Thursday, 25 June 2015

Agile methodologies are just tools

Agile methodologies (XP, Scrum, Kanban, DSDM, Crystal, etc) are for the most part made of really simple tools, so simple in fact that anybody could be told about them and start using them almost immediately

However, in my post agile isn't I said

even if you are doing everything a methodology says you should do doesn't necessarily make you agile.

I said this because the various methodologies are really just a bunch of tools designed to help you with your work, but tools alone won't make you agile.

A comparision

I have a friend who is a master carpenter, his toolbox contains many of the tools I have in my garage, and whilst I can use my tools my results are never as good as his. The reason for this is he has years of experience and a passion for working with wood, not only can he pick the best tool for the task at hand but understands how best to use that tool to get the result he wants. Combine this with his knowledge of wood, techniques for working with wood (joints, cuts, etc) and he will always be better at working with wood than I am.

I could practice with my tools and gain insight into how best to use them and which one to chose but it wouldn't make me a master carpenter, for that I also need the other knowledge my friend has about wood and techniques for working with wood.

Could I learn this? of course but from seeing my friend work and talking to him I can see its not something you learn out of a book, you do it by practice and feedback from more experienced people.

How's this relate to agile then?

As I mentioned before agile methodologies are made up of what are really simple practices and people read books, blogs or attend a 2 day course and belive that they understand how to use these practices.

However, this is no different to me having a bunch of tools in my garage, I know what they are and I have a basic idea of how to use them but I'm not going to get the best results.

This brings me back to my quote from my earlier post, unfortunately a lot of agile implementations out there are done by people who only have the tools they have no passion for it and haven't tried to learn more about not only the practices they follow but agile itself, leading to some of the nightmares you hear about agile in the work place.

It takes time, experience and practice to take these simple tools and understand how best to use them to get the most out of them, you need to spend time listening to other experienced practioners, talking to them, asking questions, etc.

Although my post Agile is... outlines some of the concepts underlying agile, at its heart its about people and tools will only help they won't do the work for you.

Friday, 5 June 2015

What does a good agile environment give you?

In my last post I talked about what made a good agile environment and the post before that I mentioned practices that people often mistake for being agile.

When you put the concepts and practices together the result should be an agile environment with everybody aligned towards the same goal


One of the biggest gains that you get in a good agile implementation is the increase in trust between the organisation and the team.

Historically projects have failed and organisations have lost trust in teams to deliver so they then started implementing tighter project management in an attempt to ensure projects succeed, which frequently leads to more project failures.

The trust can be restored with the team increasing communication with the rest of the organisation by being transparent with the work (tracking work on a card board, charts showing progress, meetings, etc) and through collaborating with the rest of the organisation about work being done and the work to be done.

If the team is trusted to deliver the focus moves from when something will be completed to what is to be completed and how soon it is needed.


Frequently work is prioritized based on a vague idea of what the business thinks is best for the product but often without any context of how simple or complicated a feature will be, collaborating with the team can feed in to discussions around the cost of developing that feature versus the benefit the business anticipates it will get from it.

Value itself is a tricky thing to pin down: is it about money, cost, user experience? Recently there have been some posts (notiable this post and this post) discussing value and showing that it is a very nuanced area.

The good thing about this is value will need to be considered and discussed amongst everybody involved across the business so that everybody understands value around the work being done it can help them with making decisions themselves as to priority of the work.


One reason that organisations want to adopt agile is because they are under the impression that it will speed up development, they are then disappointed when following a methodology the work isn’t being done any faster.

The potential increase in speed, and it is only potential, comes about mainly because with better engineering practices, building the features that are of most value and focusing on creating working software the business ends up being able to use/sell the software in days/weeks rather than months/years.

The people aren't actually creating the work any faster than they were before it is just that they are having to do less rework which means the software is available to the end users sooner, and there should be less defects meaning less time needing to be spent on support all of which combines to allow more time to be spent developing the product in the first place.


Good agile implementations don't restrict themselves to looking at the team creating the work, they look at the organisation as a whole and work out everything that is involved in creating a product, how the work flows through the business.

Once you start seeing the team as only part of the overall flow it may become apparent that there are both upstream and downstream activies which are causing problems for the organisation and it is these areas that need to be tackled to improve rather than just focusing on the team.

A side effect of this can be to question the need for estimates, if you have a team that is reliably deliverying work and is only 1 part of a bigger process what value does the business get from estimates from the team? This isn't to say the business isn't going to want estimates but it is understanding the value of that estimate in context to the larger organisation versus the value from the thing they want to create.


Creating an agile environment is hard, you can take a team and implement an agile methodology but that is only a small part of the story. Done right agile will involve a lot of the organisation (if not all of it), it changes the way people in the business look at the work and focuses them on what is being produced and the part everyone has to play in getting that work done.

More than that though it can prompt people to look more deeply into various aspects of running the team/organisation such as value, reward, motivation and learning to name a few.

Hopefully these posts have given you an idea of what agile actually looks like rather than the perception that it is a methodology or just a few meetings, charts and user stories.

Wednesday, 3 June 2015

Agile is...

My last post covered a lot of things that people associate with agile and how just doing those things don't make you agile, this time I'm going to cover things that good agile implementations tend to display.


The keystone of all good agile implementations is good communication.

The last post listed meetings & visual work tracking which are are all ways to communicate but what is key is that an agile team doesn't restrict themselves to just these methods, they will communicate whenever they need and not restrict themselves to "set" meetings.

For me agile revolves around communication - communication between members of the team, the team and the larger business, the business and the customer, etc. From communicatin we get


Another key part of a good agile implementation is collaboration, it forms the basis for working in the team and for the team working with the customer/business.

I mentioned capturing user stories as requirements in the last post, these should involve collaboration between the customer/product owner/product champion and the team. The team shouldn't be separate from these discussions it should be a part of them able to provide not only technical insight but frequently a different view on the product as the team has unique domain knowledge from working on the "product".

If there is no collaboration then people tend to fall into silo's & lose the view of the whole product as well as fostering the mindset "it's not my area/problem/issue/job".


Whilst collaboration is about working with other people toward a single goal,cooperation is about working with others towards your own goals. Think about this. In a team everybody is doing their own work to pursue the larger goal that the team is working towards, but an agile team co-operate understanding that everybody needs to complete their work for the team as a whole to succeed.

What does this mean in practice? An example isif one team member finishes their work and can see other people still working on theirs they offer to help them complete existing work before starting anything new.


For a long time developers were told what they had to work on, when they should work on it and often how long it should take them to complete it, they had no influence over what they did or when.

Agile turns this on its head, collaborating with the business the team should be involved in all aspects of the work from generating user stories, prioritising the work, developing, testing, etc.

This apparent reversal of traditional control is done to allow the team to be able to make decisions quickly, rearrange the work based on changing circumstances/priorities to enable them to deliver the work.


Work needs to be completed & users using it for it to be of any value to the business, people working in an agile environment understand this.

Don't confuse this with the team committing to complete work in a sprint from the older Scrum guides, this is the business being able to trust the team to complete the work, this doesn't mean working ridiculous hours to finish by the end of an iteration(although the team may decide to do that), it is knowing that when a piece of work is started (barring any changes in priority or direction) the work will be completed and delivered.


Agile isn't static, when you first start you may have a framework that you work within but as a team becomes more experienced they use techniques like retrospectives to determine how they can improve what they are doing.

Should you wait till a retrospective to make a change? No. Just like you shouldn't wait until a meeting to talk to other people there is no reason why you can't make changes to the process outside the retrospective.


People will often point to the agile manifesto and say "that's agile" but it doesn't tell the whole story, other people will point to the agile principles which provide more examples of how an agile team should work but still can't convery the subtleties of working in an agile environment.

Underlying both the manifesto and the principles are the concepts I've outlined above, regardless of if you are practicing an agile methodology or not, you can tell if a team/organisation is agile by observing them and seeing if their culture embodies the concepts listed here.

In my next post I'm going to go into how some of the benefits of being agile, often listed as reasons to go agile, are in fact side-effects of having the right culture and mindset based on these concepts combined with the practices from my last post

Monday, 1 June 2015

Agile isn't...


Various agile methodologies outline specific meetings but holding these meetings doesn't make you agile

Capturing requirements as user stories

You aren't writing large requirements documents, you're capturing the details using a form of "As a <user> I want/need <functionality> so that <desired result> This is a good way to focus on functionality that you need for the system (it does have its pitfalls and nuances) but doesn't make you agile

Visually tracking work

Having a board that shows progress of work is really good to help people understand what the team is currently doing and what they've done, similarly charts summarising progress can help people understand trends but doing this doesn't make you agile.

Engineering practice's

Your team practice TDD, pair programming, continuous integration/delivery/deployment, that's awesome and will help you a lot in your day to day work. These practices aren't restricted to agile working so this doesn't make you agile either.

Following an agile methodology

I can hear people say "we do all that, and we have people dedicated to ensuring we follow the process etc" and I would not doubt you but from experience I can still say that even if you are doing everything a methodology says you should do doesn't necessarily make you agile.

So what is agile?

The items listed above all have value in them and I wouldn't suggest you stop if you're doing them already. Just don't think that by doing those you are agile.

In my next post I'll detail what things are common to successful agile implementations that I have seen.

Monday, 30 March 2015

Emailing support

A while back I was working with a colleague and we ran into an issue that required us to email a company for support.

My colleague expressed an interest in how I went about structuring the email and asked if I could write up my process for them, hence this blog post.


The way I structure any of my mails, to support or otherwise, is along the lines of:

  • Explaining reason for the email
  • Provide any information I may have on subject
  • If asking for something clearly state what I need/would like

In addition if I am able to provide anything to help, such as test project with failing tests, they should be attached or easily accessed.

One of the additional things that I take into account when writing an email for support is thinking “what would I need to help


Below is an example email written about a problem with a fictitious component aBxZ:

Hi Support,

Here at we are using your aBxZ component and are concerned about the number of duplicate calls it is making to the api and can see no way to stop this.

We have been using Fiddler to profile aBxZ’s behaviour on our QA environment whilst testing with a similar load to our production environment and have attached the data we captured.

If you look at the captured sessions you’ll see that aBxZ makes a call to for every request, this is happening even when we don’t call aBxZ explicitly.

We read the documentation and ensured that we have followed the instructions and guidance on how to use aBxZ and we posted on StackOverflow about it but nobody was able to provide a solution.

We managed to replicate the behaviour in a test project, which is attached, and it appears to be happening when aBxZ is instantiated.

Please can you let us know:

  1. Are we using the component correctly?
  2. If using it correctly is there anything we can do to stop this?
  3. If it is a bug with aBxZ when you think you might have a patch/fix

If we should be doing something different to what we do currently please could you alter the attached test project and send it back to us.

Looking forward to your reply

Analysis of the email

Looking at the mail it follows my structure and breaks down like this:

  • Paragraph 1, say who we are and what the problem is
  • Paragraphs 2 – 5 provide details about what has been tried and provide to them additional information in form of captured fiddler sessions and a test project
  • Paragraphs 6 & 7 ask, hopefully as clearly as possible, for help to resolve the issue.


Hopefully this is of use to you, for me it seems to be very clear but I’d be interested to hear from other people how they structure their emails and if you think I’m missed something out.

Thursday, 26 March 2015

Simple.Data Asp.Net Identity Provider

ASP.Net Identity is Microsoft's latest implementation of a membership and identity service which will replace ASP.Net Membership, SimpleMembership and Universal Membership.

The out of the box implementation of ASP.Net Identity is, unsurprisingly, coupled to Entity Framework (EF) which means if you do want to use the new identity functionality but your app doesn’t use EF for data access you either end up having to include it in your project anyway or not use Identity.


On a project I was part of towards the end of last year we ran into this exact problem, we were using Simple.Data for data access but wanted to use the new features of Identity.

One of the advantages of identity over previous membership implementations is that it is possible to provide a different data access layer allowing you to use the database you want but keep the rest of the functionality provided by identity e.g. password encryption, OAuth, etc

The advantage of using Simple.Data is that you aren’t limited to connecting to any one database as you simply include the relevant provider for the database you want to use (if it is supported).

The Provider

The result is a Simple.Data provider for ASP.Net Identity which should be able to dropped into an existing application using EF with little to no change.

You can install it from Nuget using the package manager console with the following command:

PM> Install-Package Simple.Data.AspNet.Identity

If you prefer to use the “Manage Nuget Packages” GUI it is easiest to search for Simple.Data:


If you want to just look at the code or feel like building it from code yourself the repo on GitHub can be found here.


There are a few limitations with this implementation that its worth knowing about:

  • Due to ASP.Net Identity being designed around generics you must use concrete classes with it, you cannot use dynamic.  To help the provider comes with 2 classes IdentityUser and IdentityRole classes for you to use.
  • Database tables – currently you need to use database tables that match the IdentityUser and IdentityRole classes although they don’t have to be named the same.  The Nuget package includes a script for Sql Server to create these tables.
  • Relies on referential integrity – when deleting data ASP.Net Provider does delete any data that may be effected, instead it relies on db referential integrity to clean up linked records.
  • Ids stored as strings – although ids for users, roles, etc are generated as Guids they are stored as strings, this mimics the standard EF model.
  • No IQueryable support – Simple.Data doesn’t support IQueryable this means that the Users property on UserStore and Roles property on RoleStore will throw not implemented exceptions.  If you need to query Users and Roles you need to do that yourself
  • Only works with Simple.Data v1 – a version is coming for Simple.Data v2 (which supports async await) but this is still a work in progress.
  • Db Connection – the provider supports using Simple.Data default connection or a named connection, it currently doesn’t support taking a connection string directly.

Although you need to use the identity classes in the provider you can inherit from them just like the EF code.


To try and make it as easy as possible to get started with this provider I have built 2 sample applications based off of the normal MVC project template:

  1. MVC 5 Standard Sample – this is the standard MVC project template the only change is that it uses Simple.Data for data access.
  2. MVC 5 Sample – this sample is an extension of the standard sample as it includes using roles with identity to be able to control access to pages of your web application

The easiest way to work with these samples is:

  1. Create new ASP.Net Web application
  2. Select the Empty project template
  3. Once the project has been created add the appropriate Nuget package

If you want the standard sample then at the package manager console type:

PM > Install-Package Simple.Data.AspNet.Identity.MVC5.Standard.Sample

Or if you want the sample that includes roles:

PM> Install-Package Simple.Data.AspNet.Identity.MVC5.Sample

When you install these packages you will have all the necessary classes and references added to your project and it will create a database in localDb so you should be able to build and run the code to start working with it immediately.

Both sample projects are in one repo which you can find here if you wish to play with the code that creates the samples although installing the Nuget package is by far the easiest way to get working with the code.

The future

This version is 1.0.0, it is stable and should work without problems (taking limitations mentioned into account).

Going forward there are several improvements that I intend to make:

  • Simple.Data v2 – complete work on this implementation
  • Support connection strings – to match normal Simple.Data connection options add the ability to be passed a connection string.
  • Support different id variable types – normal ASP.Net Identity allows for different variable types for ids via more generics, I’m looking to add this in.
  • Automated builds – looking to use AppVeyor to build and test the code
  • Automate Nuget publishing – once AppVeyor is building the code look to automate generation and publishing of the Nuget packages
  • More samples – add samples showing use of the provider in different ways e.g. not tied to Owin context


This provider should make it easy for people that wish to use ASP.Net Identity with Simple.Data right now.

With the planned future improvements the additional functionality will bring it up to parity with the full EF model.

Tuesday, 24 February 2015

Node School–Bristol

Last Saturday I got to attend the first NodeSchool to be held in Bristol.

The event was held at the MixRadio offices in the heart of Bristol who were sponsoring the day by hosting the event, the room was at the top of their building looking out over Bristol which was brilliant.

The Day

I, like most attendees, started arriving about 9:15 finding somewhere to sit and getting their laptop ready.

Just after 10 the event began with an introduction telling us all about the day, what we could expect, how to get help and thanks to the sponsors that had made the day possible.

The format for the day was to follow the tutorials in the learnyounode workshop from NodeSchool (which you can install using npm install –g learnyounode) the difference being that instead of having to do it on your own, and try and complete the exercises simply by Googling for solutions, there were “mentors” on hand to help if you got stuck.  The mentors were volunteers who had experience with node and could help solve the exercises in the workshop.

The original timetable for the day had attendees having a break at 11 but most people were heads down working on the exercises with mentors helping out as required.

Lunch was supplied by JustEat, who were also sponsoring the event, who provided some awesome pizza for everyone to eat.

After lunch there was a short update on what the plan was for the afternoon and other tutorials that you could look at before we carried on with the exercises.

At 4:30 we stopped for a talk from Adam Butler on “5 Things to build with NodeJS” the highlight of which was Adam flying a couple of different drones using node.

After the event there people went for drinks at Big Chill which unfortunately I wasn’t able to attend.

The workshop/exercises

The learnyounode workshop code is structured into a series of exercises that take you from “Hello World!” through multiple async http calls to creating your own http server.

Everything is done at the command line with learnyounode providing a list of all the exercises and marking each as completed when you successfully complete an exercise.

When you start an exercise you get told what it is you need to do and then get a “Hint” giving you more information about how to accomplish what it is you need to do, once you’ve written your solution you verify its correct using the learnyounode inbuilt test harness. 

If your code doesn’t pass the test you get test failure output showing what was expected and if you pass you get shown the “official” solution to the problem.

My Day

During the day I managed to complete the entire workshop, which was the intention for everybody.

I also managed to complete the Express workshop, ExpressWorks, and what was interesting (to me at least) was how noticeably easier it was to use node with a framework like express after having had to code at the lower level for most of the day. 


It was a really good day in a fantastic venue, we were fed watered and looked after really well throughout the day, I must give a big thanks to Katja ,Tristan, Tom & Adam for organising the day.

Although I wouldn’t say I’m now a proficient node developer I feel I have a much better idea of how node works and also through doing the Express workshop can see how easy it would be to put  together an application on node.

I know that they are planning future events and I would recommend attending one if you want to explore the world of node but were unsure where to start.