Re: Bombardier Recreational Products picks Dell for revamp (Jan. 11)

What is the point of this story? It seems that is really only providing free advertising for Dell!


suggests that after “a white sheet approach” that they decided on an end to end Dell solution.

By only concentrating on the details of the Dell deal, we are offered no information of who else was considered, or what criteria was used. We cannot use this information to avoid similar pitfalls or make better decisions. All I see is a testimonial that I would expect to read on Dell’s own Web site.

This type of reporting offers no value to our company or our clients , and suggests that Dell is the answer.

I am greatly disappointed in this, since Dell is the largest player in the market already and really doesn’t need blatant free advertising. The market is extremely competitive, and it’s hard for the thousands of small companies like ours to vie for contracts and opportunities as it is. Please consider that it many cases Dell is not the only answer and that by promoting the perception that they are you make it harder for us to survive. Reporting should be unbiased and complete so that the reader receives the whole story. Why was Dell picked? Why were others not? What were the key deciding factors?

I am not anti-Dell, and we have used Dell products in many of our projects. The bottom line is Dell doesn’t support us is any way, except that they do supply solid products at competitive prices. Since we are always looking out for our customers’ best interests, we do often include them. Many other manufacturers support us in many ways, but they don’t always have the answers or products that we need. It goes both ways.

I guess that I am concerned that “Dell wins another big contract” is worthy of “Breaking News.” Is there a message I the story? Not really (other than use Dell). I’m sorry, but it touched a nerve!

Don Smith
Pro-Tech Network Solutions Inc.

Shane Schick responds: Typically when we ask enterprises about how they chose their vendors, the answers are the same — they go through a competitive bidding process where they examine cost, the vendor’s expertise and the quality of their proposal. This case was no different, though as Deslauriers mentioned, the firm’s previous relationship with Dell probably helped somewhat. The story was not really about Dell but about Bombardier Recreational Products, how it is upgrading its desktop infrastructure and why it needed to achieve consistency across its operating systems. When there is a unique process around vendor selection or usual attributes that led to a relationship, we always highlight those details in our coverage.

Re: The future is Wile E. Coyote (Jan. 10)

We need to take a lesson from Wile E. Coyote’s experience with the fine work done by the QA team at ACME, maker of Wile E.’s electromechanical aids. Seegrid’s robots will probably be programmed by recent community college graduates working under time constraints for the company which put in the lowest bid. Then we live with the erratic results forever, or until the robots rust out, or their operating system is no longer supported. Hans Moravec needs to read “A Hitchhiker’s Guide to the Galaxy.” Get Sirius.

Robert Cosh
Fisheries and Oceans Canada

Re: Canadian teen sentenced in Randex virus case (Jan. 5)

Sophos security analyst Gregg Mastoras is quoted as saying that “”he’s not out for profit or to damage large corporations. It was more for his own personal knowledge”” and “”Most of the time it’s done for nothing other than bragging rights . . . They don’t realize who they’re hurting or what’s happening until it’s actually shown to them.”” Either Mr. Mastoras is unbelievably naive or he just doesn’t know much about teens. It is impossible that these young hackers aren’t perfectly aware of what effect they’re having — after all they have access to all the news on the subject, just as we do; and if bragging rights are at stake then the more damage they can do, then the more prestige they will get among their fellows.

Nine months probation with no restrictions on his Internet use — what sort of deterrent is that? Putting him to work as a data entry clerk until he’s paid off some portion of the damage he’s caused would bring it home to him. If hacking of this sort is considered undesirable and worthy of criminal investigation then the perpetrators need to be made to understand this! (I won’t say slap on the wrist, I won’t, I won’t . . .)

Peter Lacey

Re: Canadian teen sentenced in Randex virus case (Jan. 5)

How many times have you heard this . . .

The damage caused by such “fun-loving kids” is widespread, costly to those infected, perhaps dangerous for those in wheelchairs, disabled or handicapped — those who rely solely on their computers to enjoy the outside world.

And the courts let him off! No computing ban, no naming, miniscule probation, don’t want to spoil his future plus have the audacity to tell me he seems like a “nice kid.”

Graham Hobbs

Re: The future of prediction (Jan. 5)

We read with great interest your recent article.

Rather than group all of the technologies offering predictive capabilities together and dismiss them as “assumptive,” why not take the opportunity to educate your readers about the benefits of the combination of people, process and technology — specifically, some of the benefits of predictive technologies when combined with appropriate skill sets and well-thought-out processes?

First, one needs good data to support good analysis and therefore good decision-making. You allude to this in a few places: where needs are extremely well documented and understood; the problem with factoring in (or rather not) the intangibles, etc. The best data are not always available — some data are simply not collected or are problematic to collect. The data may be and often are “dirty” at the outset. Regardless of whether the data are entered by hand or electronically generated, error is inevitable. Numbers are transposed, sections of forms are left blank, etc. But, this is the world we live in and the question arises: Just because we have imperfect or lacking data, should we still not make the most of the data we have?

Second, when building predictive models, the goal is often to have a good model that generalizes — one that performs well on new data. Say we build a model to determine the likelihood of a credit card transaction being fraudulent. To do this, we use cases of known fraudulent and legitimate transactions so that a new transaction can have a likelihood associated with it and that in assigning that likelihood we will be making a good “statistical guess” as to the possibility of fraud. It is not possible to cleanly assign all transactions as 100 per cent fraudulent or 100 per cent legitimate — some transactions will have attributes that may make them appear to be fraudulent (e.g., a cash advance transaction at a casino), but are legitimate and vice-versa. Setting an acceptable threshold is where careful human judgment is needed. If we build a model that is “suspicious” of every transaction, we run the risk of offending customers; if we build a model that is too “trusting” we tolerate more fraud than we should. Setting a wise threshold is far better than doing nothing or tossing a coin.

Assumptions, unpopular as they may be, are more often than not necessary in decision-making. Example: Exploratory data analysis reveals a spike of 93-year-olds in the customer database. Closer inspection reveals that almost all of these customers have a common birth date, November 11, 1911. “11/11/11” is the date format and how the dates were entered into the system-this is the easiest-to-enter, valid birth date. The integrity of these dates (and therefore age for these customers) is in question-birth date is treated as a required field. Assumptions have to be made to deal with this data and can vary depending on the task.

But perhaps your intent was to take aim at vendor hype (e.g., “no thinking required — just push this button . . .”) around predictive technologies. If so, please consider this reinforcement of that effort and let us go on record as saying that predictive modeling technology is an enabler — those who use it wisely (human thought required) reap the benefits.

Anne Milley
Director Analytical Strategy

Re: Software skills overtake hardware on top salary list (Dec. 7)

Your article missed one important point: when firms “”downsized from 100 to 15,”” those 85 people left the profession. Finding “”senior”” software skills is now near impossible.

H.A. (Rick) Stomphorst
VP Operations
Incubed ltd.

Letters to the editor must include the writer’s name and company name along with an e-mail address or other contact information. All letters become the property of Editors reserve the right to edit submissions for length and content.

Share on LinkedIn Share with Google+