HR-1 was officially retired over two weeks ago, and we are still going with Seeker. I thought you might like to know a few of the differences.

Seeker is a web-based application which, in our case, has an Oracle back-end. There are about 150 normalized tables holding essentially the same information that was in HR-1. Almost everything is delimited by date effectivity. This makes it possible to enter future transactions quite easily. Unfortunately, it also makes queries for current data a bit more complex because effective date information must be used with all queries.

Views and updates are roll-based so that security can define which roll can see which information. However, unlike HR-1 where security is typically defined on a field or row basis, roll-based security takes into account the relationship between the person doing the query/transaction and the person/org/whatever represented in the data. So, for example, you cannot see salary information for your peers.

Eventually, the product will be rolled out to the entire company so people can update their own personal information and managers can update information regarding their direct reports. For now much of this will still be updated by the service center.

Canned reports are generally faster in Seeker than in HR-1 even when joining 20 tables together is necessary. However, the complexity of the database makes ad hoc queries much more difficult.

Updates and navigating through data screens is very slow. The norm is for transactions to take 10-20 seconds.

Architecturally, all this is managed through a multilayer approach, with a user layer, and application layer, and a database layer. In order to achieve the current performance, five (count them) application servers are used. This week employees in the IT department were notified of the availability of the application for making changes to employee information. Shortly after the email notification went out, 4 of the 5 servers went down due to the heavy load. (an interesting side note is that the one server that remained online was the most recently added server – it's previous duty was as a staging server for HR-1)

Many users continue to use HR-1 for queries because the data is not readily available in Seeker. However, with appropriate security definitions, this problem should eventually fade.

Anyway, although this is not specifically Arev related, it does demonstrate one of the possibilities of converting away from HR-1.

View this thread on the forum...