|
|
|
|
The Value of Open Process Standards
by: Thomas Meloche
Proprietary Development Processes Software development organizations often realize, after a few spectacular failures, that they require a more formal approach to software development. Often they may assign one or two individuals the task of creating their own process and sometimes they go to large consulting companies to get help. Happy to oblige, many of these large consulting companies expend a great amount of time and effort developing their own internal proprietary development process which they offer as evidence of their programming prowess. Attempting to avoid failure is a laudable goal, but this approach suffers from multiple failure points. These include:
This paper will review these concerns and propose an alternative to creating or using a proprietary development process. Failure Point 1: The Uniqueness of the Process A proprietary and internal development process is by definition non-standard and unique. It likely has its own vocabulary, rules and definitions that are not used anywhere else in the industry. Is it a good process? Who knows, as it has never been open to public scrutiny and evaluation. A unique and private process is difficult to evaluate. Case studies are not public. Criticism is silenced rather than requested. I am aware of one large consulting company that had 35 people off by themselves doing product visioning and requirements for an e-Commerce exchange start-up. It was a waste. It was a bad process. Proprietary methodologies result in constant special training and retooling as new team members are added. And because the knowledge is so specialized, those who know and understand the process often charge exorbitant rates. Words to heed: "Caveat Emptor" (or Buyer Beware). It is our observation that big consultancy processes seem more designed to spend your money than deliver your software. Failure Point 2: Dearth of Complimentary Documentation Proprietary processes have limited documentation. If a participant does not understand the process adequately from the training or the original set of manuals, there is nowhere outside the organization to go for help. Experience tells us that it often takes three or more exposures to new information before that information can be internalized effectively. Internal proprietary processes often only provide one or at best two exposures to that information. Failure Point 3: Limited Training on Processes A process proprietary to an organization also has limited training opportunities. It is a matter of record that internal training initiatives are always the first to suffer in budget allocations and corporate restructuring. Also, since training is an expense that doesn't generate direct revenue, it is typically one of the least respected internal initiatives. Proprietary internal processes suffer from a lack of formal training options. Typically there is only one class or curriculum to choose from to learn the process and if it is poorly developed and conceived there are few remedies. Failure Point 4: Narrow Experience of Process Developers The narrow experience of the process developers results in fatally flawed processes that are incomplete, inconsistent, and do not fit the needs of many projects the teams must deliver. Large organizations do hundreds to thousands of different development projects over the life of a development center. A process developed by a small group of internal process developers is not likely to be large enough in scope or vision to encompass a large organization's needs. Failure Point 5: Lack of Supporting Tools Internal proprietary processes often lack the supporting tools necessary to automate and implement the process. Therefore externally provided tools must typically be used. Since they were developed without the knowledge of the proprietary process, they are often a poor fit for the prescribed process workflows. Of course, if the tools are developed internally, they will likely suffer from the narrow experience and incomplete process on which they are based. As a result of these limitations, proprietary processes often end up partially implemented, ignored or abandoned. It is reasonable therefore to conclude that these types of processes should not be the choice for your organization. Open and Standard Processes Fortunately, there are ready alternatives to proprietary development processes including, The Rational Unified Process® and Extreme Programming. These processes are open, published and supported standards. We feature a more detailed description of these processes in our papers: The Rational Unified Process® (RUP) and Extreme Programming. If you are interested in learning more about these development processes consider attending one of the process overview classes provided by The Menlo Institute. We have extensive classes such as Managing Projects with RUP-Lite and Introduction to Extreme Programming that may be of interest. A list of all classes are outlined here. ------------------------------------------------------ Interested in Learning More?
Tom, send me Rich's free white paper "Secrets of Software Success: Adapting Projects to an Accelerated Society" We will not resell your contact
information. Privacy Policy: No Resale Zone. Spam Policy: No Spam Zone. |
|
||||||||||||||||||||||||||||||||||||||||||||||
|
Menlo Innovations |
||||||||||||||||||||||||||||||||||||||||||||||||
|
Articles | White Papers | Newsletters | Speakers | Links | Reading List |
||||||||||||||||||||||||||||||||||||||||||||||||
|
Menlo Innovations LLC (c) 2006 |
||