Prizes. The prizes in Startup Weekend events are not that appealing – you don’t get any material trophy, just fame & glory and something to help you continue with your startup.

Teams. In Startup Weekend events the teams get formed on the fly, with people that just met a few minutes ago. If you are in it for the win, this doesn’t work. You’ll want to have a well formed team that already knows how to work together.

Ideas. As mentioned in this blog post (“A bolder idea?” part), you’ll want to know more about the viability of the idea you’ll be working on so you’ll choose a project with more chances of success. A 100 persons debate on the market share or competition (things that the teams will do anyway) for all the 40-70 presented ideas would take all night and probably become unmanageable.

Collaboration. Different teams collaborating or helping each other is a common sight in Startup Weekend events. If the focus is on the competition, this would not happen – if you’re interested in winning there is no point in wasting time helping a rival team.

The fact that there is a winner and that there are prizes can be misleading to the participants in a Startup Weekend. The real value you get is the experience (from the “elevator pitch” all the way to the feedback received from the jury) and the networking. Everybody wins that!

I still think teams that do an exceptional job in one or all of the areas should be congratulated by the jury and get their “fame & glory”. Have perks for the teams to help them build their startup further.

Just don’t make it about the competition – that is another type of event, with a different dynamic and interaction between the participants.

As a former participant in the Startup Weekend events in Mexico City and Guadalajara, it was very interesting to me to see the event with the detachment and global view of an organizer. So, I’ll be posting some of the random thoughts I had.

(Part 1 here)

Canvas time

From what I’ve seen most teams spent almost all Saturday working on their Business Canvas. I guess the fact that we had printed canvases waiting for them on their work desks has something to do with it, I hope it wasn’t a bad idea. The downside was that only few got to do some “get out of the building” style customer development. And I think about half of the teams didn’t really get to validate their hypothesis at all.

Still lost on Sunday morning

There were teams that still struggled to find a model that could work on Sunday morning, with only a few hours to go until the final presentation. It always happens, I guess. It amazes me that they always seem to find a way and get the presentation done on time.

Design

I found it weird that a couple of teams decided to concentrate on great design and worry later about the business model. In the end they did come up with “the numbers” but I doubt they got to validate their value proposition with real potential clients. And in this type of events the judges will probably question your business, not the design or implementation, sorry.

Non profit? You still need the business model and customer development

There were 2 projects that declared themselves as “non-profit” or at least “not in it for the money”. What are they doing in an event that is focused on creating businesses? It turns out they need to validate their hypothesis about client needs just as a for-profit business. And, more importantly, they need to be scalable with few available resources, an even more difficult task in their case.

Community

One thing we knew when we started was that there was no organized startupper community in Monterrey that we could contact and spread the word. In fact one of the reasons we organized the event was to get this kind of people together. I hope I’m not wrong, but I think that this is the greatest success of the Startup Weekend Monterrey: entrepreneurs now know they are not alone, they are in contact, they want to do meet-ups and other events and to take it to the next level.

Monterrey pride

Unfortunately, Monterrey is going through some rough times. The drug-related violence of the last couple of years has made it a dangerous place to live. Just 2 weeks prior to our event, an attack on a casino resulted in 52 deaths and left the city and the whole Mexico in shock. That’s why I think there was a parallel agenda in our event. This was part of the citizens’ response. You could feel that all the participants at Startup Weekend Monterrey were proud to show that good things are being done here, that there are still a lot of people here that build instead of destroying. Once again in my time in Mexico, these people surprised me with their resilience. I felt humbled and proud to be part of it.

As a former participant in the Startup Weekend events in Mexico City and Guadalajara, it was very interesting to me to see the event with the detachment and global view of an organizer. So, I’ll be posting some of the random thoughts I had.

Most pitched ideas don’t seem so great at first

I’ve noticed that in the other events I took part too, but I blamed the fact that I was too concentrated on how to make my sale pitch and didn’t give them enough of my attention. Anyway, this makes the final presentations even more amazing, seeing what a good team can do with one of those ideas I thought was unclear or generic or impossible. I’ve learned my lesson to keep an open mind, and I realized I even came to expect the “miracle transformations” during the weekend.

Most people are so concentrated on pitching their idea that they are oblivious to everything else

We had a simple routine, where after pitching the person would go to the back of the stage and in order to write their idea on the panels used later for voting. It was funny to see one guy forgetting to do that, people making signs to redirect him, and then the next and the next, most of them failing to register what happened before their turn while waiting in line.

Put passion in your presentation

I’m  still working on this for myself. Also, please don’t use enterprise sounding long words, not with this crowd. Keep it real.

Teams form fast

After the voting, I was surprised to see how fast teams formed, in spite of the fact that, due to the architecture of the place, the team leaders were kind of spread and not visible. Before I knew it, the work areas were full and the teams already started to work.

“Deserters” happen

I really didn’t notice this in the events I took part as a participant, but apparently it happens all the time. A few people left because they wanted to work only on their idea. It made me sad to see among them a couple of my friends. Some had advances on their project and didn’t want to waste time working on another idea. That brings me to the next point:

New idea vs. project already on the way

Startup Weekend is about Lean Startupping, about doing customer development, business model generation – basically being able to change and adapt your idea. If you have a partial developed product, you’ve lost a good deal of your flexibility, the things you can change without throwing your past work away are limited. It’s way more difficult to get people to join your project because they would feel like they’re missing out on that experience and also not be able to assume ownership of the idea.

(part 2)

[MetadataType(typeof(JobMetadata))]
partial class Job : IGuidID
{
	public class JobMetadata
	{
		[Required]
		public global::System.String Title { get; set; }
	}
...
}

When using ASP.NET MVC’s ModelState.IsValid and the default MVC validation, everything goes as expected – the [Required] attribute is respected and you get the appropriate validation error message.

Now, if you need to run the validation manually, using Validator.[Try]ValidateObject(…), you’re in for a nasty surprise: the validation passes, ignoring your attributes.
After a lot of searching, I finally narrowed it down to MetadataType, got the right search terms and found the answer on Stack Overflow (of course): you need to manually register the MetadataType provider.

static Job()
{
	TypeDescriptor.AddProviderTransparent(
			new AssociatedMetadataTypeTypeDescriptionProvider(typeof(Job), typeof(JobMetadata)), typeof(Job));
}

Here is the link for the full answer.

This is an update to my Syntax Highlighting TextBox in Silverlight using MVVM post.

First of all, i’d like to thank Jeff Willcox for his port to Silverlight of Drew Miller’s syntax highlighting engine called ColorCode and Morten Nielsen for putting it together with my overlaying TextBox idea.

Morten’s SyntaxHighlightingTextBox control is just for demo (he mentions that a polished version will be included in their SDK), and the performance goes down when the text is bigger than just a few lines of code.

I rewrote the control in order to improve the performance – the goal was to get instant update when typing.

Changes:

Used a RichTextBox in the background instead of a TextBox. They both use Inlines for the content so the change is trivial, but from my observations the RichTextBox is faster, especially for scrolling.

Fixed a couple of bugs in the Colorizer.

Problem:

Syntax highlighting is done on the UI thread at every key press; refreshing the syntax caused the whole content do be cleared and rebuilt.

Solution:

Use in memory entities for formatting instead of updating the Inlines collection directly. This decouples the syntax highlighting code from the UI and also enables background processing. Also, it enables partial updating – basically updating only the Inlines that changed instead of clearing and reloading all the content for the RichTextBox. Even more, it is possible to update only the style (color) for existing Inlines if the text is the same.

Problem:

With the syntax highlighting in the background, the UI continues responsive, but typing takes place in the invisible overlaying TextBox and it is not visible until the syntax highlighting updates the background RichTextBox, and it is confusing for the user. (For big texts, the syntax highlighting can take seconds).

Solution:

Use a DiffTextBox as overlay. A DiffTextBox handles all key presses and updates a list of text transforms (insert, remove, move caret). Using that information, I can update rapidly the affected RichTextBox Inlines and get instant update for the background. The syntax highlighting can now be delayed so that it takes place after the user stops typing (configurable HighlightDelay).

There is still a limit of text size that the text box can handle – if the text is very big (dozens of pages), the Silverlight TextBox and RichTextBox slow down considerably. I’m thinking this could be fixed by managing “virtual pages” and loading only the current context (several pages) in the control so that it keeps being responsive.

You can see the control in action in this code sharing site I built: CodeReturn.com. You can upload code that you want to share and / or embed in your pages. The best part is that you can edit the code (and soon: post replies to other people’s code) directly in the page where it is embedded! You just need to login with any OpenID account!

Like this:

Also, you can grab the source code here: SyntaxHighlighting.zip (99.63 kb).

Working on a project that includes a database implies dealing with the schema changes and versions. The problem is that the tools that can help you are pretty expensive and (from my experience) not 100% reliable. It was so bad that my team and I developed our own database versioning mechanism and even started working on a tool to automate it.

What I want

I want to be able to deal with the database schema objects the same way i deal with the code files – specifically, I want to be able to see a svn history for them.

I want to be able to deploy a new version of the database to test and production servers AUTOMATICALLY, without worrying about data loss.

I want to be able to restore an old backup (like one with specific test data) of the database and update it to the latest version.

I want to keep the developer, test and production databases in sync (that is, the schema is the same for the same version).

I want the configuration data (read-only data like states and cities) to be included in the versioning system. [more]

VS 2010 Database Projects

With the latest release of Visual Studio, Microsoft has finally added a functionality that was loooong overdue: REAL Database Projects.

There are some nice posts by Deborah Kurata covering the basics here: Introduction, Using an existing database, Adding tables, Building and Deployment, Adding stored procedures.

I tried the projects for a new database, added a few tables, and then imported a script for various objects from another database.

I was pleasantly surprised to see that everything worked as I expected – the extra objects were added without compromising the existing structure!

For the next step, altering a table, I was a little baffled: should I edit the table on the development database and then compare the schemas, create a script and import it, or edit the CREATE TABLE in the sql file for the table?

The answer: any option would do! Of course, my favorite is the last, editing the sql file just like any code file. You can then get the changes script by comparing to another database (using a Schema Compare), or just hit “Deploy” and let Visual Studio deal with it!

So far so good

The scripts for the schema objects are maintained in individual files, so all you need to do is add them to your code repository and get object-level versioning!

Updating an existing database works like a charm (although I’ve only tried it with relative simple schemas so far). If you need manual touches, you can edit the changes script before deploying.

You can also add pre- and post- deployment scripts to handle your configuration data, for example.

My conclusion so far: Excellent work! What took you guys so long?!?

In the Plan Driven approach, a project is successful if it goes according to plan, so in software development it depends on requirements stability, on having clear and fixed requirements. As you probably know, that is a luxury most software projects don’t have, and a first approach would be to apply techniques like Change Management or Sign In Blood to contain the changes. Unfortunately that leads to unhappy clients and unusable software.

The Agile approach is to break the dependency on requirements stability and come up with a process that takes into account changes. It does that by using Adaptive Planning and Evolutionary Design.

Adaptive planning implies going through the project cycle many times, re-planning and re-adapting often.

Evolutionary design can be achieved with the help of practices like Self Testing Code, Continuous Integration, Refactoring and Simple Design.

If this made you curious, I highly recommend that you view the complete talk here:

Pourquoi, pas comment

USI 2010 : conférence incontournable du l’IT en France
Rendez-vous annuel des Geeks et des Boss souhaitant une informatique qui transforme nos sociétés, USI est une conférence de 2 jours sur les sujets IT : Architecture de SI, Cloud Computing, iPhone, Agile, Lean management, Java, .net… USI 2010 a rassemblé 500 personnes autour d’un programme en 4 thèmes : Innovant, Durable, Ouvert et Valeur.
Plus d’informations sur www.universite-du-si.com

• - add Config Transforms to the web.config (like web.Debug.config)
• - duplicate the sections with the values you want to customize
• - set the xdt:Locator attribute for the nodes that need to be identified by an attribute (like xdt:Locator=”Match(name)”)
• - set the xdt:Transform to control how the values are changed (common uses – xdt:Transform=”SetAttributes” to update some attribute values, or xdt:Transform=”Replace” to replace the whole node
• - you get the transformed web.config when publishing or creating a deployment package (or you can create it from command line).

More details on how to use the Web.Config Transformation here.

However, there is a nasty bug that you won’t probably notice before publishing for the first time.

Let’s say that you want to replace a value that is not an attribute, but the inner text of a node.

 <value>Default value here</value>

You can write the transform like this:

 <value xdt:Transform="Replace">Production value</value>

The problem is that in the generated web.config file the value will appear like this:

 <value>Production value

 </value>

The extra spaces (and newline) will be passed to the value and can generate problems if the setting represents a file path for example.

According to Microsoft, this is a known bug and will be fixed in the next Service Pack for Visual Studio.

The workaround is to trim the value before using it in code, but that won’t help you if the setting belongs to a 3rd party component that you can’t modify.

At TED2010, mathematics legend Benoit Mandelbrot develops a theme he first discussed at TED in 1984 — the extreme complexity of roughness, and the way that fractal math can find order within patterns that seem unknowably complicated.

Silverlight for Symbian

Silverlight includes a runtime that is optimized to display content on memory-constrained devices. Silverlight support for Nokia S60 5th Edition devices includes:

• The ability to view Silverlight applications in the mobile browser.
• Tools to build Silverlight applications that target devices

Tips and Tricks for INotifyPropertyChanged

As a WPF or Silverlight developer, you know that your models must implement INotifyPropertyChanged and it can be a pain. To do it safely, you really need to check to see if there are any registered handlers, then raise the event. To add insult to injury, the event arguments take a string, so if you mistype the property name you’re out of luck. Some clever individuals have created nice code snippets to generate the needed plumbing, but it doesn’t help with refactoring.

One common solution is to create a base class that provides the plumbing for a raise property notification.

Import Art from Photoshop and Make into Silverlight Controls

In this tutorial, we’ll take graphics created in Adobe Photoshop and Illustrator, import them into Expression Blend and then quickly turn the visual assets into interactive Silverlight controls.

Create a Custom Control – Inheriting from TextBox

When a control almost does what you want it to – if only it had another button or behaved slightly differently – you may be able to extend it by writing a custom control. Custom controls let you change an existing control or write a completely new control.

1. Set the DataContext in XAML to a StaticResource (like a ViewModel) created for design only, and change it to the real thing at runtime.

Good: use the existing class structure, XAML intellisense and autocompletion.

Bad: the design-time object will get created at runtime too, cannot access private set or readonly properties, must have a parameterless constructor.

2. Create sample data in Blend (XML).

Good: can be created by designers, can use private set or readonly properties, no runtime penalty.

Bad: must be kept in sync with the actual ViewModel, no compile-time checking, no intellisense support.

DesignData

The DesignData combines the advantages of the two options – you can create a XAML with your design values, based on the existing class model.

You can take advantage of intellisense and autocompletion and you’re able to set readonly properties, instantiate classes with no parameterless contructors.

The objects will be created only at design time, so there is no runtime penalty. You also get compile-time checking – you get an error if you try to set an invalid value to a int property, for example.

This functionality has been available since Blend 3, but it was (and still is) kind of hidden – in Blend 3 / VS 2008 you had to manually edit the project file to get it to work.

Fortunately in VS 2010 you don’t need to edit the project file, but there still is no Item Template for DesignData files.

[more]

In order to add a DesignData file, you can add a generic text file to the project and set its extension to “.xaml”.

Then change its Build Action to “DesignData”.

For this demo we’ll use a ViewModel that has different types of properties (public, private set, readonly, collection).

For the Entity class, we’ll have no public parameterless constructor:

We can go ahead and create our design-time data – notice that you have full intellisense support, even for the readonly properties.

If we add a TextBlock to our view, initially we have no databinding support.

We need to set the design DataContext for the container to use our DesignData source.

Now we can set the databinding easily:

And instantly get the values in the designer:

We get full support in Blend, too:

Of course the ItemsControl has no ItemTemplate set, so we get this:

We can create a template and take advantage of the databinding support in the template editing too:

We have a full image of how our design looks like with data, without having to compile and run the application.

In Visual Studio too:

Conclusion

Although still “hidden”, the DesignData functionality is the best option if you want to take advantage of the design-time support in Visual Studio 2010 and Blend, especially if you are using the Model-View-ViewModel pattern.

Update (thanks to Adrian Vinca):

You can create DesignData easier in Blend 4 – select “Create sample data from class…” in the Data pane:

Select the class you want to use:

You’ll get the generated schema:

And random data created depending on the property type: