Develop Reference

How to avoid rewriting a VB6 from scratch when migrating to .Net

In our company we develop and sell a VB6 application and we think it's about time to migrate it to .NET.
The main reasons are:
We expect VB6 runtime support to end at some point in time, and we do not want to start the migration just then since it's probably gonna be a lengthy process.
There is just 1 1/2 VB6 developers left. The half one being me.
More and more customers asking for features like cloud and mobile device support.
I know that rewriting an application from scratch is the least recommended way for migrating to .NET. I totally aggree with that! Throwing away over a decade of code feels just wrong and would be such a waste of money spent, that I have a hard time recommending and justifying it towards our management.
But right now I don't see another way to do it.
Let me tell you a little bit about the application:
Like I said it has been developed for over a decade. There have been numerous developers working on it, most of them rather unexperienced at that time. We have one developer left from the initial team. That application has been his first and biggest software project and by now he realizes that many of the architectural decisions made over last 15 years have been horribly wrong, others were right at that time but have not been refactored to meet changes made in other parts of the application and so have become wrong at some point in time. This application seems to be a showcase example of code rot.
We are talking about an application of about 150 KSLOC, all in one single executable. It uses about 15 external DLLs, some of them third party ActiveX controls, some of them are our own .NET assemblies.
Adding new features to the application is still possible and being done, but takes ages compared to our other .NET applications. The reason is that every little change in the codebase requires changes all over the place. The only reason why changes are possible at all is because that one developer simply knows most the dependencies and quirks of the application. As you might have guessed the rate of unexpected side effects and bugs is quite high.
My first thought about migrating that application was to first clean up and refactor, then migrate/convert possibly using tools from Artinsoft/Microsoft/WhoEver and then refactor again to get a nice and clean .NET application.
But I see some problems:
There seems to be no way of refactoring the old application. There is no automated testing whatsoever, not even a formal method for manual testing. Every little change requires manual testing by experienced users who just know where defects might hide.
on the other hand I have established a process and set of tools for testing of our .NET applications which gives us a solid base for making refactorings
Converting that code to .NET without major refactoring feels like: Garbage in, garbage out. Even though I hate calling the old application garbage because somehow it works and has proven itself useful.
Our management has a habit of explicitly demanding quick and dirty solutions, disregarding the effects it has on the productivity and against all recommendations from the development team which has at some point started to deny the existence of quick and dirty solutions in order to be able to do things right. That does not mean that we polish features, but we do include the time to write tests and do refactoring in our estimates. So knowing this, I suspect that once the code is converted to .NET and fixed to the point where the application starts and seems to work, the refactoring-phase will be canceled and the application will be shipped to some customers.
So. What I think is that, despite the fact that rewriting from scratch will take a lot of time and resources, it might still be our only option.
Am I missing an option? Do you see possibilities of not having to rewrite that application?

I suggest that you take a step back and read this paper by Brian Foote & Joseph Yoder (University of Illinois). It provides some architectural insight into the problem you have and options to solve it. It's titled 'Big Ball of Mud' (please don't laugh, it is a serious paper). Here is the abstract:
While much attention has been focused on high-level software
architectural patterns, what is, in effect, the de-facto standard
software architecture is seldom discussed. This paper examines the
most frequently deployed architecture: the BIG BALL OF MUD. A BIG BALL
OF MUD is a casually, even haphazardly, structured system. Its
organization, if one can call it that, is dictated more by expediency
than design. Yet, its enduring popularity cannot merely be indicative
of a general disregard for architecture.
These patterns explore the forces that encourage the emergence of a
BIG BALL OF MUD, and the undeniable effectiveness of this approach to
software architecture. In order to become so popular, it must be doing
something right. If more high-minded architectural approaches are to
compete, we must understand what the forces that lead to a BIG BALL OF
MUD are, and examine alternative ways to resolve them.
A number of additional patterns emerge out of the BIG BALL OF MUD. We
discuss them in turn. Two principal questions underlie these patterns:
Why are so many existing systems architecturally undistinguished, and
what can we do to improve them?
BTW, I think your best option is to use the current application as your Requirements and rewrite everything in VB.NET or C# using a proper design.

There are four main options when you have an application like this:
Do nothing: this is always an option, as everybody knows, if it ain't broke don't fix it. However this might not be an option for several reasons such as needing to comply with some security requirements at the company, or simply because one of the components doesn't work in new platforms.
Rewrite: This would be the dream, right? being able to get rid of all the bad practices and duplicated code and so on? Well, it might be that way, however you have to think all the risks involved in developing a new application from scratch. Do you have all the formal requirements? what about test cases? do your team know every little detail in the code or would you need to go line by line trying to figure out how why that if is there? Also, how many bugs do
Buy something off-the-shelf: Since you are an ISV this won't be an option.
Migrate: Of course you'll be bound by the programming practices you used for the original development but you'll get to a new platform faster, all your business logic will be automatically migrated, you can actually hire developers for the new platform and you can get rid of the legacy elements. From here you can also take advantage of all the tools available to refactor code, continuous integration, unit testing, etc.
Also, with an automatic migration you can actually go further than just WinForms. There are also tools that can take your C# code all the way to the web using a modern architecture.
Of course, I work for Mobilize.Net (previously Artinsoft) and this is my biased perspective.
We've been working on this for around 15 years and have seen dozens of clients who come to us after trying to re-write their application and fail after months or even years of struggling without being able to deliver a working application.

Continuous Integration for stack with Visual C++ and C#

Please recommend a good continuous integration that would build and integrate with the .net stack and the visual c++ as well.
Some recommendations I have got are
Jenkins
CruiseControl
Teamcity
Because of the polyglot nature of the project, which continuous integration solution would you recommend?

I have used all three over several years. Some of the answers below state that most of the work will be producing your own build scripts. This has been true in my experience as well. We use a combination of MSBuild and Powershell scripts for our build process, which can be run under just about any CI tool, so picking one comes down to what you're looking for in terms of customization, integration with other systems, performance, and ease of use.
Short answer:
I recommend Jenkins. So far it seems to be the best combination of the above qualities. It has a ton of plugins, some localization and is actively developed by the OSS community.
Long answer:
I started with Cruise Control .Net. It was easily configurable with a text file and I found it highly reliable. However, we moved away from it because Thoughtworks was moving toward a paid product (Cruise, now Go) and future development was in question. A new team has since forked the project but there is little word about future development since.
We moved to TeamCity, which is free and has a great ajax-y UI. It is easy to setup and get going and has a lot of features for distributed builds. We quit using TeamCity for several reasons. The server does a ton of stuff and it was a bit overkill for our basic needs. Even so, it was not very customizable (see Time Zones and notification contents) and we often found the administration UI confusing. That was all still okay, but we also had steadily worsening performance problems. We started with a standard HSSQLDB out-of-the-box, moved our installation to SQL server when we started experiencing degraded performance, then had to quit using the server at all as performance continued to degrade over time. I'm not sure what the culprit was but I couldn't find any cleanup to do that would explain the constantly worsening performance as the Tomcat web server fought with SQL Server for resources, even when there were no active builds running. I am sure it's my fault and I was missing some crucial setting or needed to feed the server more memory, but this is a shared utility box, we did not have these issues with CC.Net, and most of all, I am not a Java/Tomcat guy and don't have a lot of extra time to keep fighting with these issues.
We've moved to Jenkins now. It seems to be working fine so far but we've only been with it a short while. It was easy to set up, does not seem to be taking nearly as many resources as TeamCity and has a ridiculous number of plugins. The only downside so far is like many OSS products, it does not seem to have the best documentation and it does so much that I may be tweaking knobs for a while to get it set up the way we want.

Between CruiseControl and TeamCity, TeamCity is faster and easier to set up, but you may need to check on licensing for it. I can't speak to Jenkins, never having used it.

Jenkins has the big advantage of being very extensible (currently over 400 plugins), which allows you to combine it with a huge number of other tools. So it gives you complete freedom in your other tool choices. I recently read that this is one problem of TeamCity, that you get locked in using the whole stack of tools (e.g. using SVN or Git as version control system will not be possible).
I am using Jenkins myself for our projects which has both Java and C++ code, and I am very happy with the tool. We had CruiseControl before, and have not once regretted the switch.

I have tried both Cruise Control and Jenkins, and Jenkins impressed me with very fast and user-friendly set up.
The three you list are all sensible choices, and the main problem will be producing the build script(s) needed to do produce the build artifact(s). If you manage to make them do everything needed, changing CI system shouldn't be a big issue.

After implementing all three in different shops, I'd chose all of the above. Pick one.

Planning a programming project by example (C# or C++)

I am in the last year of undergraduate degree and i am stumped by the lack of example in c++ and c# large project in my university. All the mini project and assignment are based on text based database, which is so inefficient, and console display and command, which is frustrating.
I want to develop a complete prototype of corporate software which deals in Inventory, Sales, Marketing, etc. Everything you would usually find in SAP. I am grateful if any of you could direct me to a books or article or sample program.
Some of the question are :
How to plan for this kind of programming? should i use the concept of 1 object(such as inventory) have its own process and program and have an integrator sit for all the program, or should i integrate it in 1 big program?
How to build and address a database? i have little bit knowledge in database and i know SQL but i never address database in a program before. Database are table, and how do you suppose to represent a table in a OOP way?
For development type, which is better PHP and C++ or C# and ASP.NET? I am planning to use Web Interface to set form and information, but using a background program to handle the compute. .NET is very much integrated and coding should be much faster, but i really wonder about performance if compared to PHP and C++ package
thank you for the info

This may not answer your question directly, but I thought this might help you get started in some way. So here it goes: I would say, "think through the process". This means, think through the software development process:
Gather requirements
Identify and define the problem.
Get as much information/facts as you can. (turn on green light, think about everything that you want to go into your software)
Come up with a baseline (turn on red lights, what you really want? the minimum functionality your software "must have" - cant live without)
Analyze
Know what you don’t know, what are the missing facts?
Evaluate your information or lack of it/reliability of information source.
Infer facts that you don’t know.
Form an assumption, opinion, or possible solutions.
Consider alternatives and implications of each solution.
Form an action plan.
Identify technology pros/cons.
Decide technology
Comeup with a functional specs.
Research
Dig into stuff that you would want to know (Best database, ORM, design practices, code samples - gather everything, read about inventory systems that are already there)
Design
Develop
Test
Fix
Prepare deployment plan
Release the product
Gather user feedback
Analyze user feedback
Plan for items in next release.
Repeat steps
And Enjoy!

Before I start this is a shallow answer to a deep question.
1) It looks like you have a reasonable grasp of the major components of your target application. As a .net developer I'd build assemblies that matched broad areas of functionality (not sure what the equivalent is in PHP) and then you can use those assemblies together as a single large app, or seperately as required. It's unlikely you'll get it right first time, so build it how it feels right, and then do some ruthless refactoring to make it better once you've got a handled on the problem.
2) This whole area is covered by Object Relational Mapping - ORM, NHibernate is the best of the bunch in the .Net world. BTW if you learn that you'll be way ahead of the game come graduation/work time. Raw sql is so last decade. I guess you know that Sql Server Express is a free download?
3) For development go with the languages/environment you feel most comfortable in. My preference is .net, and the integrated coding is much faster. Performance is definitely good enough, especially as this is learning project - SO runs on .Net and that supports a gazillion users pretty well.
Enjoy

I don't have any good recommendations for SAP-like projects in particular, but in general the best examples to use for things like this are well-established open-source projects. Anything else is going to be a "toy" example in one way or the other, and will be simplified and cleaned up. It's the "cleaned up" that makes it most unrealistic -- one of the really key things that makes real-world large software projects different from university examples is that the real world is messy, and real-world requirements are messy, and collaboration between lots of people with not quite the same priorities is messy, and real-world software projects have to adapt to and thrive in this messiness.
In answer to your specific questions, though:
1.) Do things in a modular way. This means you have something you can test and work with as soon as you get the first module done. That's especially important when you're learning, because (a) you probably won't have time to actually finish the whole thing, (b) you'll learn a lot from writing the first bit that you'll want to apply in future bits and then you'll probably want to rewrite the first bit, and (c) you'll learn even more from using the first bit.
2.) There are many views on this, and many online articles and books. I can't answer that in an answer here (except to note that in some cases trying to represent it in an OOP way is the wrong programming paradigm -- be careful about overconstraining the answer by the question you ask!); the right answer is to find things to read and spend some days reading them.
3.) You do not care about that sort of performance issue here. Successful programs have been written in both forms. You care about what will teach you the most, and what you are comfortable working with. Either one should be fine. You'll probably find more open-source pieces to look at with PHP and C++.

Your question pretty much covers the whole gamut of planning for a project; a whole thesis might be written (+:
Keep in mind what your team and your teaching-staff want out of the project.
1) Modular is my choice. It'll force you to address the application one module at a time and keep you focussed, but that is subject to
The familiarity of your team with the preferred/recommended language for this project.
Time in hand
Remember that modular means you will necessarily have to provide for module integration too.
2) C++ or C# ? Whichever offers the more learning experience. My own experience with both mentioned technologies is limited, but I remember there used to be a Database Template Library (DTL). C# on the other hand will probably be faster to develop. I could be wrong. There are any number of free DBMS engines available on the net. Unless the assignment explicitly recommends using a text, opt for one of these.
3) I concur w/Brooks up there ^^^
Good Luck!

You are a university undergraduate. And you are talking about complete inventory system.
I suggest building a blog application first with all the best practices (like blogengine), then move to e-commerece sites (nopcommerce, dotcommerce). And then do whatever you like.
This is a common problem with undergrads like you, of jumping way higher without building any simple projects first.

As a full time PHP developer, PHP sucks! ASP.net is okay (mmm... no it sucks too), but it locks you into proprietary licenses.
If you're starting from scratch, go for node.js. It's c++ and server-side javascript. Yes, it's new, but it has engineering promise. It'll be more commonplace in a few years.
And if you're worried about performance, don't. Javascript in V8 is extremely fast.
http://shootout.alioth.debian.org/u32/which-programming-languages-are-fastest.php
Here are some node.js links to get you started:
http://www.delicious.com/homer6/nodejs
Enjoy.

Starting wxWidgets C++ need a gentle nudge

So I've been learning C# for like a year now (I'm 20 years old) and I'm getting pretty confident with it. I've also been meddling with C++ every now and again. For example just recently I've been following the Nehe OpenGL tutorials for C++ and I find it a fun way of learning.
I want to start looking at creating cross platform GUI software after I stumbled across this library called FLTK (fluid something rather). After finally and painfully getting it to work I found it refreshing to know that there are solutions for GUI creation in C++, however I think FLTK looked pretty old.
So I googled around for some newer GUI frameworks and decided to start playing with wxWidgets (decided against Qt because of licensing).
I downloaded it, compiled it, and looked to see if there were any IDE plug-ins for RAD development, as you can imagine going from drag and drop a component onto a form in C# I was hoping for something similar.
I learned that code::blocks has something of the sort so I tried that out. It was alright but the thing that turned me off was the horrible code completion; it would only show members and methods in the current object and nothing for the #included header files. I understand that code completion/IntelliSense isn't easy for C++ but Visual Studio 2008 handles it pretty good. I did find some other RAD tools like wxFormBuilder but it costs money, not something I want to do for simply learning.
So my TLDR question is if anyone has had experience with wxWidgets? Do you just develop in whatever IDE you're comfortable with and just code the GUI? Meaning no visual helpers? Perhaps you could give me a nudge in what direction I should be going :)
Thanks, this is also my first post on this site albeit I have read many threads before that have helped me solve copious problems. Cheers!

My suggestion is to learn how to do GUI layout with wxWidgets in code, then when you get good at it learn how to use the GUI tools.
Doing this kind of work manually for a while gives you the understanding about what you need ("Ok, I need a wxSizer, vertical, to put these two horizontal wxSizers into, where I put my a wxStaticText and a wxTextCtl for each line ...")... where as I think if you started out with the GUI tools you'd just tend to get annoyed because (last time I looked) none of them were Drag And Drop editors like you get with .NET.

Definitely give Code::Blocks another try. It is a WONDERFUL environment to work with wxWidgets in. It comes with a form designer and templates for wxWidgets projects, so I can't imagine working without it.
Also, for a good beginner's introduction to wxWidgets, try this page. It helped me alot when I started with it.

I use wxWidgets without using a drag and drop designer. There are obviously drawbacks to that approach but an advantage is that you don't have any horrible automatically generated code to deal with. In the past I've found having such code in the middle of my project has caused various types of grief - especially if you decide it needs some serious changes it is often necessary to start from scratch because form designers are so much better at "writing" as opposed to "editing".
For simple applications you can assemble a UI using a mixture of customized and standard widgets without too much difficulty. Check out the wxWidgets samples and demos, which use that approach. My chess app (see my website) uses this approach too.
A more elaborate UI could possibly be implemented by writing something special at a higher level of abstraction. The kind of thing I am thinking of would be a subsystem that accepts high level flexible requests to provide different types of user interface functionality, and then sorts out the details of the controls to create and the positioning etc. of those controls itself. That is kind of a technical fantasy of mine but I haven't ever really attempted to do it. Actually, a fairly primitive facility of this type is already available and used by the demos I mentioned; It is called a "Sizer" (class wxSizer) and is basically a control container.
Finally I would point out that problems with the Qt licence have basically gone away since Nokia bought Trolltech and made it "more" free (LGPL licence). Many people think Qt is the way to go these days. I am pretty happy with wxWidgets but will definitely evaluate Qyt seriously one day. Good luck with your projects.

I have developed a number of application GUIs using wxWidgets - you can see screenshots at http://ravenspoint.com/
IMHO, you should distinguish between designing a GUI and implementing it. Expecting to use the same program to design your GUI and to automatically emit all the code to implement the design is expecting too much.
A design program should be fast and simple. I recommend http://www.balsamiq.com/
Once you have settled on the design, then you can turn to the coding. Personally, I find that placing widgets on a panel using the wxPosition and wxSize parameters of the wxWidget constructors to be trivial - easier than trying to nudge widgets into their exact positions using the mouse.
The trouble with using some kind of form builder to emit all the code for a significant GUI is that you end up with a morass of automatically generated code that is hard to find your way around in. If the code has been handwritten, then you already know where everything is located.

wxFormBuilder but it costs money
wxFormBuilder is Free and Open source
Also Visual Studio is best for C++ coding in wxWidgets, but you will run into lots of issues on Linux due to the way Visual Studio handles a lot of things, like file names you have to take extra care for case sensitivity.
Also regarding the GUI, add the wxFormBuilder to the Visual Studio solution, add the file type handler to Visual Studio. Now just double click on it and add your dialogs, like you do it in C#.
Just add the generated code files to the Visual Studio project once and you are done.
It's very easy to use Visual Studio with GUI programming in wxWidgets.
Also don't forget to watch the CodeLite editor if you want to port on Linux. It is very much compatible with the Visual Studio projects and shortcut keys.

There is wxGlade and wxFormBuilder.
I don't know if there is another one, but the wxFormBuilder I know is free and open source.
You will find links to other tools on the homepage
of wxGlade (http:// wxglade.sourceforge.net/).
Most of these programs are able to output xrc files, which are are an XML representation of your GUI. It address the issue of complexity of the generated code but is not as powerful.
Yet personally I don't use any of these tools, except sometimes to see what it's going to look like before I do it but never to actually generate code.

I just finished doing a hello world program in both wx and fltk using the netbeans IDE. Microsoft Visual Studio WinForms or MFC was much easier than either. I needed cross platform capability so I could not use it for this application. I finished fltk in much less time than wx. The program was simpler, smaller, faster, and easier to write. Wx had pretty disjointed documentation and organization. Fltk was designed back in the "old days" when small fast and efficient was valued. That's probably why it seems "old" to you. I'm an old guy so it worked for me. YMMV

Should you obfuscate a commercial .Net application?

I was thinking about obfuscating a commercial .Net application. But is it really worth the effort to select, buy and use such a tool? Are the obfuscated binaries really safe from reverse engineering?

You may not have to buy a tool - Visual Studio.NET comes with a community version of Dotfuscator. Other free obfuscation tools are listed here, and they may meet your needs.
It's possible that the obfuscated binaries aren't safe from reverse engineering, just like it's possible that your bike lock might be breakable/pickable. However, it's often the case that a small inconvenience is enough to deter would be code/bicycle thieves.
Also, if ever it comes time to assert your rights to a piece of code in court, having been seen to make an effort to protect it (by obfuscating it) may give you extra points. :-)
You do have to consider the downsides, though - it can be more difficult to use reflection with obfuscated code, and if you're using something like log4net to generate parts of log lines based on the name of the class involved, these messages can become much more difficult to interpret.

Remember that obfuscation is only a barrier to the casual examiner of your code. If someone is serious about figuring out what you wrote, you will have a very hard time stopping them.
If you have secrets in your code (like passwords), you're doing it wrong.
If you worried someone might produce your own software with your ideas, you'll have more luck in the marketplace by providing new versions that your customers want, with technical support, and by being a partner to them. Good business wins.

At our company we evaluated several different obfuscation technologies, but they all had problems. The biggest problem was that we rely a lot on reflection, e.g. to dynamically create grids based upon property names.
So all of the obfuscators rename things, you can disable it of course, but then you lose a lot of the benefit of obfuscation.
Also, in our code we have a lot of NUnit tests which rely on a lot more of the methods and properties being public, this prevented some of the obfuscators from being able to obfuscate those classes.
In the end we settled on a product called .NET Reactor
It works very well, and we don't have any of the problems associated with the other products.
"In contrast to obfuscators .NET Reactor completely stops any decompiling by mixing any pure .NET assembly (written in C#, VB.NET, Delphi.NET, J#, MSIL...) with native machine code. In detail, .NET Reactor builds a native wall between potential hackers and your .NET code. The result is a standard Windows based, not MSIL compatible, file. The original .NET code remains intact, well protected by native code and invisible for prying eyes. The original .NET code is not copied on harddisk at any time. There is no tool which is able to decompile .NET Reactor protected assemblies."

The fact that you actually can reverse engineer it does not make obfuscation useless. It does raise the bar significantly.
An unobfuscated .NET assembly will show you all the source, highlighted and all just by downloading the .NET Reflector. Add obfuscation to that and you'll reduce very significatively the amount of people who'll be able to modify the code.
It depends on you are you protecting yourself from. If you'll ship it unobfuscated, you might as well open source the application and benefit from marketing. Shipping it obfuscated will only allow people to relatively easily generate modified binaries through patches instead of being able to steal your code and create a direct competitor. Getting the actual source from obfuscated code is very hard, depending on the obfuscator, of course.

I think that it depends on the type of your product. If it is directed to be used by developers - obfuscation will hurt your customers. We've been using the ArcGIS products at work, and all the DLLs are obfuscated. It's making our job a lot harder, since we can't use Reflector to decipher weird behaviors. And we're buying customers who paid thousands of dollars for the product.
So please, don't obfuscate unless you really have to.

Things you should take into account:
Obfuscation does not protect your code or logic. It just makes it harder to read and understand.
Obfuscation does no one stop from reverse engineering. It just slows the process down.
Your intellectual property is protected by law in most countries. So if an competitor uses your code or specific implementation, you can sue him.
The one and only problem obfuscation can solve is that someone creates a 1:1 (or close to 1:1) copy of your specific implementation.
Also in an ideal world reverse engineering of an obfuscated application is economical unattractive.
But back to reality:
There exists no tool on this planet that stops someone from copying user interfaces, behaviors or results any application provide or produce. Obfuscation is in this situations 100% useless
The best obfuscator on the market cannot stop one from using some kind of disassembler or hex editor and for some geeks this is pretty good to look into the heart of an application. It's just harder than on an unobfuscated code.
So the reality is that you can make it harder and more time consuming to look into your application but you won't really get any reliable protection. Regardless if you use a free or an commercial product.
Advanced technologies like control flow obfuscation or code virtualization may help to make understanding of logic sometimes really hard but they can also cause a lot of funny and hard to debug or solve problems. So they are sometimes more like an additional problem than a solution.
From my point of view obfuscation is not worth the money some companies charge for their products. If you want to nag casual developers, open source obfuscators are good enough. If you want to make it as hard as possible to look into the heart of your applications, you need to use cryptographic containers with virtual execution environments and virtual filesystems but they also provide attack vectors and may also be a source for a bag full of problems.
Your intellectual property and your products are in most countries protected by law. So if there's one competitor analyzing and copying your code, you can sue him. If a bad guy or and hacker or cracker takes your application you are pranked - but an obfuscator does not make a difference.
So you should first think about your targets, your market and what you want to achieve with an obfuscator. As you can read here (and at other places) obfuscation does not really solve the problem of reverse engineering. It only makes it harder and more time consuming. But if this is what you want, you may have a look to open source obfuscators like e.g. sharpObfuscator or obfuscar which may be good enough to nag casual coders (a List can be found here: List of .NET Obfuscators on Wikipedia).
If it is possible in your scenario you might also be interested in SaaS-Concepts. This means that you provide access to your software but not the software itself. So the customer normally has no access to your assemblies. But depending on service level, security and user base it can be expensive, complex and difficult to realize a reliable, confident and performant SaaS-Service.

No, obfuscation has been proven that it does not prevent someone from being able to decipher the compiled code. It makes it more difficult to do so but not impossible.

I am very confortable reading x86 assembly code, what about people that is working with assembly for more than 20 years ?
You will always find someone that only need a minute to see what your c# or c code is doing...

Just a note to anyone else reading this years later - I just skimmed through the Dotfuscator Community Edition (that comes with VS2008) license a few hours ago, and I believe that you cannot use this version to distribute a commercial product, or to obfuscate code from a project that involves any developers other than yourself. So for commercial app developers, it's really just a trial version.

...snip...
these messages can become much more
difficult to interpret
Yes, but the free community edition that comes with Visual Studio has a map functionality.
With that you can back track the obfuscated method names to the original names.

I've had success putting the output from one free obfuscator into a different obfuscator. In Dotfuscator CE, only some of the obfuscation tricks are included, so using a second obfuscator that has different tricks makes it more obfuscated.

It's quite simple to reverse engineer a .net app using .net reflector - since the app will generate VB, VC and C# code straight from the MSIL, and it's possible to pull out all kinds of useful gems.
Code obfuscators hide code quite well from most reverse engineering hacks, and would be a good idea to use on proprietary and competitive code that adds value to your app.
There's a pretty good article on obfuscation and it's workings here

This post and the surrounding question have some discussion which might be of value. It isn't a yes-or-no issue.

Yes you definitely should. Not to protect it from a determined person, but to get some profit and have customers. By the way, if you reach a point here someone tries to crack your software, that means you sell a popular software.
The problem is what tool to choose for the job. Check out my experience with commercial obfuscators: https://stackoverflow.com/questions/337134/what-is-the-best-net-obfuscator-on-the-market/2356575#2356575

Yes, we do. We use BitHelmet obfuscator. It's new, but it works really well.

But is it really worth the effort to select, buy and use such a tool?
I found Eazfuscator cheap (free), and easy to use: took about a day.
I already had extensive automated tests (good coverage), so I reckon I could find any bugs that are/were introduced by obfuscation.