why you think it is not a good idea to perform a step, and thus refuse to implement their recommendation, they will not proceed in handling the SR. Catch 22, or better, deadlock.

It would help if all of these analysts had some on-site experience. For all the money customers pay for support from Oracle, this is not asking too much.

Before i end this flow of thoughts, let me stress too that some analysts simply do their job very well, and just play their role as intermediate between customer and development.

After a while, you start knowing them by name. One hint, if your SR is randomly assigned to one of these that think that collecting 10 Giga of data is necessary before contacting dev, try to escalate untill a knowledgable analyst takes over.

retropay, i hate it, i love it

I scanned my blog for retropay. I found back a post of september 2003. At that time we worked out in the Netherlands a solution around reversals. Standard retropay was not an option at that time, since the specifications of the customer were too specific, volumes were huge, and the NL localisation and core solution were not yet 100% finished.

Q: Would i do it again, more then 3 years later?
A: Yes.

Main reason is performance. Retropay is designed to recalculate exceptions. When massive retropay is needed, you can start it very early in the payroll cycle. And run it an other time to process the last minute exceptional cases.

key flex fields

It is not the first time they ask me to "decode" a flex structure. Till now, i havent found a generic way of doing that, but the supported package fnd_flex_ext does the trick nicely for key flexfields. The function get_segs returns the concatenated display value off all segments.

Please drop me a line when you can point me to a similar function for descriptive flexfields.