Letters to the editor
To comment on articles you've read in our magazine or on our Web site, send an e-mail to our editors.
- A reader responds to “Needed: Technology Investment,” Editor's Note, Test & Measurement World, November 2005.
- A reader responds to “Needed: Technology Investment,” Editor's Note, Test & Measurement World, November 2005.
- Asset InterTech responds to “Moving beyond boundary scan and inspection,” Tech Trends: Manufacturing Test, Test & Measurement World, August 2005.
- LogicVision responds to “DFT drives yield improvement,” Viewpoint, Test & Measurement World, August 2005.
Rick, I am not sure what you are arguing for in the latest editorial [Needed: Technology Investment" (November 2005)] about the 11:1 pay ratio for US vs. Indian engineers: Salary hikes for them or pay cuts for me?
I see that 11 of them might be hired for one of me, but I don’t believe my foreign counterparts to be as well trained or as capable.
Of course, many of my US counterparts have left engineering to manage, start other lines of work, or whatever, since there seems to be a dead end for US engineers about five years into their careers. Salaries become stagnant and benefits woeful (which starts with W, just like another American institution of higher ignorance and greed) while the cost of living goes way up. And then editors like you toss in the bull about outsourcing.
If you want US engineers, money is not all it will take. Gaining the respect offered to doctors, lawyers, and other professionals takes education, licensure, practice, and a bunch more than money. People still confuse me for the guy who drives a train and makes more than I do. All I would like is respect, money commensurate with the amount earned by other professionals, and for people to know that I--like other professionals--am educated, skilled, and capable.
The writer is an EE and registered Professional Engineer.
Name withheld by request.
Dear Rick Nelson, I recently read your editorial titled "Needed: Technology Investment" (November 2005). It's not clear what your particular position is, although the reference to an 1100% salary premium seems to suggest that Americans are overpaid by that amount, and the corporations hiring them don't get what they pay for. If that is true then you've overlooked some key points.
There are thousands of American engineers who do indeed possess the skills, the education, expertise, and experience to compete with engineers from any other country. True, we may not produce as many, but neither are we a country of 1 billion plus people. What you have overlooked is that because engineers are cheaper overseas, corporations have chosen to bypass their American counterparts. They've been laid off in droves and displaced by their cheaper Asian counterparts. Conditions are ripe for corporations to exploit the large wage discrepancy, and in the end America and Americans will pay a larger price than the current wage differential. One such condition is the existence of the H-1B and L-1 visa programs, which has allowed corporations to exploit cheaper wages and displace American engineers. In spite of the language governing the H-1B program attempting to prevent such exploitation, it is routinely done. You need only visit any on-line database of Labor Condition Applications (LCAs) to see that the wages offered to holders of H-1Bs are not the prevailing wage.
I have a BS in civil engineering and an MS in computer science. I was laid off from Network Associates and was unemployed for over 13 months. Five H-1B employees (three Chinese and two Indian) were kept in an office of 20 while five Americans lost their jobs. I eventually found a position at two-thirds my former wage (60% if you include medical and other benefits). I have friends and colleagues who have been unemployed for as long as three years. In all cases, those lucky enough to have found a job are working at a lesser wage.
So what is it you're advocating? So what if we put all the necessary infrastructure in place to produce more mathematics, science, and engineering graduates. Do you think anyone would actually pursue a curriculum that is comparatively difficult to graduate at a wage that competes with that of third-world nations? Is the lack of opportunity in your chosen career and at a lesser wage supposed to be encouragement?
Certainly any job that could be done at a computer here can be done much, much cheaper overseas if corporations are allowed the opportunity to do so. That not only includes the long-gone manufacturing sector but also call-center, IT, computer science, R&D, technology, financial, medical, legal, accounting, etc. Would you argue as these careers flee the 'over-inflated American wage scale' that we ought to offer some educational retrospective rather than the simple fact that they've fled because we've allowed them to?
The H-1B and L-1 programs and offshoring have severely cut into the fabric of the American economy. Why? Because they're ripe for abuse. The engineer is a vital thread for a productive and vibrant economy. If the current trend is allowed to continue America will surely find itself in the position of the third-world nations to which these jobs have been so generously offered by our corporate lobbyists and our corrupt politicians.
Unfortunately, the H-1B and L-1 programs and the offshoring effort have gathered so much momentum that it will be difficult to stop. I have already seen jobs advertised that now require fluency in Chinese and Korean, which are no doubt offered to the H-1B and L-1 recipients working here in America. So it is now even easier to exclude the American engineer from consideration. I had two separate interviews at Motorola and was interviewed by a Chinese woman, three Chinese men, and an Indian woman. Not a single qualified American in house?
Of the 65,000 H-1B visas snapped up by corporations this year, do you honestly believe that not a single American was available or qualified for any of these positions? If there is a shortage of qualified personnel to fill these positions, then why are wages being driven down? Why is unemployment in the computer-science field greater than the national average? Why is the Senate under the guise of budget-deficit reduction allowing an increase in H-1Bs? Why are L-1 visas allowed with so many unemployed American engineers?
Maybe I'll just have to put in 12 times the hours of my Asian counterparts to earn my 1100% wage differential to pay for my $800,000 2200-sq-ft, 4-bedroom home on a 9000-sq-ft lot and $13,000-per-year taxes.
Sincerely,
Joe Stepien
Dear Editor:
In his column, “Moving beyond boundary scan and inspection” (August 2005), Steve Scheiber made several very valid points, including his assertion that test engineers “must think outside the box.” Unfortunately the box he placed around boundary scan was severely misplaced.
Steve asserts that “contract manufacturers have to get away from the notion that a single test strategy can address every situation.” You might think that as the supplier of a certain test technology, we would disagree. Far from it. We have been telling our boundary-scan customers for many years that no one test technique--not boundary scan, not in-circuit test (ICT), not optical inspection, not any--is a panacea. In fact, this belief has driven our company’s strategic direction for some time now. That’s why we’ve formed strategic relationships with companies like ICT-supplier Agilent Technologies and the processor-based functional-emulation company mentioned in Steve’s column, International Test Technologies. We believe that the strengths of boundary scan complement the strengths of ICT, in-circuit emulation, functional test, optical inspection, flying probers, manufacturing defect analysis (MDA) systems, and other types of test technologies.
Steve erred when he wrote that boundary scan was once seen as the “ultimate solution” and that now “manufacturers are facing the limitations of boundary scan to verify a board’s function…”
Boundary scan was never a silver bullet. Its limitations and strengths have been apparent since the IEEE 1149.1 standard was first accepted over a decade ago. Boundary scan is a digital structural test technology that can increase test coverage to an extremely high level. Most often, this is done in concert with other test techniques. Although boundary scan is not and was never intended as a functional test technology, it is often used as the point of access for chip-level functionality, which comes into play during functional test.
Steve also wrote that “boundary scan makes boards more susceptible to security breaches.” If this were the case, then any programmable device or the presence of any embedded firmware would jeopardize the security of a system. From our perspective, security rarely, if ever, enters into the discussion when boundary scan is being considered as a test or onboard programming method. In the few cases where security has come up, it has never discouraged a manufacturer from deploying boundary scan as a test or programming technology. One of our customers, a large manufacturer of a video game console, deployed boundary scan to test the first generation of its product during design and development, but disabled boundary scan because of unwarranted security concerns when the system moved into production. Now, the next generation of the same video console is being shipped with the boundary scan port enabled because the manufacturer wants to use it to troubleshoot, support, and service the system.
There is considerable insight in Steve’s point that we must move beyond boundary scan. In fact, boundary scan has moved beyond boundary scan. In recent years, a plethora of new standards have piggybacked on the IEEE 1149.1 standard. Rather than replacing 1149.1, new standards such as the 1149.6 boundary-scan standard for advanced digital networks, the IEEE 1149.4 mixed-signal test bus, the IEEE 1532 standard for in-system configuration, and the Nexus 5001 standard are all built on the original 1149.1 boundary-scan infrastructure. Why? Because it works, and many OEMs and contract manufacturers know firsthand the value and benefits that boundary scan has brought to the electronics industry.
Steve is right. Thinking outside the box is essential. That’s just what the boundary-scan industry is doing. We’re building on the solid foundation of a proven and highly capable test and programming technology, yet extending it with new capabilities and innovative ways of applying it. In fact, an article in the September issue of Test & Measurement World, “Boundary scan goes underground,” by my colleague at Asset InterTech, Dave Bonnett, describes how boundary scan is being extended beyond board-level operations to embedded system-level test and programming functions.
Best regards,
Alan Sguigna
VP of Sales and Marketing
Asset InterTech
Dear Editor:
I read with great interest your interview with Robert Hum of Mentor Graphics, “DFT drives yield improvement” (August 2005). In response to your question, “What are the prospects for logic BIST taking on functions that scan and compression now perform?” I understand why Robert attacked BIST, as Mentor has failed in its attempts to develop a BIST product that met the customers' requirements. But what really surprised me was his criticism of the quality aspect of BIST. It is generally accepted in the design and test community that quality is a strong benefit of BIST. In fact, there are various customer testimonials and technical papers that confirm it. Companies including IBM, Intel, AMD, and Toshiba have embraced BIST for 90 nm because it also reduces their life-cycle costs.
Robert also mentioned LogicVision by name with regard to a missed window of opportunity. First of all, we at LogicVision appreciate the contribution Robert made in helping launch the company. His comments on BIST, for the most part, were correct in the late '80s and early '90s, but the environment has changed dramatically in the last few years. As 90-nm production volumes start to build, the associated life-cycle costs are skyrocketing. For this reason, LogicVision is focusing on “yield learning” with BIST as the core technology as the way to improve quality and reduce product life-cycle costs.
Robert claimed that the industry in 2000 or 2001 started to turn its back on BIST. He claimed it did so because BIST no longer had the resolution necessary to test for faults that actually occur in manufacturing; faults such as shorts between two lines. It is quite ironic that, from our perspective, the industry started seriously embracing BIST, if not in 2000, certainly by 2001 and 2002. The window of opportunity was opening. BIST was being adopted for solving major problems due to incomplete testing of factors such as crosstalk, impedance effects of interconnect structures, problems associated with clock structures, etc. Take for example the “shorts between two lines” mentioned by Roberts in the interview.
Making a general statement that logic BIST doesn’t detect defects that require “correlated patterns” and using shorts as an example is somewhat naïve. In fact, this example, in simple terms, is just wrong. The detection probability for bridging faults is generally higher than that for stuck-at faults because there are two possible sites from which the effects of the fault can be observed, whereas there is only one site from which the effects of a single stuck-at fault can be observed. There are two references for this: Touba (ICCD 97) and Waicukauski (ITC 88).
The real issue is controllability, and logic BIST solves this inherently, a fact established over 10 years ago. Waicukauski (ITC 88) has shown that the fault coverage of shorts tracks the stuck-at fault coverage obtained by applying random patterns. The same conclusion was reached for transition faults. The high coverage of shorts and transition faults is due to the large number of faults detected multiple times during logic BIST testing. LogicVision’s technology provides a high degree of controllability and observability of the entire circuit. Logic BIST represents a more practical solution to address all these faults compared to ATPG that would need a very large pattern set to achieve the same result.
There is not sufficient space here to go into to all the reasons why the industry is at a tipping point, tipping from ATPG to BIST. Companies that have sophisticated test and test-coverage strategies are increasing their use of BIST. In many cases, mandates from top management or from customers have been the drivers for the use of BIST. Robert's reply was very skillfully done. Unfortunately, his remarks are false, but many readers may believe him. The best approach to understanding why BIST adoption is so important, and happening now, is to understand the reduction in product life-cycle costs it affords through economically achievable yield learning.
Sincerely,
Jim Healy
President & CEO
LogicVision
To comment on articles you've read in our magazine or on our Web site, send an e-mail to our editors.