Wednesday 10 October 2012

Model for Software Release Schedule Process


In keeping with my wish to share the tools I've come to use in managing my projects, in this post I plan to share a simple, yet effective template for planning, coordinating & managing the Software Release Schedule (a.k.a Bill of Materials BoM).

This process is typically owned by the party responsible for System Integration or producing the final Software Deliverable. Again, my context is a Set-Top-Box (STB) Software project, where the system basically consists of a number of software components, typically provided by independent component providers. For instance, the common building blocks in a STB Software stack are:
  • Physical Target STB Hardware Device (Decoder) with Platform Software
    • Platform Software is basically:
      • Firmware - Bootloader
      • Operating System - nowadays mostly Linux Kernel
        • Customised by Chipset Provider, e.g. Broadcom's own Linux Toolchain
      • Device Specific Drivers complying to a Middleware Interface (Macro - OS)
  • Middleware Component
  • Virtual Machine Component (VM Engine)
  • Electronic Programme Guide (Primary User Interface / Application component)
  • Interactive Applications
If you would like more information on what goes on in DTV Software Projects, just search for "DTV" on this site. I have previously written about DTV Architecture, STBSI and Agile Development Methods.

Because these components may come from a variety of sources, there is usually a need for assigning the role of [STBSI], or Set-Top-Box System Integration. This team is responsible for producing the system software build that will eventually be deployed by the customer, the [PAYTV OPERATOR].

[STBSI] in my opinion, is core to the success of any project. They have ultimate responsibility and accountability for ensuring a fit-for-purpose product is assembled and deployed. They have the freedom to drive and manage the independent [COMPONENT VENDORS], often delegated the role of customer, i.e. [STBSI] has the [PAYTV OPERATOR]'s best interest at heart, and often makes the necessary judgement and priority calls for the customer. However, this is not always the case, as some [PAYTV OPERATOR]'s have a strong Product Management team, where the [PRODUCT OWNER] works with [STBSI] in determining the priorities, etc.

So what's a Release Schedule or BoM anyway?
I've seen projects executed through the classic methods of Waterfall, Spiral/V-Model and more recently more Agile, Iterative models of project delivery. Once the early development / integration phase is reaching completeness, with STB projects in particular (despite using Agile) there is generally a massive effort of further integration & stabilisation. In this phase, the component releases need to be carefully controlled, and clear communications must be maintained with all parties, to ensure the correct issues are worked on, the priorities are made clear, and that everyone understands the goals [STBSI] is ultimately aiming for.

So there is a need to maintain a sense of a Software Release Schedule or Bill Of Materials, that is executed in a timeous, methodical manner, in producing software component releases. There needs to be some structure and method, that all parties need to understand and agree on as a way of working. This then becomes a blueprint, a regular heartbeat that is executed timeously - removes ambiguity and provides a sense of clarity and control.

Brief Walk through of the Tool
I wish I could take credit for conceiving the original visualisation, but this is actually something that I picked up from a previous project from four years back, however I've since then modified & added my own enhancements in producing a generic model for my PM Toolbox, reusing on all future projects:
Generic Model for Software Release Schedule (BoM) Process
This visualization should be self-explanatory. If you don't think it is, then I've failed in communicating the essence of the process, which will make it difficult for people to pick up and reuse. If you need more info, please get in touch.

This model assumes a timeline, as shown by the main arrow in the centre - showing a timeline of working days for the period leading up to a specific release iteration point (a.k.a. sprint or release cycle), during the cycle, and post the release cycle.

More projects are adopting some form of Agile, the basis being iterative, incremental release points. The typical STBSI release cycle is [TWO] weeks, it can be more, say three weeks, or less, about a week. It all depends on the strength of your STBSI team, the maturity and stability of your components, and the willingness of your component vendors to be flexible in producing regular release drops. Traditionally DTB Component Providers, especially Middleware providers have been very poor in delivering continuous incremental releases - which has been the bane of many PAYTV Operator's projects because the Middleware providers lock the customer in to their own processes, rigid and slow, rather than bend to the norms of current trends of flexibility. Times are changing now, and these component vendors are realising their mistakes because PayTV operators are seeking alternatives - features must be deployed in the order of a couple of months rather than at least 6-12 months...

Coming back to the process, in a nutshell:
  • Assumes a Release Cycle is in the order of TWO working weeks, or 10 man-days
  • The work or backlog for the next release cycle is governed by the Release Schedule or BoM
  • The BoM is produced leading up to the Next Cycle, a few days in advance to allow Vendors to plan
  • The BoM is baselined at the start of the Release Cycle
  • The BoM is subject to change based on any new Showstopping defects the Project uncovers
  • Vendors are continually monitored to track the progress with burning down the issues assigned
  • Any deviation to the agreed BoM is subject to Programme/Management interventions
  • Vendors release their components at the end of the Release Cycle subject to meeting the Acceptance Criteria as defined by SI
  • [DTT] - is a placeholder for your own Defect Tracking Tool
  • [STBSI] - is a placeholder for your Integration Team
  • [Vendor] - can be used as a placeholder to replace if you only have one vendor, otherwise keep it generic

Is the Tool/Process Effective? Surely this goes against Agile? Is this Not PM Overkill?
I believe the complex, variable and unpredictable nature of a STB Project calls for some processes to provide some sense of control. This is especially true for STBSI teams who are new the to SI process and are in a state of chaos. I've seen this on past projects, and on introducing the tool, it brought a sigh of relief to the project as it allowed for all players to settle on a common understanding and goal. So yes, I believe the tool, a visual representation is always useful way of communicating a method of working.

Is the Process Effective - it is as effective as the team responsible for execution. Like any process, there is some management, administration and diligent participation required from all impacted players. It certainly can make a Release Manager's life much easier, and planning more predictable. There is a time and place for instigating this process, so used prematurely, this BoM process might not be very effective, and could become a hindrance more than anything else.

Agile: First, I don't believe Agile or Scrum works well in the context of a STB project delivery, without a massive investment in management, contractual agreements between vendors, etc - such that the delivery is managed, end-to-end in an Agile manner. I will expand on this in a future post as to why I think Agile will not work in Digital TV Systems projects. The BoM process is iterative, and is aimed at producing incremental releases of sound functionality and stability. There is a Backlog, called the release schedule. Work is broken down and executed in release cycles, two week sprints. So the concepts of Agile are present, but the fundamental issue of leaving bug/fixing/stabilisation toward the end of the project cycle, after the initial development/integration cycle, indeed disobeys a fundamental rule of Agile/Scrum as delivering shippable, stable product at the end of each iteration - and doing so, would not call for a separate stabilisation/bug-fix stage...again, this is a topic for another post.

PM Overkill? To some yes, it might be. But to others, this brings a sense of clarity and purpose to a project delivery team. It sets the stage for Release Management, communicates the rules and behaviours expected from all parties, and establishes a baseline for communication and project escalation. So no, I don't believe this is Project Management Overkill!

Monday 8 October 2012

Meeting Tom Gilb & his take on Competitive Engineering


Last week I attended a two day masterclass workshop on Competitive Engineering by the world-reknown, and much respected Tom Gilb.  Taking a paragraph out of the training brochure from Secolo Consulting.  Book  Competitive Engineering: A Handbook For Systems Engineering, Requirements Engineering, and Software Engineering Using Planguage

Tom Gilb is a competitive engineering consulting mainly serving multi-national clients in improving their organizations and methods. He works with major multinationals such as Boeing, Bosch, Qualcomm, HP, IBM, Nokia, Ericsson, Credit Suisse, Sun Open Office, Microsoft, US DOD, UK MOD, Symbian, Philips, BAe, Intel, Citigroup, Telenor, BAA, Den norske Veritas, Schilumberger, Tektronix, Thales, GE, GSK; and many others - including smaller organizations such as Confirmit, University of Trondheim IT, Enlight, Avenir, Clearstream, UECC....Gilb is the author of 9 books, his latest book is "Competitive Engineering: A Handbook for Systems Engineering"...

I have heard of Gilb before, and have, for a very long time, a few of his books on my Amazon Wishlist, namely Software Metrics & Software Inspection, as these books come well recommended by other famous people in Software. But admittedly I had held off buying Gilb's books because they were written way back then, pre-1990 stage, before my time - surely there can't be anything of relevance in those books?? They must be out-dated, filled with stuff around structured and sequential programming techniques, classic traditional Waterfall methodology of project estimation, implementation & delivery...why bother with these old relics, better learn from what's current and trending...


How naive was I? Just as I my eyes were opened when I read the Mythical Man-Month, realizing that the challenges that were faced in the early days are still very much relevant today, my eyes again opened on being enlightened by the ideas presented by Gilb in person, not through reading his work, but with face-to-face, personal human interaction. It was truly an honor & privilege to have spent two days with Tom...


So what did I take from this course then?
The workshop was really around navigating through the book, Competitive Engineering, discussing ideas and spending time on a few sample case studies and Tom's war stories. It was basically a seeder, an introduction to share this concept, to instigate new ideas and ways of thinking - and introducing the fundamental mindset changes required from people and organizations alike.

We had no time to do any practical tutorials - it was much a one-directional push from Gilb, we just had to listen, absorb and catch any ideas worth following up.

Of course, I need to give the book a read at least two attempts, it is very dense and probably requires a third read to be fully grounded. At the same time, I need to choose a few concepts and start working on them on my current projects. Whilst the ideas are fresh in my head, my take aways are:
  • Value-driven project delivery.  The idea of delivering incremental value is not new, I've always tried to work in this fashion, especially with my uptake of Scrum & Agile.  However, the fact that Gilb started this technique off back in the seventies or eighties was enlightening, Agile is not a new concept.
    • The tools & methods used to derive at kicking off the value-based delivery project is really interesting and requires more focus and experimentation.
    • Impact Estimation Table - seems to be a very powerful, multi-dimensional view of establishing the core values of the project/architecture/design. Everything starts off with the Impact Estimation Table
    • I need to look at synergies of applying some of the rigor and discipline of generating the Impact Estimation Table whilst simultaneously mapping this with existing  Scrum/Agile methods of creating the Product Backlog
  • Quantification is Key. Everything can be quantified. You're either too lazy or too stupid, or really ignorant (or can afford not to) to spend the time and energy in quantifying your core project requirements / architecture / design. If a thing is important and of value, then it must be Quantified
  • Evolutionary Project Management or Evo. This is the concept invented by Gilb which really implements the techniques for iterating on value-based deliveries. Deliver incremental value that can be easily measured on a week-by-week basis. Involves clearly defining the definition of done, which again is borne from the Impact Estimation Table.
  • Concepts such as Credibility, Evidence, Past Metrics. Gilb's tools forces you to dig deep, to really drill down into the details, quantifying everything as much as possible. I was going to share my generic Risks Register Template that I use for most of my DTV projects, actually I will share the first version, and then follow-up with an updated version to introduce concepts in the Risks Register such as: How credible is this risk? Do you have proof this risk happened before? What was the past project's benchmark, if any? All of this factor into determine the overall value of a risk. Isn't it obvious the project should really do all it takes to mitigate a risk if it came from someone who was expert in the field and hence very credible, has enough evidence to back it up and a sample of data backing up the risk??
  • Planguage. Planguage (Planning Language), pronounced "Plan-guage", is an invention by Gilb that is really a culmination of his 50+ years experience in the industry. He has created a language, that is complete with a syntax for both words and pictures, that is used to describe his method of planning, and used in all related tools.  There is a lot of handbooks out there, standards and configurations, templates and processes - but Gilb firmly believes these are all ok, some might be downright dismal, but they all lack a fundamental attribute of quantifying quality...
I have a lot more background work to do now, but I believe there's enough value and merit in testing these ideas out. I had in real-time, during the course, googled for any expert critics on Gilb, refuting his principles and methods - and came up short.  Gilb has an unbelievable track record and tenure, he is well respected in the industry, his work has pioneered much of the quality processes in Systems/Software Engineering - if you're serious about Engineering principles, in delivering value across the board, then you're stupid not to give this guy's methods a chance...

My tweets during the course...






Monday 1 October 2012

Milestones Template for Managing & Tracking a Digital TV Programme


Do you keep a PM Toolbox close to hand?
One of my most common questions I ask candidates interviewing for Project Management or Programme Management positions is the subject of maintaining your own personal set of tried-and-tested tools you've come to depend on in your career with managing projects: Do you have any tools or templates that you keep with you as a standard reference such that you hit the ground running on your new project? If so, what kind of tools do you most value and cherish, and why??

Surprisingly though, most of the people I've interviewed either settled for the generic tools that are sold by the various PM Methodologies (Prince2, PMI, PMBOK, Scrum) - I've come across very few people who keep an arsenal of personal tools ready-to-hand to pick-up and use on their next project.  I too am guilty as charged, very often relying on memory and previous project documents to pick-up and use as a starting point.  I have however, started to collect my own set of tools that I'm willing to share with everyone, in the interest of knowledge-sharing and collaboration, the first starting with this post on a Template for Tracking Milestones on a Digital TV Programme.

When I used to be a developer back in the day, I had my own set of code samples that were generic enough to be re-used, my recommended debugging tools, libraries, parsers, open source projects, and so on. Along with your trusted book on Algorithm Design, Software Architecture Patterns and System Utilities - necessary tools for the job, like any other trade.  So too, for people involved in various aspects of Project Management, you could keep your very own toolbox that can help speed-up the much repeated, but often required aspects of the job.

A Preview of my PM Tools Backlog
Because my background really is around Digital TV Systems Projects, an area that has still a lot of life to survive the next 10-15 years, especially since Digital TV (DTV) is really only just beginning to fire up in emerging markets of Africa, India & China, there will still be a need and demand for such Project Managers.  I aim to share as much knowledge as I've gained through practical experience, by making available the tools I've come to trust and use in both past and current projects alike:
  • Generic Model for Tracking Milestones in a DTV Programme (this post)
  • Programme Model for Staged Deliveries for a DTV Launch
  • A Modeling tool to Estimate & Predict Realistic Time to Market (Climb to Launch)
  • Set-Top-Box Software Integration Release Schedule Process
  • A Generic Risk Register for a Typical DTV Programme
  • A Template for a typical DTV Project Charter or Software Development Plan
  • A Template for Managing a DTV Project based on an Agile Product Backlog
  • RAG Template Options
  • Models for Programme Organizational Structure
  • Methods for Controlling Quality: Effective Defect Management
  • Methods for Managing STB Software Development Processes
  • Template for Defining Bootloader Development Work-Stream
  • DTV Projects Matrix - A guide to understanding the permutations & combinations of system impacted components in a variety of different projects, arranged by complexity and rough estimations of project life-cycles
So I intend to release a tool as often as I can, in the hope that other Project Managers can use, learn, benefit and adapt these tools in their own day-to-day projects...

PM Toolbox #1: Generic Milestones Template for DTV Programme Management
This tool is all about summarizing and tracking the various Milestones that must be met in order to achieve a successful DTV Project Delivery. It is essentially the fifty thousand foot view of the high level work packages (or Work Breakdown / Sub-Projects) that essentially make up the various streams of delivery in the overall Programme.

The tool serves the purposes of communicating clearly the status of the overall Programme, which is what Senior Stakeholders are interested in.  It also serves as the master Project Plan for the entire programme, around which dependent sub-projects or work streams must deliver to.  Whilst it might come across as a simple table, list with columns, it is nevertheless a powerful tool in managing and tracking the overall project.

Download the Tool!
Please click here to download the tool. It is simply a Microsoft Excel spreadsheet that provides a list of milestones, for each milestone, we track the following criteria:

  • Milestone Name / Title - The actual event that manifests as an identifiable achievement of a goal
  • Company Responsible - The company or vendor who is assigned responsibility and accountability for delivering the result
  • Actual Plan - Tracks the current forecasted planned delivery date
  • Original Plan - Tracks the baselined original dates at the inception of the project
  • Acceptance Criteria - For each milestone, the agreed acceptance criteria (at high level)
  • Dependencies - For each milestone, detail the associated dependencies the milestone has, preventing successful completion
  • RAG Status - An agreed RAG convention to be used to summarize the status of the milestone
  • Recovery / Mitigation Actions - A set of actions to be taken to implement recovery scenarios
  • RAG Comments - A note on the current state of affairs
  • Project Work-Stream Owner - The name of the Project Manager or Owner assigned responsibility and accountability for delivering on that milestone

The table is pre-populated for you with the standard set of milestones I usually expect from a typical DTV Project. I have also pre-filled in scenarios for Acceptance Criteria, Dependencies as well as listed your typical Recovery Scenarios.  The tool comes with an information page that instructs you how to go about replacing the template with actual project data.

What is a Digital TV Programme: Players, Concepts, Milestones?
In order to appreciate the nature of a Digital TV Systems Project, you have to first indulge in a brief overview of what a Digital TV System entails, the players involved and the typical project configurations expected:

So now that you've read this background material, we can start exploring the make-up of the Milestones Tracking for a typical DTV Project:

The template assumes a DTV Project that impacts the entire end-to-end system of a PayTV Operator's value chain. Not all DTV projects encompass such massive changes across the board, there are some projects that involve just Set-Top-Box (STB) Software-Only changes, other projects might only involve introducing a new STB hardware device (decoder) with no changes to software. There are also Headend projects dealing with routine maintenance and enhancements as well.  I plan to create and share a matrix of your typical DTV project-based scenarios one day...

The tool assumes a big-bang project that involves the following players:

  • [PAYTV OPERATOR] - This is generally the PayTV Operator that is implementing the project to deliver or enhance their existing systems (e.g. BSkyB, DirecTV, Multichoice, etc.).
  • [STB MANUFACTURER] - The designated company responsible for the hardware design and manufacture of the client Set-Top-Box (STB) (a.k.a. Decoder), supplying related device Software (Drivers).
  • [MIDDLEWARE PROVIDER] - The company or service provider responsible for developing the software operating system for the STB, including providing value added services that enables rich Application Development for STB User Experience (i.e. the Electronic Programme Guide / EPG).
  • [EPG UI DEVELOPER] - The company or service provider a [PAYTV OPERATOR] usually employs to develop the STB client application or user experience. There is a growing trend with PayTV Operators electing to do this development in-house.
  • [INTERACTIVE APPLICATIONS DEVELOPER] - The company or service provider a [PAYTB OPERATOR] employs to implement value-add interactive applications that are built on the [MIDDLEWARE PROVIDER]'s platforms. Increasingly PayTV operators are also electing to do in-house development, or to contract this work out to the same [EPG UI DEVELOPER] or indeed have a third-party application development house contracted.
  • [HEADEND PROVIDER] - The company or service provider responsible for developing the backend or headend systems. Typically this is the same as [MIDDLEWARE PROVIDER] since the technologies are inter-related, but there are some PayTV operators that prefer open standards and choose to use a variety of suppliers to manage their risks (i.e. not to have all their eggs in one basket).
  • [SYSTEMS INTEGRATOR] - The company or service provider with sole responsibility and ownership for proving components & systems are well integrated and deliver on the desired functionality and expectations of the programme.  There are many levels of System Integration, depending on the [PAYTV OPERATOR]'s budget, etc. Typically though, it seems that [PAYTV OPERATOR]'s are aiming to be more independent from their service providers in taking the plunge of owning not only the System Architecture but also the System Integration delivery aspects of the programme.
With these players in mind, the Programme can then be translated to obey the following high-level concepts  in translating detailed deliverables or milestones (The detailed breakdown of the plan is a topic for another post):
  • STB Manufacturer Hardware Delivery
    • Hardware Design Specification
    • Remote Control Design Specification
    • Packaging & Artwork Specification
    • Hardware Acceptance Testing
  • STB Software Delivery
    • Device Firmware Bootloader Software Ready
    • Device Hardware Platform Driver Software Ready
    • Middleware Component Functionally Complete
    • UI / EPG Component Functionally Complete
    • Interactive Applications Functionally Complete
  • Backend / Headend Delivery Functionally Complete
    • Broadcast Components Operational
  • Systems Integration
    • STB Integration Functionally Complete
    • Headend Components Integrated
  • Launch / Deploy Product
    • All Acceptance Criteria Met
    • System has acceptable level of defects
    • Business decision granted for Launch
This is just a summary of the deliverables one can track. The tool provides a sample of 34 Milestones that drill down into the finer details.

In the next post I will spend some time explaining what some of these milestones mean, and how one should realistically go about planning the various work-streams, such that dependencies are minimized whilst simultaneously running multiple sub-projects in parallel.


Summary
This is a first draft, a brief attempt at sharing what I think is a powerful tool to aid Project Managers in managing Digital TV Projects.  I have for now, assumed the reader is familiar with the territory and therefore have excluded explanations of describing what the different sub-projects entail, etc. My aim was to share this tool with already practicing Project Managers in the DTV trade, to solicit feedback and share new ways of working, etc.
Some people might think I'm giving away some of my crown jewels by sharing this knowledge, that I'm forsaking my competitive advantage, etc...but I don't think so. I've laid out a simple framework that can be useful to others. There are many templates and tools out there to aid Project Managers, the success of a project delivery lies not in the templates, but in the project manager's knowledge, wisdom and tenacity of execution and implementation....