Wednesday, March 17. 2010Fitnium 2.0 and SLiM
I now have a version of Fitnium working with SLiM, version 2.0 to be release shortly will include support for the existing Fitnesse DoFixture plus support for the SLiM Script Table
You will be able to continue using your existing scripts with no change, but also start developing in SLiM Wednesday, March 10. 2010Agile & Diving
I've recently started a new Agile coaching gig in the Telco space. The team are already into the development cycle but had been struggling with scope, quality and communication, hence the call from the programme manager one day
They already have a system in production and have ongoing support requests, plus a decent sized team 16+ people actively seeking work each day, there is not the option do a drains up coaching exercise and get everyone trained in the basics before we start cutting code, testing and delivering, there is a business to run. The approach I've taken reminds me of how you learn to dive. Start simple, head under the water and breath, then introduce the concepts one by one, mask clearing, BCD ( buoyancy control ), proper finning etc piece by piece, but only after the team demonstrate that they have mastered the current activity. Basically I am brain training them, getting them to accept and adopt a specific practice to the point where it becomes second nature. Just like with diving you don't progress to open water until all the safety procedures are pretty much second nature. When it goes wrong you don't think, you just act and therefore you don't drown. We started with planning, poker and standups and did not change anything else about their current process. Just these 3 aspects got the teams talking more, got developers and testers talking to analysts and demonstrating finished functionality in small easily understandable and therefore testable chunks. People saw the changes added value and actively started seeking more advanced Agile concepts, automated build and testing, mock objects etc, but always under the control of the coach at first to make sure they were doing it right. This reminded me of when I did my first pool session, I was taught the basics, the instructor making sure I did them all correct and then let me play in the pool, but never so far away as he couldn't keep an eye on me and correct anything that he saw was wrong. I practiced full mask clearing ( only been taught partial mask clearance ) and I practiced buoyancy control through breathing in and out ( we'd only be shown how to use the BCD ). This analogy reminds me that the concepts of agile are not new, just best practice and easily learnt and if taught in small easily digestible parts allows the student to progress to new areas at their own speed. Tuesday, July 7. 2009After the Goldrush - When The Consultancies Finally Leave
You are on your first Agile project, and you are struggling, so you decide to bring in a consultancy, one with a really shinny new website and a bunch of people who can blog for their country as an Olympic sport. They arrive in a whirlwind of new terms and behaviour, coaching your team on all things Agile, pressing the point that Test Driven Development is the way forward, showing you burn down graphs, story cards, planning poker and something called a retrospective. Unfortunately the money soon dries up, quite often well before the project does and off they trot with a hearty slap on the back, a fond farewell and that "don't worry, it will all be OK, we've coached you", look in their eye.
This uncomfortable truth happens all too often, and all too often it can leave a project in disarray, its like the first time you have the stabilisers taken off your cycle, unfortunately its more like been given the keys to Learjet after 4 lessons. You sort of get the gist, you known what you should do, but you are at 10,000 feet and an engine just blew out. Its at times like these that so many projects revert back to old ways wasting all the time and money invested in showing them there are new better ways to work. I've seen this happen myself, quite often the customer had a limited budget and therefore when then money dries up, no amount of persuasion or convincing is going to get them to pay for more and even in these enlightened times who wants to work for free. Alternatively I've seen situations where projects managers under pressure for budgets think because its all going so well their teams must have really picked it all up and can run on their own now, so why not claw some of that money back early. Inevitably problems occur, not big ones, but because its all new and different they often seem big at the time. So for all those currently going through some form of Agile/Lean process improvement here are a few guidelines that I have seen work well for others. 1. Don't suffer alone You don't have to give away the company secrets, you don't even have to sign on with your company email address, but get yourself signed up for a few news groups, and forums. These are excellent sources of information with people at all skill levels willing to help out for free. A search on Yahoo Groups or Google Groups will bring up the names of some great forums such as AgileDevelopmentManagement, extremeprogramming and scrumdevelopment, while LinkedIn has a wide range of Agile related groups such as Agile Agile & Scrum UK, Agile Alliance and extreme programmign (XP) If you are really lucky there might be an actual Agile group in your local area, where you can meet up socially chat about problems, listen to talks from experts and generally have a shoulder to cry on ( over a beer or three ) when it gets really hard. 2. Be prepared for mistakes and learn from them Mistakes will happen. Agile is a learning exercise that is best experienced through repeated practise. That means sometimes you'll get it wrong, but the key is not to see these as failures, but as learning points. Use the concept of retrospectives the coach would have taught you to look at why the problem occurred for your organisation and what you could try and do to ensure it doesn't happen again. 3. Continue the education If you are paying £20,000 - £40,000 for a month or two of an Agile coach, what is £3000 for a Scrum Master or Product Owner course for 2 or three of your staff. I've seen this work well for teams that have been left on their own for a short period. The certification allows them to take along specific problems and experiences to share with the class which benefits everyone There are also plenty of books on Agile, make sure you coach leaves you with a decent list of the best. Again £500 spent now on a good library can save you £1000's down the line. 4. Work the network It seems that everyone these days is adopting agile, and in many ways they are, whether its hard core Agile XP or Scrum adoption, or looking at Lean process improvement, or just dabbling with better coding and testing through TDD and CI. It probably means ex-colleagues and friends in the industry are also going through the same growing pains, or may have done and have stories to share, its time to pull out that old address book, or review that LinkedIn profile and work that network. 5. Ask for support Any good consultancy, even the single owner start ups should provide you with a level of support free of charge. If they don't you need to question why you are using them. They should at least offer to respond to email queries in a reasonable amount of time, may even provide a support forum for clients, or even a phone number to call when it gets really tough. Suffice to say any good company will not leave you abandoned and should providing a level of back up support once they leave, but make sure you check before engaging them. 6. Prepare for them leaving This may sound stupid but in heat of all the new things Agile brings, many companies forget to plan for the next biggest event to the go live of their product and that is the consultancy leaving. In my experience the best time to plan this is during the negotiations of them starting. Just like your project has a goal, so must the enablement and this has to be planned in from day one, so that it can be measured and reported on. You never know if you achieve your goals early, there is nothing wrong with the consultancy taking any early exit !!!!, but joking aside you need to know when it is going to happen and what needs to be place when it does. Tuesday, July 7. 2009Agile exposes scaling problems
A blog entry on scaling Agile, and how its not Agile which doesn't scale, but the organisation under going agile probably didn't scale anyway, and therefore Agile just highlighted the problem
Tuesday, July 7. 2009New to agile? What to do when you are behind
Excellent blog entry by Agile Bob on tracking time and what do when things eventually slip behind
Wednesday, November 19. 2008When Agile Projects Go Bad
James Turner writes about When Agile Projects Go Bad
"Your software development projects can benefit from Agile - assuming it's really what's used. Learn about the sins that have been committed in the name of "Agile." With input from two of the Agile manifesto signatories." Tuesday, November 18. 2008
Jakob Nielsen on Agile Development ... Posted by Keith Sterling
at
12:09
Comments (0) Trackbacks (0) Jakob Nielsen on Agile Development and Usability
Jakob Nielsen writes about Agile Development and Usability and how agile can pose new threats to to user experience quality.
However he does go on to show in the article how Agile can be used in enhance user experience by not approaching agile in a narrow fashion, but open it up to embrace usability testing and end user involvement in design. Monday, November 17. 2008The Decline and Fall of Agile
James Shore has written an excellent article The Decline and Fall of Agile where he points out that Agile projects are failing because too many of them are only adopting parts of the agile principles. Lots of projects are doing short iterations and increasing planning, but too few are doing more of the esoteric or softer practices correctly.
This follows on from my previous post that too many people are being churned out of CSM courses with effectively 2 days Agile training and then trying to apply this skill on major projects. Therefore anything not taught on the course is not applied because it was not on the course, unfortunately good development leads are not created then are forged out of years and years working on successful and not so successful projects. Just as their software grows, as does their understanding of what it takes to build great software. I know so many good agile dev leads that either don't have CSM or have because their company paid for it, but never needed to go on it in the first place. They are good because they have always been good and will continue to be pragmatic and agile in their adoption of all that is good in agile itself Sunday, November 9. 2008
Scrum Ceritification - Atlast it ... Posted by Keith Sterling
in Scrum at
09:47
Comments (0) Trackbacks (0) Scrum Ceritification - Atlast it might mean somethingInfoQ: have a interesting discussion about the value of the Certified Scrum Master certification. I've always been trouble by the value of CSM certification, not because of the course per say, but the fact everyone who turns up for the course is awarded it, and the the certification is being used by people to suggest that are experienced at running Scrum projects. However the big players in Scrum are starting to take notice and are discussing the need for a test at the end which would ensure that every candidate has the required level of understanding of Scrum before they can use the CSM certification Sunday, November 9. 2008
Agile development: It isn't just for ... Posted by Keith Sterling
in Agile at
09:15
Comments (0) Trackbacks (0) Agile development: It isn't just for small projects
Scott Ambler and Damon Poole discuss how IBM scales Agile across large teams ( sometimes in excess of 500 people ), in an article Agile development: It isn't just for small projects
While a brief article it does touch on one or two of the really key problems facing agile scaling; how to size and split teams into management units. As Ambler says "The architecture should be a system of subsystems." and therefore according to Ambler "the structure of the teams should be aligned with the architecture, and ideally the sub-team members should be co-located." I have to admit that I always saw greater perceived value in creating teams that attack vertical slices through the architecture, ultimately these teams deliver end to end stories, and this worked extremely well on a major project I did for AOL. However at one of my current clients, this is just not feasible at this stage, the political makeup of the organisation and the inherent compartmentalization of existing teams, mean that its has been a major ( and ultimately impossible ) struggle to get a manager ( who once controlled his whole team ) to seed various functional teams with one or 2 of his assigned developers. We have therefore grown into a suite of teams split by architecture. Time will tell if this will actually work Thursday, November 6. 2008Selenium API Spreadsheet
While I think Selenium is a great tool, the documentation can be a little difficult to work out the specific command. As part of a training course I have been developing, I have created the following spreadsheet of Selenium IDE API calls which you might find useful
http://spreadsheets.google.com/ccc?key=pq3bDQaoQ3jphqQfHrfKRYw Shared on Google docs Tuesday, November 4. 2008In Search of Stupidity - A Review
I love this book, In Search of Stupidity by Merril R Chapman
If you grew up with computer companies in the 80's, and remember MS-DOS, Lotus 1-2-3, Wordstar, dBase, Ashton Tate, or just want to know why Microsoft got big and Borland suck then this is the book for you. The author's personal experience on WordStar is used to dissect many other of the software industries great marketing fowl-ups. If you are involved in writing or producing software products you have to read this book and make sure you don't make the mistakes he describe, such as Wordstar trying to sell 2 competing products, yes thats right, 2 word processors, both different, both word processors. You laugh now, but they really did try and do it, and thats why Wordstar doesn't exist and we have to use Word I was on the OS/2 team in the late 80s and it was light a trip back to memory lane as I watched IBM become totally shafted by Microsoft, and do nothing about it because it thought it was better than Microsoft. lol, how we laugh now....... Tuesday, November 4. 2008Dreaming In Code - A Review
I've just finished reading Dreaming In Code by Scott Rosenburg
While trying to write a review on Amazon I struggled with whether to give it 0, 3 or 5 stars, let me explain First the book is about the development of Chanlder, a next generation PIM sponsored by Mitch Kapor, ex Lotus Chief and with some of the big names in software development and open source, including Andy Hetzfeld whold wrote most of the original Max UI code. What strikes you while reading this is what a complete disaster the project was, they seem to spend weeks if not months "thinking" about the design, navel gazing like never before. The book seems to suggest the designers never once tried to use actual customers or possible customers to understand the problem domain but instead came up with wierd and wonderful designs from the UI down to the lowest level code, most of which were near impossible to implement The author then goes on to suggest, many times in the book, that software is hard and thats a fact. My god, its hard the way these people tried to develop it. On the back of the book I downloaded Chandler, and what a waste of several man years, and several million $$'s its basically a very bad PIM that is barely intuitive, slow and to be honest a bit c**p. So back to the review, if you want to read about the trials and tribulations of a complete mess of a software project, buy this book, 5 stars If you want to read about how not to design and develop software in the current internet age, buy this book, 5 stars If you want to read about how some of the apparent great minds in open source, are not really that great at working as a team in a real company, buy this book, 3 stars If you want the author to describe every computer programming term in the most basic definition, buy this book, 2 stars If you want to learn how to write great software, don't buy this book, 1 star Poor people, you really did get it wrong Monday, November 3. 2008
Agile, NFRs and the role of ... Posted by Keith Sterling
in Agile at
10:09
Comments (0) Trackbacks (0) Agile, NFRs and the role of Architectural Governance
Time and time again I come across the question and related issues of how does Agile deal with NFR's ( or Non Functional Requirements ). Those requirements which are often impossible to capture as stories. I have seen some teams ( and even tried my self early on ) to write stories such as As a <
I've become increasingly in favour of the use of architecture governance where by there is a specific role defined within the team who owns NFRs, and is responsible to ensure that they are adhered to within each story. The best solution I've found is that the Architect responsible for NFRs have an responsibility to ensure that each story is reviewed at some point with a view of the NFRs and the specific requirements are then capture as Acceptance tests on the story. Any story with data can state a critieria demanding Oracle, any story with UI can refer to usability tests Issues with this approach 1) Large agile teams may require more than one named architecture resource 2) Small agile teams may have this as a part time basis 3) Inexperienced architects may apply all NFRs to all Stories to "cover their backs" Monday, November 3. 2008Agile, NFRs and the role of Architectural Governance
Time and time again I come across the question and related issues of how does Agile deal with NFR's ( or Non Functional Requirements ). Those requirements which are often impossible to capture as stories. I have seen some teams ( and even tried my self early on ) to write stories such as As a <
I've become increasingly in favour of the use of architecture governance where by there is a specific role defined within the team who owns NFRs, and is responsible to ensure that they are adhered to within each story. The best solution I've found is that the Architect responsible for NFRs have an responsibility to ensure that each story is reviewed at some point with a view of the NFRs and the specific requirements are then capture as Acceptance tests on the story. Any story with data can state a critieria demanding Oracle, any story with UI can refer to usability tests Issues with this approach 1) Large agile teams may require more than one named architecture resource 2) Small agile teams may have this as a part time basis 3) Inexperienced architects may apply all NFRs to all Stories to "cover their backs" |
Calendar
QuicksearchArchivesCategoriesSyndicate This BlogBlog AdministrationTop Referrers |
|||||||||||||||||||||||||||||||||||||||||||||||||
