Certification Proposal

[chet hendrickson]

Hello All,

This is an outline of the program Ron and I have been contemplating.

We have found that in order to achieve excellence a technical member
of a Scrum/Agile team one must obtain experience and knowledge in
seven areas:
• Product
• Collaboration
• Business Value
• Supportive Culture
• Confidence
• Technical Excellence
• Self-Improvement

While each technical role will demonstrate these values in a different
way, they represent a common path to achieving a level of
craftsmanship that is far beyond basic competency. The demonstration
of this sort of craftsmanship is the key to creating a certification
that will be embraced by the developer community.

A developer wishing to be certified at the Certified Scrum Development
Practitioner level will be assessed by an authorized auditor. The
auditor will not only observe the candidate, but work side-by-side
with them for between three and five days. After which the candidate
will receive a radar chart scoring their level of achievement on a
scale of 0-10 in the seven skill categories. The results of the audit
will also be published on the Scrum Alliance website.

In order to be eligible to receive an audit, the developer must first
successfully complete a five-day gateway course. The gateway course
will demonstrate the importance of the seven areas and provide the
participant with the means of charting a course to obtain practical
experience in them.

To signify their entry onto this learning path, the developer will
receive certification as a Certified Scrum Developer. This
certification will expire unless superseded by the CSDP within no more
than four years.

--
Best regards,
Chet Hendrickson mailto:ch...@hendricksonxp.com

[Jim Shore]

I agree with the general thrust of the proposal. I'm still uncertain whether public evaluations of any sort are a good idea. And the more I think about it, the more I'm convinced that this idea is only as good as the evaluators, and that as a result its reliability and quality will degrade due to pressure to field more evaluators.

One suprising bit of feedback I got from Eric Lefévre-Ardant (I think) was that the Scrum certification is North America-specific. He was quite bitter about it--he felt that the Scrum certification created a bias in the market towards people with certifications and that, since there were no French CSTs, it made it harder for the French to compete with foreigners. His argument could easily be extended to any group that is qualified but biased against certification for some reason... and those groups include the highly qualified but persnickety developers who naturally distrust certification but in fact could be our greatest resources.

I also heard concern from Ward, as well as Eric, that any sort of certification program will stifle innovation.

Ward also mentioned that we already have a sophisticated certification program in place that has spent far more time and money on this than we could... university accreditation and degrees. If they can't get it right, what makes us think we can? If they can get it right, why are we inventing our own?

Now, for some specific suggestions:

- Rather than using a scale of 1-10, use the "Shu Ha Ri" scale. By using a less precise scale, we're more likely to have consistent evaluations. Plus, "Shu Ha Ri" says worlds about what we're looking for, which (for me) would make evaluations much easier.

- Strongly desire to remove requirement for gateway course prior to evaluation.

- Recommend that the evaluator not have any prior association with auditee, particularly not training or financial.

- I continue to be intrigued by the idea of only evaluating teams, not individuals. Although the Shu Ha Ri scale seems more appropriate to the individual, somehow.

Jim

--

James Shore - The Art of Agile

recipient of Gordon Pask Award for Contributions to Agile Practice

co-author of The Art of Agile Development

voice: 503-267-5490

email: jsh...@jamesshore.com

blog: http://jamesshore.com

[chet hendrickson]

Hello Jim,

Thanks, this is very good input. The Shu Ha Ri also resonated with
me. It seemed 'right' as soon as I heard it.

There is much here to think about, I am sure you will see it reflected
in our refinements as we move forward. If in fact, forward is where we
should move.

chet

Check out our upcoming CSM Plus courses @
http://hendricksonxp.com/index.php?option=com_eventlist&Itemid=28

[Elisabeth Hendrickson]

I really like Jim's Shu-Ha-Ri suggestion. 

And I totally agree with "Recommend that the evaluator not have any prior association with auditee, particularly not training or financial" if the evaluation is done outside the training. So a consultant would not certify ppl at a client site as part of some other consulting engagement.

But I'm not convinced of the practicality of doing audits.

I've been thinking about scalability. I've heard estimates of the number of CSMs minted ranging from 20K CSMs certified to 50K. If we're talking that kind of scale, as desirable as one-on-one evaluation is, I just don't see how to scale up the number of auditors to meet the demand.

Further, the cost of the evaluation would be so high that it would be a significant barrier for most companies & individuals. I believe that certification schemes succeed when people feel pressure to get a certification "because everyone else has it." If the barriers to getting it are *too* high, the certification won't catch on. And that would leave the door open for something evil to take over the market.

So I have a suggestion:

What if, like the current CSM, you get the certification by attending the class? BUT, unlike the CSM, performance in the class affects the results.

So imagine there's a 5 day class. 

In class, we spend mornings on the content, and afternoons working on a realistic project. 

That means they'd be writing real code to implement realistic stories in a simple enough domain. Perhaps there would be some starting code in place. The goal would be to have the project be something that any group with members who have rudimentary development skills *should* be able to succeed in delivering. Like my WordCount simulation, the complexity would not be in the system being implemented but rather in the interactions between the members implementation team. 

So...if the team manages to deliver a project that meets the needs of the "customer" (played by the instructor/facilitator with standard guidelines in place for what constitutes "acceptable"), then everyone who participated in creating the final result is certified.

Active participation of some kind is mandatory. If someone consistently skips out - plays hooky - they will be warned and ultimately will be exempted from the class. One of my gov't clients does something like this: if a participant fails to show up for at least 80% of the class, they do not get "credit" for the class in their HR file.

In addition, instructors can award a "With Honors" designation to individuals who distinguish themselves. Guidelines for awarding With Honors would be based on the seven areas y'all already developed to help ensure consistency. And the record of someone's achieving Honors would include the name of the instructor who awarded it.

This suggestion has 3 important implications:

1. The team as a whole succeeds or fails. They become keenly aware that they will not be certified if they can't figure out how to work as a team. The instructor may have to coach team members who don't play well with others. But in general, the team -- whatever random collection of people end up in the class -- will have to demonstrate that they can collaborate. I expect that the vast majority of classes will succeed & be certified - and I also suspect that for some who don't know how to work well with others they will gain deep and necessary learning about how to do so that will make them better Agile developers.

2. The barriers to getting the basic certification are low, increasing the likelihood that this certification scheme will gain popularity and thus will ultimately meet the market demand for a certification. 

3. There is room for individual recognition to those who deserve it, and that recognition is based on the individual's performance in a realistic situation. So, it's not as good as a one-on-one assessment, but I think it's more scalable.

Whadya think?

Elisabeth

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

Elisabeth Hendrickson

Quality Tree Software, Inc.

p 925-426-9726

c 925-989-9634

http://www.qualitytree.com

http://www.testobsessed.com

[Sebastian Kübeck]

I think this idea is great, however, picking out people from a team to
award or punish isn't really in the sense of Agile or Scrum, is it?
How about an up front multiple choice test to cover basic technical
and agile skills to get rid of the people who aren't motivated enough.
Then, you do a very simple five day project as you defined it.
I think this is close to what CSM certification is now.

Sebastian Kübeck

> c 925-989-9634http://www.qualitytree.comhttp://www.testobsessed.com

> >> Chet Hendrickson                          mailto:c...@hendricksonxp.com

[mheusser]

>I believe that certification schemes succeed when people feel
>pressure to get a certification "because everyone else has it."
>If the barriers to getting it are *too* high, the certification won't
>catch on. And that would leave the door open for something
>evil to take over the market.

I know I'm late to the party here ...what is 'success' for the
certification?

The best certs I can think of come from the Martial Arts, where there
are mutliple disciplines, and multiple schools and certification
authorities for each discipline.

To take that back to SwDev, Joe might be a certified ScrumCoder and
Sally might be a certified XPCoder and Jim is the lead committer on
the OpenSQL Open-Source project - all three might make compelling
candidates.

Of course, this defines 'success' as something other than a monopoly.

I honestly don't know the goals of the Scrum Alliance on this. I'm
open to contibuting along the lines desired if I can, leaving quietly
if I can't.

regards,

Matt Heusser,
CSM, Member of Scrum Alliance,
Honest And Serious Software Craftsperson (TM), etc.

[jediwhale]

On Aug 29, 12:44 pm, Elisabeth Hendrickson <e...@qualitytree.com>
wrote:

>
> I believe that certification schemes succeed when people feel pressure to get a  
> certification "because everyone else has it." If the barriers to  
> getting it are *too* high, the certification won't catch on.

I believe that certification schemes *fail* when people feel pressure
to get it 'because everyone else has it'. If the barriers are too
low, it becomes another meaningless certificate to pin on the wall. A
certification should require some hard work and commitment to get -
that's the kind of filter I want to apply to prospective hires. Think
MD, PEng, even journeyman electrician, not MCxx.