CLDR Sample Exam Question

[Roy_F]

I'm a manager in LabVIEW R&D, but have been involved in the recent revamping of the Certified LabVIEW Developer - Recertification exam (CLD-R), so will submit my thoughts in response to the many posts that were recently circulating here and on info-LabVIEW. I mostly want to answer what we're trying to do, how we're doing it and suggest ways for you to help make it better.

Certification, in general, personal opinion:

Professional certification programs are no different than more conventional certification programs, like a high school or college diploma. The issuance of the diploma is a statement by the certification authority (the school) that the person has demonstrated the minimum level of expertise or accomplishments required to obtain the certification. You can make many assumptions from this, like assuming that anyone with a high school degree can read and write (6th grade, 8th grade, not sure what this is, but it isn't 12), they can add and subtract, make change at a register, follow medium level directions or processes, etc. Someone with a college degree can read and write at a higher level (10th or 12th grade level, again, not sure), do basic research, follow more complicated instructions or create medium level instructions, etc. Are all high school graduates equal - no. Are the ones with the higher grade point average always the best worker (however you want to define "best") - no. I'm certain there is a strong correlation between grade point average and performance, but there are no guarantees and there are always exceptions. One of the best software developers I ever met had a Chemistry degree. Some very successful people did not graduate from college, contemporary examples being Bill Gates and Michael Dell. But they're the exceptions, not the rule...

Certification Questions

We heard the rumblings from the field that our exams had "issues", so Customer Education decided to rework them, with significant involvement from several departments within the company. There was much discussion around content, pass/fail rates, "new" feature questions, etc. We created categories of questions, allocations for each category and solicited questions from R&D developers, Systems Engineers (SEs), Application Engineers (AEs) and Customer Education folks. We filtered and massaged these questions to come up with over a hundred exam questions and many example questions. We then had internal users take the tests for a pilot run, looking at pass fail rates for the exams and specific questions. When satisfied all was well, the new exam questions were released.

In general, we expect to have a high pass rate, but not 100%. We expect very high pass rates on some core questions (e.g. loops) and not quite as high on others (e.g new features). We want a broad coverage of features and topics, which means we can't have too many on any one topic. You should have to maintain knowledge of the product and its features lest your level of expertise drop below the threshold for recertification. Do you have to know or have used the most current release to maintain certification? No, and we guarantee this by putting at most two questions (of 40 total) related to New features from the last release in the exam. A specific area has at most 10% of the questions (i.e. 4/40), where an area is something like loops or debugging or error handling. Since you need only get 28/40 to pass the exam, there's some room for deficient areas and/or a few specific misses (like Feedback nodes - don't want to use them, don't want to learn about them - OK, you'll still pass). Perhaps you never deal with File I/O, or you do a lot of File I/O, but always use spreadsheet files and never use TDMS. It's OK if you miss a TDMS question, it won't keep you from passing. "Should" you know about feedback nodes and TDMS - we contend yes, but ensure there is some amount of wiggle room for "world of many colors". The more you know of the product, the better your decisions will be. If you are not aware of the existence of a feature or function or tool or don't understand how they work, then you will not consider whether to use them when it would be appropriate to do so.

Should developers using LabVIEW 8.5 keep up with what is available in LabVIEW 8.6 and 2009 - absolutely. Should they be able to pass a recertification exam without keeping up - yes, but only for a while. Eventually (3 or 4 years, 6 years?), not keeping up with changes should result in a real challenge to keeping certification. The test has to be hard enough to reflect some minimal level of ability and it needs to be broad enough to reflect some general assumptions, albeit no single assumption might be correct. For example, we have questions on LVOOP. Assumption - Everyone who is certified knows LVOOP. <annoying horn sound> Wrong! The test has questions about loops. Assumption: Everyone who is certified knows how For Loops and While Loops work. <pleasant sounding bell> Correct !

 
Issues with Questions:

As for the certification exam itself, please feel free to submit issues with specific questions to certification@ni.com. State the exam, the question as best you can remember it, and the issues you have with the question. Some suggested adding he ability to submit feedback as a part of the exam itself. Our tools support this, and that is exactly what is done for the pilot, but this works best when you receive an immediate response to your answer (right/wrong) and can then submit feedback with that in mind. You could obviously provide feedback without that response, but it is very helpful. My hope is that our reviews and pilot testing will avoid a bad question getting through and our monitoring of the fail rate for specific questions will alert us if one does make it through.

Silver lining to a bad question - bad questions are the ones that spur a lot of discussion on a feature or function. Although the initial focus is on the question being bad, it invariably includes discussion on the feature and many people learn things they did not know before. The enlightening discussion on Feedback nodes being a case in point. We obviously don't want to create a bad question to spur discussion, just making lemonade out of some lemons.

Example Questions:

We readily agree there are some example questions which would not be suitable as exam questions - that's why they are examples and not actual exam questions. The vast majority of example questions are questions we reviewed for the exams, but rejected, either because of some nuance with the question itself or because we already had enough coverage and we kept the ones we deemed best. None of the examples should be wrong, but they might be a bit more controversial than what you should expect to see on the exam. The value of keeping these as example questions is to help users focus on the features and functions they'll be tested on to help them prepare. If you don't know about feedback nodes, you might want to read and play with them a bit before taking the exam. The value of this effort is that you will be more knowledgeable when done and might consider using them in the future as a result, not because they are a cool new feature, but because there are places where they truly are helpful.

Localization:

We're planning to localize the exams for more languages, but resources and costs are impediments. We considered localization when reviewing the questions, so we are hopefully well prepared for expanding this soon.

A significant value to taking an exam is the knowledge you gain in preparing to take the exam. Successful people want to succeed. People who could take the exam and pass with no preparation will still study or practice to prepare for the exam, the fear of failure or not wanting to miss even a single question spurring their efforts. Anything that motivates us to do better is a good thing.

This response is much longer than I had planned - my apologies to those who suffer through it and conclude they wasted their time.

Roy

[NIquist]

Good response Roy, you make some great points.

Sorry if any of us were getting "testy".   

[NIDave]

Well put Roy. 

Your thoughts and views make sense and the feedback node question did exaclty as you intended.  I messed around with them for a few hours after that and, well I made some lemonaid.  Will I use it in my current situation, doubtful but I'm much more aware of its capabilities and should the need arrise I may just recognize that it's the perfect tool for the job.

-Dave

[johnsold]

Roy,

Thank you for the information about the exam preparation process.

One other comment:  Many of the sample exams include platform-specific questions.  This is a situation which is somewhat different from latest version features or items simply not used by some programmers.  A non-Windows user CANNOT have any experience with DAQmx although questions about it have appeared on sample exams.

For the higher level exams (the ones requiring programming) what are the options for a Linux or Mac user?  Trying to take a major exam on an unfamiliar platform is a significant disadvantage.

Lynn

Using LV on the Mac since version 1.2 

[muks]

Thanks for the useful information roy..

[StephenGerads]

It is still feedback; feed-forward is something else.

[JoeC]

I just took the CLD-R, and other than my dry erase pen only working half the time, passed with no issues with a grade of a B.

The approach I took was to do the review tests, and determine where my knowledge was lacking.  Using the LV help covered a lot of details that I had not previously known.

I really liked Roy's post.  He did a great job of summing up why the test is there, and what the benefits are of the recertification quiz.  No time wasted in my opinion.

[JoeC]

FYI, been using LV since v2.0 on a Mac!

[Trouble]

One thing to condsider...  Instead of just saying that someone is certified as a LabVIEW Developer, perhaps the title should include the year of certification.  This would give a clue to the person(s) looking at the title, the level of experience/knowledge with the most current version of LabVIEW.  If my title had 2008, then a person looking at my title would have a clue that I may not be the most efficient at version 2011, whereas if my title included 2012, then they can conclude that I am familiar with version 2011.  Then the importance of getting re-certified would be less for some, but they would have to say "I was certifed way back in 20xx" (although this would not be the most beneficial for NI's $$$ for the certification program, but may relieve the stress some are feeling about this program).  The people that want to continue can.  And I think you said it best:

A significant value to taking an exam is the knowledge you gain in preparing to take the exam. Successful people want to succeed. People who could take the exam and pass with no preparation will still study or practice to prepare for the exam, the fear of failure or not wanting to miss even a single question spurring their efforts. Anything that motivates us to do better is a good thing.

I think I will post the bolded section on my wall and in my house for the teenagers to 'hopefully' read and possibly think about. 🙂

BTW - I did pick up on a couple of items going through the sample exams that I was not aware of.  A couple of features that may be useful for me.

Jeff

[SolPS]

Hi everyone,

I'm currently going through the solution for the ATM CLD sample exam but I don't understand how the text file is being read.  Using execution highlighting, I could see that the CLD solution has a 4x4 2D array output while the vi I created had a 4x1 2D array.  I'm trying to parse the 4 elements in each row of the text file into the appropriate control. Also, the old "Read from spreadsheet file" vi was used instead of the new "Read delimited spreadsheet" vi. Can some explain what's going on and how to implement it correctly?  I've attached code and text file screenshots.

Thanks