[Post Event] Web Codecamp event in Cluj-Napoca, September 28, 2013

Yesterday, in Cluj-Napoca we had “Web Codecamp”, an event organized by Codecamp Romania for web developers. We had 4 great session where we talked about Backbone.js, AngularJS, SignalR and continues delivery of web application. There were more than 100 attendees that stayed almost until the end of the event – thank you for your patience.

Presentation content
Backbone.js - Dragos Andronic
Slides
Source code
Video

AngularJS - mai multa actiune, mai putin cod - Simona Avornicesei
Slides
Source code
Video

Automate all the things! - How continuous delivery can save your sanity - Attila-Mihaly Balazs
Slides
Source code
Video

Real-time fluent communication using SignalR and Cloud - Radu Vunvulea
Slides
Source code
Video

Pictures from the event

Special thanks for our sponsors:

Read More

Big companies, Software Engineer and Partnerships

Some weeks ago I talked with a friend that used to work in a big company. He told me two interesting stories.

Story 1: You are working for a client on a project for 1-2 years. Things are going very good, you’re company decide that they will allocate another person in the team. The new developer will be a junior and he will take your place on the project. Because they allocated a new person in the team, you will be allocated only 2.5 days per week on the project. The rest of the time you will work on another project. They ask you to join the scrum every day and pretend that you are working all the week on the original project. The new member of the team will cover you. The client knows that the team is now n + 1 (the new member). So you begin to join the scrum every day and pretend that you are working…
First problem is sincerity. Going in a scrum meeting and pretend that you are working on the project on a specific day, but in reality you are working on another project is wrong. From the relationship perceptive you are destroying the trust that you have with the client. Yes, you think that he don’t knows, but in time he will observe that something is not okay.
Don’t forget that at the end of the phone another person like you exists. When you are pretending that you are working on the project – you are lying him.  
A software engineer should never accept a request like this from the company because a thing like this affects him professionally. You cannot name yourself a professional when you are doing things like this. The company cannot force you to lie.
If you need to work for one or two days one another project you should tell the client that you will not be available on that days. But, in this situations your company will request you to not tell them the true.
The best thing that you can do in this situations is to tell them that you don’t want to work anymore on that project. You cannot accept this “business” model.

Story 2: Your company has a very good partnership with another big company. Your partner comes and ask you to investigate and search a solution for their problem. After 2 days of investigation, your boss is coming to you and say that the both solutions are good, but you should hack the “investigation” in such a way that the second solution to be selected. This should happen because the team that has knowledge on the given technologies doesn’t have any project on the queue. You know that the first solution is better, so what should you do?...
This is a very common situation and I heard people saying that this is the way how businesses are made. Maybe this is true, but when you propose and push a solution that is not valid for the client you will suffer a lot – and you deserve to suffer.
When you will be ready with version 1.0 or in the moment when you will add some new features you will realize that is impossible to accomplish and the costs are very high.
In this situation, the guilty is divided between your company and you. First of all they push a wrong solution and now they will need to pay in one way or another. If the project was a fixed price project then they have big problems. Otherwise, there are chances that the client will pay without knowing the true, but in the end the partnership between them will suffer.
The software engineer that pushes the wrong solution usually is the one that will need to solve the new problems and it will not be easy for him. From the beginning he is making a big mistake, listening the management “recommendation”. The best thing that can be done in this moment is to talk with the client and explain the mistake – the current solution is not the best one and they should use another solution.

Read More

New team member and code guidelines

Being a developer doesn’t mean that you are staying alone in a small room, in front of a computer and writing in a mystic programing language. Being a developer means working in a team, with real people, where you need to communicate and being part of a team. The code that is written by a developer will end to be modify by another developer from the same team or company. This means that is very important to write the code to respect the team and project standard.
But what should you do when one of guidelines are not part of your ‘believes’? For example adding a comment to each public class/interface or method.
There are developers that will NOT respect the guideline, will fight for it until the end of the world and will say NO – ‘I will never respect this guideline because is bullshit’.
This is an extremist approach and I think that this approach is not better than the one when you accept everything. In this cases the most important thing is communication. You need to understand why that guideline exist (it came from the team, architect, PM, company or client).
You should explain why the guideline is not good, but be aware, one time is not enough, because you may have people that use that guideline for 10-15 years and for them is a religion. Small steps are the best option in this cases.
But if none of the solutions works and all the team say that the guideline is good then you have two options:
Respect their decision, write code using their guideline (you didn’t lost the fight, but you need more mitigation – time is your best friend)
Leave the team and the project (the extremist developer : - ) )
The worst thing that you can do is to write code using your own guidelines. Even if you don’t like the guidelines, you should not ignore their guidelines. If you have 1 million lines of code written using a specific guideline, starting to write a part of the code with your own guidelines is a big mistake.
It is okay to write a small PoC where you can show them why that guideline is wrong, without making a push to the main repository.
In conclusion we could say that being a new member in a team can be hard, but you need to be a team player and accept other believes also.

Read More

Types of Subscriptions Filter on Windows Azure Service Bus

Service Bus from Windows Azure is a great services to distribute messages to multiple subscribers. A topic can have one or more subscriptions. A message added to the topic will be send to all subscriptions that are register in that moment.
To be able to control what kind of messages are send to each subscription we can define filters. A filter is added on each subscription and specify what messages will be accepted by subscription.
The types of filters that can used are:

Correlation Filter
This is used when you want to accept messages that has a specific correlation value. This value can be set directly to the BrokenMessage and is great if you have messages with different priority or rank.

CorrelationFilter filter = new CorrelationFilter(“low”);
namespaceManager.CreateSubscription(topicPath, subscriptionName, filter);

BrokenMessage message = new BrokenMessage();
message.CorrelationId = “low”;
…

SqlFilter
This filter give us the ability to specify what kind of messages will be accepted in a subsection based on a custom rules. This rule can be related to the properties list of a BrokenMessage that are specific by the producer or system properties (SessionId for example is a system property).
The syntax is very simple and is based on SQL92 standard and T/SQL. If you know to write a query in SQL than you will not have problem to use this filter. The syntax give us the possibility to use = != < > <> IS IN LIKE EXISTS ESCAPE NO + - * / AND OR …
For a normal use this is enough to define rules over subscriptions.

SqlFilter filter = new SqlFilter(“CustomProperty == 1”);
namespaceManager.CreateSubscription(topicPath, subscriptionName, filter);
…
BrokenMessage message = new BrokenMessage();
message.Properties[“CustomProperty”] = 1;

The “user” or “sys” prefix need to be used when you refers to properties that are in different scope. The “sys” refers to properties that are on BrokenMessage (like Correlationid or SessionId).

FalseFilter
This is a filter that can be used to block all the messages to the given subscription. This can be used if you create a subscription and you don’t want to accept message to it (temporarily). Also, this can be used when you have a subscription and you don’t want to receive message anymore to it, but you don’t want to delete it.

FalseFilter filter = new FalseFilter()
namespaceManager.CreateSubscription(topicPath, subscriptionName, filter);
…

TrueFilter
This filter is similar with SqlFilter. It will accept the message that respect the specific rule.

We looked over what kind of filter can be defined over a subscription. I think that FalseFilter can be very useful when you are in production and you need to control what messages are receive by one of subscription. For example in a scenario when a client didn’t pays the monthly subscription and you want to stop temporary his subscription. Is more simple that deleting it or set a SqlFilter – “ 1 <> 1”.

Read More

Web Codecamp event in Cluj-Napoca, September 28

In September 28th, Codecamp is organizing in Cluj-Napoca an event that is dedicated to WEB developers. We will talk about Backbone.js, AngularJS, SignalR and how to make continuous delivery for web projects.
You can register here: http://codecamp-cluj-sept2013.eventbrite.com/
Agenda:

09:15-09:30	Sosirea participanților
09:30-10:30	Backbone.js Dragos Andronic	Backbonejs - the library (not the framework). What can you expect from it: short introduction to its components.  Where does its role stop: what extensions are there on the field. Some thoughts after using Backbonejs in production code
10:30-11:30	AngularJS - mai multa actiune, mai putin cod Simona Avornicesei	AngularJS este o librarie javascript construita de Google si multi alti contribuitori din lume in jurul ideei de simplificare a muncii dezvoltatorului web. In aceasta sesiune vom construi o aplicatie in ASP.NET MVC si AngularJS si vom discuta despre cum poate fi introdusa in proiecte deja existente, cu impact minim.
11:30-12:30	Automate all the things! - How continuous delivery can save your sanity Attila-Mihaly Balazs	The talk will present solutions to automate the boring and repetitive aspects of software development (build, test, quality inspection, deployment). We will look at concrete software solutions (Jenkins, Git, GitHub, Gerrit, SonarQube) and how they can be integrated together to save software development teams countless hours in repetitive tasks.
12:30-13:30	Real-time fluent communication using SignalR and Cloud Radu Vunvulea	This session will explore how SignalR can make our life easier to create real-time application that works on any kind of device or browser.

Read More

Why we have two accounts key for Windows Azure Services?

In today post I want to talk a little bit about the account keys from Windows Azure. There are a lot of services from Windows Azure that can be consumed using an account name and key. The account name is only one and in some situations is represented by the service namespace. In all cases the account key is duplicated. There are two account keys that can be used in parallel. Both of them can be used in the same time without problems.
But why we have two accounts key?
Well, we have two accounts key with a very good purpose. Let’s imagine the case when one of the accounts key is compromised. In this case you need to generate another key and distributed it to the clients. From the moment when you generate the new key, until you distributed it to all the clients, the specific service will not be able to be used – if you use only one key.
In the situations, when a key is not valid anymore, clients could fall back to the second key and start to use it. This mean that all the clients should access to the both accounts key.
The same scenario can be used when you need to regenerate the accounts key at a specific time interval (every 3 months). In this cases you can define a process to regenerate the accounts key without affecting the clients.

Read More

Service Bus and Shared Access Signature (SAS)

Until few months ago, Shared Access Signature (SAS) could be used only wilt Azure Storage (Blobs, Tables and Queues). From now one we can use SAS with Service Bus. We can define SAS to use in combination with topics, queues or notification hub
This security feature is pretty great, especially when you have an application that use 3rd party. You don’t need any more to share with 3rd parties the account name and key. From now one you can give them a unique token that give them access to some part of your namespace services and also limit what kind of operation they are allow to do.
In this moment you are allow to define only 12 rules in a namespace, but in the near future I expect to be able to define more than 12 rules.
The access rights that can be controlled in this moment are:

• Listen – to be able to receive message
• Send – to be able to send messages
• Manage – to be able to manage the resource

From Service Bus perspective, this access rights are enough and in combination with expiration date it works pretty good. Be aware, that you can set Manage access rights, only if you set also the Send and Listen rights.
The interesting thing is how the rules are defined. The rules are defined over a URL. This means that you can define a rule over a namespace or over a specific topic or queue. The SAS will be valid for all resources under the specific URL.
When you generate a SAS rule, two keys will be generated for the same rules. Both keys can be used in the same time. This is done to help customers in the moment when they need to generate new keys and they don’t want to block the access to Service Bus.
The SAS rules can be defined and manage not only from code (using REST API) but also from the management portal of Azure. The portal can be very helpful when you are in the development phase or you have some issues with SAS and you want to check what rules are already defined.

How to create a SAS from code
First step is to create a SharedAccessAuthorizationRule and set the specific rights.

SharedAccessAuthorizationRule saar = new SharedAccessAuthorizationRule(
    “myFooName”,
    SharedAccessAuthorizationRule.GenerateRandomKey(),
    new[] {
        AccessRights.Manage, 
        AccessRights.Listen, 
        AccessRights.Send }));

GenerateRandomKey method is used to generate random keys.
Once the rule is created, you can add it to the queue, topic or notification hub using the description class

QueueDescription qd = …
qd.Autothorization.Add(saar);

Don’t forget to save the name of the name and the key of the rule.

How to use a SAS key
Once you created the rule and have the name and the key of the rule, you will be able to use them in the moment when you create the MessagingFactory.

MessagingFactory mf = MessagingFactory.Create(
    "uri,
    TokenProvider.CreateSharedAccessSignatureTokenProvider"myFooName", "myFooKey"));
QueueClient queue = mf.CreateQueueClient("myFooQueue");

Enjoy!

Read More

Big team and Commitment

Managing a big team is not an easy job, especially when you have a customer that make a lot of pressure – in the end all customers make pressure, you need to deal with it.

What you should never do to a team?

Announce that all the sub-teams structure will change from today - changing the scope of each of them and also the people.

In this scenario people will be disoriented. All the tasks and plans will remain in a gray state. Even if you will continue after 2-3 months will not be the same. After you make such a change, you should allow some time to sub-teams to finished the ongoing tasks and let the sub-project in a stable phase. In this cases the team members will feel that there is not communication inside the team and the only scope of team leaders is to make promises.

When you are in the head of a team you should not accept death-lines from the customer without talking with the team. Making such commitments can generate a lot of problems:

• You will not be able to deliver in time
• The team will be stress
• A lot of overtime
• The quality of the code will decrease
• There will be a log of bugs
• Testing team will not have enough time to make the test
• And many more

And again, you will not be able to deliver in time. The client will not be happy, even if he says that it is fine for him.

In such situations, the testing team will have a lot of headaches because they will don’t have enough time to make all the testing (from new features testing to regression tests). The testing team should never accept a new version of a product without having enough time for testing. They are like a defense tower that can discover issues that in production can ruin not only the development team but also the client.

All this problem can be avoided with the simplest thing: COMMUNICATION

Read More

Task - run on UI thread (update UI)

Working with Tasks is pretty easy. You can do very easy things that would be pretty complicated to do using threads. In today post I want to talk a little bit about UI threads and tasks.
If you are working on a Windows Store App, Windows Phone or WPF application then you will have the following case:
From a task I want to be able to update a UI component (a label for example).
By default this is not possible, if you are trying to run the task from the UI thread then you don’t resolve the problem. The task will still run on the UI thread and if you have a long task then you will block the UI.
The solution is to run only the lines of code that update the UI components on a task that runs on UI. If the code that updates the UI is at the end of the task then you can have something like this:

var blueTask = Task<string>.Run(() =>
  {
    ...
    return value;
  })
  .ContinueWith(value =>
  {
    ...
    this.MyLabel.Text = value;
  }, TaskScheduler.FromCurrentSynchronizationContext());

If the code has to display a dialog box and you need a response then you can create a task inside your task that runs on the UI task:

var uiScheduler = TaskScheduler.FromCurrentSynchronizationContext;
var blueTask = Task<string>.Run(() =>
  {
    ...
     await Task.Factory.StartNew(
              () => this.MyLabel.Text = value,
              CancellationToken.None, TaskCreationOptions.None, uiScheduler);

    return value;
  });

Enjoy!

Read More

Load Test - Binding Context Parameters

The new Load Tests features that are coming with Visual Studio 2013 are great. In combination with Windows Azure and TFS Controllers, you can run load test using cloud infrastructure without having to deploy or configure something things on the cloud.
When developers start to create different web test that will be used for load tests they usually them in sub-tests and reuse them when is possible. For example very easily you can end up with something like this:

• Foo1WebTest
• Foo2WebTest
• Foo3WebTest
• BigFooWebTest

Where BigFooWebTest make calls to Foo1WebTest and Foo2WebTest.
If you started to extract context parameters from each web test you will observer a pretty odd behavior. By default it is not possible to bind a context parameter. This means that you will not be able to fetch data for a context parameter from a data source like a csv file or data base.
Even if you will write to the value of the parameter context the binding command (“{{..#TableName.CollumnName}}”) you will observe that this will not work. When you will run the test, the value of the context parameter will be the string value itself “{{..#TableName.CollumnName}}”.
In this moment it is not possible to bind a context parameter. The only available solution in this moment is to write a plugin for the test that will load the values from data source in memory. This plugin will be able to set the value of context parameter with the memory values.
For more information about load test plugin: http://msdn.microsoft.com/en-us/library/ms243153.aspx

Read More