Friday, March 14, 2008

Government waste

If you want further proof that government bureaucracies are incompetent and mishandle millions of dollars of tax payers' money, then read this:

As a consultant with the federal government, I design customized Help Desk and Asset Management solutions using an application called the Action Request System (aka Remedy ARS). Without getting technical, and in simple terms, it is a development environment that provides a layer on top of a database. Using this tool, I design forms and code to store information and process business rules. There are two approaches to providing a client with a solution: fully-customized or out-of-the-box. Fully customized applications can cost a lot in the short term, as the client has to pay for the development, from the ground up, of a full-fledged application. However, once developed, there are no licensing fees other than for the ARS layer which can be about $50,000 a year (including the licensing fees for the database). The out-of-the-box solution seemingly has one advantage, that is you can install it and work away. However, there are significant yearly licensing fees for the applications, broken into modules, as well as the ARS layer itself. These costs can be as high as $250,000 a year!

Fully-customized solutions can be problematic if there has been no proper development process. The application usually ends up looking like a patchwork of old, out-of-date code coupled with newer code...a series of bandage code holding it all together. However, every piece of it can be modified and cleaned up as per requirements, and an organization's business rules can be incorporated as they see fit. Some division needs a certain set of fields on this form to track specific asset data? No problem.

An out-of-the-box application is only grudgingly modifiable, the idea being that the client should follow the application's business rules, rather than the other way around...and heaven forbid if the organization wants to update to the latest version...any customized work will have to be reapplied, usually painfully.

But the out-of-the-box application can work if there is a proper business analysis performed and if every section of the organization buys into it. If not, it will be hell. This is where government bureaucracies have serious issues.

I have been developing in Remedy ARS since 1998. I have seen dozens of "solutions" and implementations. Some were abysmal failures...most middling so. The common element amongst the failures was the inability of the client to actually understand their role in the process. Governments are famous for this.

For example, a series of upper-level managers (Director Generals and Assistant Deputy Ministers) will sit in a room and discuss what they "need", without actually understanding the "need" since they are so far removed from people who actually DO the work, that their concept of "need" rarely translates into the needs of the department. What these managers need is something to make them feel that they are relevant. So, some smartasses come up with something called ITIL (The Information Technology Infrastructure Library (ITIL) is a set of concepts and techniques for managing information technology (IT) infrastructure, development, and operations). This money making scheme...oops...I mean concept...was created, of course, by one of the most bureaucratized countries on Earth...the UK. ITIL is simply a common framework for IT management...it is supposed to be a starting point for managers to restructure they way they provide IT services to their clients. However, as with all things IT, it has become a buzzword which has excited managers. You say ITIL to a group of managers, and the collective saliva drooling from their mouths could create a new oasis in the Sahara.

Managers in the government love "shiny" things, and since they are not spending any of their own money directly, have no qualms about pissing it down the toilet. They will never be truly held accountable, so they don't care. Sales people know this, and will happily provide rigged demonstrations of their "products" (rigged in the sense that there is only a bit of data...so the application runs smoothly...and no one else is working on the app at the same time...which makes everything really, really quick). Then they flash some nice pie charts, a view fancy reports, some other useless bells-and-whistles, and they will have successfully distracted the managers who, at this point, only see the shiny things. One of the shiniest parts of the out-of-the-box solution from BMC (BMC Remedy IT Service Management 7.x: Application) is their CMDB (Configuration Management Database...from Wikipedia, Configuration Management (CM) is an Information Technology Infrastructure Library (ITIL) IT Service Management (ITSM) process that tracks all of the individual Configuration Items (CI) in an IT system which may be as simple as a single server, or as complex as the entire IT department). A BMC technical rep gave a demo before which he mentioned that to successfully implement a CMDB, there should be months of business analysis. Roughly an hour later, a government manager asked when it could be installed. The BMC sales rep, also in the room, almost fainted he got a hard-on so quickly.

As a sop to proper business analysis, they are now paying some out of town firm $1000 a day to do a 21 day gap analysis. 21 days. Amazing.

This is what will happen: after spending several hundred thousand dollars, if not millions when you take salaries into account, the application will be put into production. People will complain that it is too slow, especially those in missions, like some countries in Africa which still use 56K lines; tons of customizations will be made; the application will only be partially used, not at all justifying its expenses; and then the managers who perpetrated this fiasco will be reorganized out, suffering no penalty for their stupidity. The team responsible for maintaining the app will take all the shit. Then they will have to upgrade to the newest version of the application, since the old version will no longer be supported. They will have to spend hundreds of thousands of dollars more to reapply the customizations...and the circle continues.

****************Update March 17, 2008****************
The consultant brought in to do the gap analysis is already 2 days behind the original time estimate of 21 days. He's been here for 4 days :)