In between, the schedule holds a single track of sessions on:

• Tools I: Support and Use
• Processes and Practices
• Management Environments I

Yesterday afternoon’s agile GSD tutorial was interesting and offered a nice preview of Erran Carmel’s forthcoming book, as well as a review of Yael Dubinsky’s HOT framework (humans, organizations, technology). However, I was a little disappointed that more time wasn’t spent on the GSD challenges of implementing specific agile practices. (It could have easily been a full-day session.)

While waiting for the keynote session to start, I enjoyed a good chat with Darja Smite of BTH, who gave yesterday’s REMIDI talk. It turns out that we have similar views on measurement in GSD, as well as knowing some mutual friends/colleagues in Sweden!

I chatted briefly with one of the organizers at registration. Attendance for ICGSE 2010 is expected to be around 70 people, lower than usual, and about 70% university attendees, 30% industry. He attributed both the low attendance and the low industry percentage to the still-weak global economy: usually it’s over 100 people, and usually closer to 70% industry and 30% academic participation.

I’d looked forward to this COOL event since I first heard about it from game co-creator Rod Nord, and it didn’t disappoint. Getting to play the game with keynote speakers Jim Highsmith (his talk this morning on agility and architecture was excellent!) and Linda Rising (I’m looking forward to her talk tomorrow and her tutorial on Friday), and Marco of SEI, was great fun. The Hard Choices game is a useful metaphor for getting people to think and talk about software development strategies and tradeoff decisions, and I’m looking forward to trying it out ‘back home’ shortly.

You can check out the game at http://www.sei.cmu.edu/architecture/tools/hardchoices/!

Some related web references:

• A group of people (including @sepgconferences, @hi11e1, @cmmirox, @willjammer, @StagesProcess, @tcagley, @agilemanager, @srosemer, @dagreerdad, @spm221, @pdnielsen, and me-@agile_teams) tweeted during the conference using hashtag #SEPGNA. This was awesome; several times someone tweeted about the same session I was in, and I was able to arrange to meet the person F2F afterwards; when they were in different sessions, it gave me an idea of what they were hearing even though I could be in only one room.
• Several people (also including me) posted blogs here on WordPress using tag sepg-north-america-2010.
• The Week-at-a-Glance schedule has links to presenter bios.
• Presentations are now downloadable from the Session Catalog, but only by attendees (more on that later).

By the numbers (of course!): According to Anita Carleton at the Tuesday opening, nearly 800 people attended, of whom 250 were newbies, and 125 were from overseas (furthest: from Nanjing). One thing that was not clear: how many people actually attended all four days of the conference?

Monday

8:30-10am – Judah Mogilensky and Hillel Glazer
“One Stop Shopping: An Appraisal Tailoring for Increased Efficiency and Improved Collaboration” (mini-tutorial)
Key takeaway: A more real-time, collaborative approach to appraisals is not only cheaper and better, it’s more fun and less work for everyone involved. Why would you not do it this way? (tweeted)

10:30-12 – Beth Layman
“Reducing the Costs and Increasing the Value of CMMI Re-Appraisals” (mini-tutorial)
Key takeaway: This is a solid “lean, green” approach to reappraisals that makes good (re)use of PIIDs. Classic Beth!

1:30-5pm – Kent Johnson of AgileDigm
“Agile CMMI: Obtaining Real Benefits from Measurement and High Maturity” (half-day tutorial)
Key takeaway: Kent’s experiences co-teaching and co-coaching at Systematic illustrate that even a high-maturity company can use agile methods effectively to help fill the ‘how’ that CMMI [purposely] does not prescribe. (On the other hand, TSP can also fill the ‘how’ nicely, and in an agile way …). Also, Kent made the often-overlooked point that “in the agile manifesto, just because it’s on the right side of the list (eg processes and tools), that doesn’t mean you ignore it; it only means that you prefer the ones on the left.” (tweeted)

Tuesday

Keynote panel: “Process Improvement on a Regional Scale”
Panelists: Barry Dwolatzky (JCSE, South Africa), Wan Peng Ng (MDeC, Malaysia), Rafael Salazar Chavez (Mexico)
Moderator: Caroline Graettinger
Key takeaway: TSP not only works on various sized projects for a very large scale organization, it’s an effective way to accelerate building a high maturity, country-wide software discipline – but great senior management support and great coaching (not necessarily in that order) are mandatory.

Keynote: Agustin de la Maza – global practice manager for Softtek in Mexico
“Best Practices for TSP Implementation on Outsourced Application Development Projects”
Key takeaway: PSP has given them a framework for having the ‘fastest developers on earth’. Keys to effective deployment?
- “To start to transform your organization, start by transforming your management style.“
- Making teams accountable for the whole lifecycle is proven to be much more effective.

Sessions:

10:30-11:15am – Gene Miluk
“Changing Behavior: The key to adoption of complex process technology”
Key take-away: “Belief drives behavior.” When you go into an organization and see behavior that doesn’t make sense, ask yourself, what beliefs are driving this behavior? Need to change the beliefs to change the behavior. Most effective way to change beliefs (based on years of psychology): show results inconsistent with the belief. (tweeted)

11:15am-12 – Kieran Doyle
“Information Security and CMMI for Services – A Pragmatic Approach”
Key take-away: CMMI-SVC + ITIL v3 provides ‘what’ and ‘how’, but not information security. ISO 20000 (“20k”) covers it and ISO 27001 outlines requirements for it. Mapping ISO standard 27001 to the CMMI-SVC framework, and leveraging existing ITIL mappings, enables an efficient SCAMPI appraisal that provides feedback across all of these areas. PIID worksheet available on request. Very logical and practical approach which illuminates a route for extensions such as safety. (tweeted)

1:30-2:15pm – Kileen Harrison and Anne Prem
“CMMI and LSS – A Dynamic Duo” (LSS = Lean Six Sigma)
Key take-away: Successful use of LSS projects to identify and implement improvements, with the “CMMI people” focusing on risks and challenges. (tweeted)

2:15-3pm – Sujata Bapat
“Applying High Maturity Practices in Agile Projects”
Key take-away: Without asking agile teams to collect any new data, they were able to apply quantitative measurement practices. What wasn’t quite clear was how or whether doing this added value to the teams or the business. (tweeted)

3:30-4:15pm – Dirk Malzahn (Germany)
“Will ontologies help to optimize process improvement? – How ontologies can be applied in a multi-model, multi-size environment”
Key take-away: Ontology mapping is a process modeling approach that’s easy to understand and ‘automize’, not driven by tools, and useful to identify circular dependencies and islands.

4:15-5pm – Pat O’Toole
“Maturity Level 4 Results in a LOT of BS” (BS=Behavioral Stimulation)
Key take-away: Sophisticated, complicated models and measurement tactics aren’t necessarily needed to stimulate high-maturity behavior. (tweeted)

Wednesday

8:30-9:15am Anita Carleton and Linda Northrop
“Keynote – New Math: 1 + 1 > 2″
Key take-away: Combining TSP (Team Software Process) and ACE (Architecture-Centric Engineering) strengthens requirements, architecture and design among TSP teams, and provides a process framework for architecture teams. Awesome. (tweeted)

9:20-10am – Judah Mogilensky and Hillel Glazer
“SCAMPI Evidence from Agile Practices”
Key take-away: Concept of interpreting CMMI is not unique to agile, it’s a general issue. If you are achieving the purpose of a PA and avoiding the risk, then the goal of the CMMI is met. (tweeted)

10:30-11:15am – Gabriela Souza and Carlo Pires, Instituto Atlantico
“Agile with High Maturity Levels: An Innovative Approach”
Key take-away: Using DMADV with agile delivered good results and impressive quantitatively measured business improvements! (improved productivity 52%, quality 28%, reqt stability in sprint 57%)

11:15am-noon – Robert Simmering and Wadi Adam Lahrim, Booz Allen Hamilton
“Leveraging CMMI-DEV, CMMI-SVC, and collaboration tools to enhance an Agile development environment and improve investment returns”
Key take-aways: They succeeded because they ‘adopted practices that focus on exposing and fixing both people and technical issues using agile retrospectives’. Their GSD team needed online tools to substitute for literal ‘cards on a wall’ – Microsoft TFS didn’t work for them because they were unable to tailor the Scrum template enough to meet their needs (would have liked details on that). They adapted their tools to include technical debt on their dashboard, and a radar chart for total quality. NB: This presenter used for reference the same VersionOne agile poster that I keep in my office.

Lunch and SEI Member Awards
The SEI Membership Awards Lunch gave me a welcome opportunity to reconnect with Gene Miluk and Tim Chick of the SEI TSP team, as well as 3 work colleagues (2 from Sweden). Congratulations to Bill Smith on winning as Member Representative, to Kathy Smith as Outstanding Contributor, and to Gary Coleman as Outstanding Advocate!

1:30-2:15pm – Hillel Glazer
“Love and Marriage: Why Agile and CMMI *Need* Each Other!”
Key take-aways: Important factors that affect perceptions of Agile and CMMI compatibility: misuse of CMMI (‘model malpractice’), and lack of accurate information on what CMMI and agile are today. His challenge to us: learn, then go out and ‘tell the truth’.

2:15-3pm – Mike Phillips
“CMMI v1.3 Plenary Session”
Key take-aways – what will be new in CMMI V1.3:
- New PA to be added at CMMI ML5: Organizational Performance Management (OPM)
- Agile interpretation guidance – 9 PA’s will have informative material updates on what the PA means in an agile environment (tweeted)
- Architecture development guidance (retweeted)

After the talk I took the opportunity to chat briefly with Mike about People CMM, which got relatively little attention at this year’s SEPG, and whether it would be aligned with the latest constellation approach. He reassured me that it’s still being enhanced.

3-3:30pm Book Fair
I met up with Bob Stoddard (my IPPSS instructor); we chatted about our ongoing work on the CMMI-Six Sigma Body of Knowledge working group.

3:30-5pm
Peer-to-Peer Sessions (photo)
1) CMMI-Six Sigma
Very small group, but a very good discussion about blending CMMI and Six Sigma. My key takeaway to bring back to the working group: people want to blend the two improvement methodologies, not just apply CMMI practices to Six Sigma improvement projects or apply Six Sigma techniques to CMMI-based improvement initiatives.

2) TSP Symposium Program Committee
I enjoyed meeting Michele and Rafael in person, and reconnecting with Bill, Alan, and Tim.

Thursday

8:30-9:15am – Hillel Glazer
“Top Ten Clues You’re Probably Not Doing Engineering”
Key take-away: “Not everything involving software is engineering” – or Development – and suitable for CMMI-DEV. CMMI-SVC fills a real need by sketching a clear path for non-engineering organizations to still benchmark their performance and work to improve. His suggestion for a new value to be added to the Manifesto: “we value Improvement over Compliance” (tweeted)

9:15-10am – Winfried Russwurm
“The multi-model and multi-appraisal quagmire – an approach to organize them by a classification scheme”
Key take-away: This classification scheme defines a way to visually portray multiple models in ‘footprint charts’ (disappointingly, the charts he’s completed do not seem to be available).

10:30-11:15am – (my presentation slot – my slides)
“Using Requirements Metrics To Guide Project Management And Assess Process Improvements”
Great audience (84 people!) – Q&A discussion focused on how often to measure.

11:15-noon – Dennis Goldenson
“The Role of Expert Judgment in Improving Software Estimation Processes”
Key take-away: Reference to a Magne Jorgensen and Barry Boehm debate on estimation, moderated by Stan Rifkin which was published in IEEE Software in March/April 2009 – must fetch/read. (tweeted)

Lunch: Hillel and I skipped the SEI Partner lunch to look for the tweetup table, but #39 seemed to be occupied by others (and we didn’t get the ‘table 46 overflow‘ tweet until too late). The mixup did give me a happy chance to talk at some length with Bill Nichols and Jim Over of the SEI TSP team, and learn more about AIM (the new IDEAL) and how they are using it to help organizations greatly accelerate their maturity progress.

1:30-5pm – Diane Mizukami-Williams
“How to Create and Deliver Great Presentations” (Professional Development Workshop)
Key take-away: Most advice on presenting is on how not to fail – not how to really succeed. Her goal: to be remembered. She offers six techniques for being different. General strategy: Break the ice and help your audience get comfortable with you as a person (I’m going to try her “Which one is the lie?” quiz in the near future).

I also took time on breaks during the week to visit the student poster area. Some had interesting ideas, although they were not at the level of quality in content or format that I am used to seeing at academic conferences – I am sure this will improve over time. I used my digital camera to capture their URLs and contact info (one student completely omitted this, oops!).
(poster 1)

Photos will be attached within the next few weeks …

Do I want to go to SEPG in Portland next year? You bet!

A PDF of the updated slides I delivered can be downloaded from the agileteams.com website publications area. Comments are welcome, here or on the agileteams blog or Twitter!

I was pretty happy with my two: Judah Mogilensky and Hillel Glazer tag-teamed on their ‘super-sonic SCAMPI’ approach that’s both leaner and more collaborative, and Beth Layman gave a nice talk on making (re)appraisals more efficient. A key takeaway both sessions emphasized: more reliance on ‘affirmations’ and less on ‘indirect’ evidence.

On the walk to lunch, I met up briefly with some SEI and external colleagues from my TSP coaching days, who I plan to catch up with further over the next few days.

Choosing an afternoon session was tough: multiple models, CMMI+TSP, requirements, or agile CMMI? Again my colleagues and I split up: one was room moderator for the CMMI-TSP session, and another wanted to see Bill Diebler’s requirements, so I flipped a coin and went to agile CMMI.

Kent Johnson offered some interesting data on how a 500-person company in Denmark made significant productivity improvements by adding agile to their CMMI maturity level 5 approach. I was a little puzzled though by some of the measures proposed, for instance the completeness of the items in the product backlog: in agile, the product backlog shouldn’t be defined in great detail too early – selected stories are clarified collaboratively with the Product Owner during sprint planning. (Plus, how to actually measure this idea of backlog item completeness was never explained.) But it was generally a worthwhile session.

Locating a seat in this crowded afternoon session gave me another lucky reconnection: I ended up sitting right next to someone I had worked with in another state over 15 years ago. It was so nice to catch up with Caroline on break and afterward! Then I strolled down to the requirements session I’d missed to meet Bill F2F, and we had a good chat.

The last good news item of the day: the Savannah river’s been reopened and the ferries are running again! I’m looking forward to riding tomorrow.

Final thought on what I enjoyed about SEPG today: to me it’s always a nice treat to participate in a technical event that has more than the typical ratio of 5-15% women – it’s been closer to 50% in my sessions today.

Biggest disappointments so far? The rooms are crowded, drink selection and availability could be better on breaks (bringing my own water bottle tomorrow), and wireless connectivity in the session rooms is too weak to use.

I took a few photos, but don’t have time to offload and attach them tonight; I’ll post a comment here when they’re available.

After circling the block twice, I couldn’t identify where to park for check-in, and no open public parking was available on the street. I saw two people standing around at an opening on the side of the block, and drove up to ask them if they knew. They turned out to be the parking valets (the only hotel parking option, but with no signs) and that WAS the place to go for check-in parking.

So I pulled my car to the side, got out to check in (which went smoothly), and came back to my car. The valet asked me, are you staying? to which I said yes (and began to wonder how often people pulled in there and then decided not to stay). He then handed me a claim check stub and walked back to his station.

I hadn’t yet finished unloading my luggage onto a cart when a different valet walked up and asked me if I had a claim check, to which I said yes; and asked if the key was in the car. I said no, and handed it to him.

Then I asked him, how will you know that this key and car go with this claim check? and he looked at me in puzzlement and said, well, because you just gave it to me. I said no, how will you keep track of it, how will you know that this is MY car? and he just didn’t seem to understand, shook his head as if I was an idiot for asking.

After a few more minutes, I finished unloading my car, then turned to talk to the valet again. But before I said anything further, I could almost see the light bulb go off above his head – he said, “wait a minute”, went back to the valet station, came back with a rearview mirror hang tag that matched up to my claim check, and put it inside my car.

Only then did I feel confident that I’d actually be able to get my car back from them when the time came.

On the surface, this little incident has nothing to do with software. But it has a lot to do with processes, customer service, and managing data.

Their claim check tag system for valet parking data management is simple and low-tech, but when well executed, should be highly effective (and in the spirit of doing the simplest thing that could possibly work). However, two problems happened here:
(1) it wasn’t well executed, when the first valet dropped the tagging job mid-stream, and
(2) the second valet dropped the ball in two ways: by assuming that my having a claim check in hand meant that the tagging process was completed properly, and by treating his customer like an idiot for asking what was actually a reasonable and important question about their process and how it was being executed for my car.

After all, the purpose of the system and how well they use it is to support them in serving my needs as a customer. Those needs include not just parking my vehicle safely, but reassuring me that it’s being well managed, and that they are keeping good track of which one is mine.

I’ll post a followup comment by the end of the week on whether I run into any difficulties in getting my car back!

• make it easier for more collaborators to join in the discussions, and
• reduce WP admin work for the volunteer webmistress (me)

Looking forward to chatting with you all!