Methods & Tools - News, Facts & Comments Edition - January 2005

*** Products ***********************************************************

* Product News

You can view the last product press releases from software development tools vendors on our forum

*** Companies **********************************************************

* Christmas Shopping

The last month of 2004 saw many acquisitions in the IT industry.

Oracle finally signs an agreement to acquire Peoplesoft for $10.3 billion. This is the end of a take-over battle started 18 months ago. Oracle declared having a 3 years plan to merger Oracle, Peoplesoft and J.D. Edwards products. Oracle will also eliminate about 5'000 jobs after the takeover, leaving the company with 50'000 employees.

Symantec agreed to merge with Veritas software in a stock deal valued at $13.5 billion, forming a company active in security and data storage software, both for business and retail customers.

Siebel put $115 million in cash to acquire Edocs, an e-billing software vendor. Siebel said that this acquisition would help to answer customer demand for better integration of front-office applications.

The IT mergers and acquisitions' market seems to be a little more active in 2004 than in the 2000-2003 period. Companies are looking for buying opportunities to fuel growth by external acquisitions or to sell themselves for a good price. In the case of Oracle, this is also part of the general consolidation trends that usually leave only one or two major suppliers in mature markets.

------------------------------------------------------------------------

* Delphi meets CMM

Borland announced this month the acquisition of TeraQuest Metrics Inc. This company was co-founded by Dr Bill Curtis, one of the authors of the Capability Maturity Model (CMM). This acquisition should enable Borland to increase the offer of its services division with process optimisation practice. Dr Bill Curtis has become Chief Process Officer of Borland and will work with Borland R&D teams to apply TeraQuest expertise to Borland's product development.

This could seem strange to see a company like Borland, which has built its reputation selling tools for the "lonesome programmer", embracing the process improvement mantra, not exactly the deepest love of code-oriented developers. In these recent years, Borland has diversified its product offering, notably after the acquisition of TogetherSoft. Currently the company gets more than 20% of its revenue from its Design software division and more than 30% from its services division. These two areas have provided the majority of the growth at Borland last year. Considering these facts, and even if services provide less gross margin than software, the acquisition of TeraQuest becomes more logical for Borland.

------------------------------------------------------------------------

* Company News

You can view the last company press releases from software development tools vendors on our forum

*** Numbers ************************************************************

* Training for Testing

The current M&T poll looks at the amount of training received by software development professionals in the area of software testing. Currently it seems that the answer is "none" for most of the respondents. Go to our [Vote] section to disclose your situation and to see intermediate results. The final results will be published in our next issue.

*** Read for You *******************************************************

* The Quality is in the Database

As Fred Brooks described in his famous "no silver bullet" paper, there is no single action in the computing community can take to radically improve software quality. There are however "copper bullets" - lesser steps that improve quality over time. One such copper bullet is the notion of software engineering, the practice of thinking carefully before immersing yourself in the minutia of coding. Judiciously applied, software engineering should improve quality. [...]

Companies now routinely assess vendor software on the basis of cost, functionality, user interface, and vendor stability, but none of these dimensions addresses the software's intrinsic quality. Database quality, on the other hand, could be a litmus test for overall quality. If a product has a flawed database, it is likely to have other quality issues, such as messy programming. In contrast, the quality evident in a sound database is likely to be present in the software's other parts.

Over the past 11 years, my colleagues and I have been evaluating software using database quality as the basis for product grading. We have found that reverse-engineering a database can help a company deeply understand the associated product. Moreover, the time to do the evaluation (sometimes only a few person-weeks) is trivial compared to the millions it can cost to buy and deploy the application. [...]

This experience prompted us to look at additional databases, after which we decided that database reverse engineering was not an odd technology, but something we should routinely perform. We started keeping records of our experiences and have amassed 11 years of data in a grading table (available in its entirety at www.modelsoftcorp.com). I believe our results represent broad practice, given that a different team prepared each database. We evaluated databases only from developers we did not advise as part of our consultant work, and we included databases only for applications that a vendor actually completed. [...]

Many articles tend to give reverse engineering a sinister image, implying that developers typically use it to re-create a product. In all our case studies, we made it clear that this was not our goal. Besides being unethical, reimplementation is usually uneconomical. Instead, we assured vendors that our focus was to assess the software's merit, to get past hidden assumptions and the sales claims and to gain a deeper understanding of the product so that we could better communicate with the vendor and use the software more effectively. In short, when we assess products, we are merely trying to determine what the vendor is selling.

When we reverse engineer a product, we openly ask vendors for their database structure and tell them why we want it. If they refuse, we tell them we will penalize them in the evaluation. We performed many of these reverse-engineering case studies under commission from large companies. Large companies emphatically do not want to devote their resources to re-create a product. Commercial software is important to them, but it is incidental to their primary business. Otherwise, they would be writing their own software, not purchasing it. [...]

Database quality is undeniably a good indicator of application quality. I have found that business leaders welcome the insights gained from database reverse engineering and use them to make more informed decisions about purchasing an application. More important, the benefits have the potential to ripple into the entire computing community. If vendors improve, in-house software development will follow suit. [...]

Source: Michael Balha, "A Copper Bullet for Software Improvement", Computer, February 2004

This seems an interesting starting point to check the quality of applications before buying them. It is true that if the development team did not take some time to design the database, there are few chances that the code was architected.

*** Books **************************************************************

* Integrating Agile Development in the Real World by Peter Schuh

In our last issue, we presented three articles on the agile approaches. This book provides excellent material for a transition from a traditional approach to an agile method. The book gives only a brief description of the agile methods (XP, Scrum, FDD, etc.), but you will find a detailed presentation of the best practices common to agile approaches. For each of them, the author exposes the purposes, the prerequisites, the implementation, the opportunities and obstacles.

The books provides an agile treatment of many of the daily problems of software development projects like database management, data conversion, test data management, project communication, documentation, end-user contacts or developers management. The most interesting point for me in this book is that the author recognises that you cannot always start with a white page and require an "all-agile" process. It provides information on how to integrate gradually agile practices in a traditional software development context.

Click here get more details on this book or buy it on amazon.com

Click here to get more details on this book or buy it on amazon.co.uk click

*** Conferences ********************************************************

* XP DAY TORONTO 2005 www.xpday.info

February 19, 2005 Westin Prince Hotel, Toronto, Canada

XP Day is a one-day event especially for those who want to get started learning about Extreme Programming from the best minds in the community. It has a little something for everyone, whether you're:

• An executive wondering how XP can improve the contribution your software organization makes to the bottom line,
• A manager looking for ways to increase your team's effectiveness,
• A skeptic, hoping to challenge the most experienced and outspoken XP practitioners,
• A programmer itching to experience a real XP project room,
• An experienced practitioner, hoping to share ideas with other experienced practitioners.

XP Day includes tutorials, a project room and the Open Space workshop "From the Fringes to the Mainstream"

J. B. Rainsberger and Ron Jeffries invite you to join us in OpenSpace at XP Day Toronto on February 19, 2005 to explore what is happening in the software development market and in agile software development today. Come to network and swap ideas, and join with our invited experts to influence what happens on a local scale.

Register before February 1, 2005 to take advantage of a 16% discount. Go to www.xpday.info to register today.

------------------------------------------------------------------------

3rd Canadian Agile Network Workshop: "User Interaction in Agile Teams"

The Banff Centre Banff, Alberta, Canada, March 13-14, 2005

Building on the success of the first and second Canadian Agile Network Workshops, this workshop will examine Usage-Centered Design and testing in the context of Agile Methods.

Feedback and evaluation from participants will be provided on the CAN-RAC website following the conference.

See you in Banff!

http://www.agilenetwork.ca/ws2005