Accidental Technologist

Try out IronRuby, Interactively in your Browser

Rob Bazinet — Wed, 14 May 2008 12:47:00 GMT

Curious about Microsoft's IronRuby implementation? Try it out in your browser.

Similar Posts

Twitter Broadens my Network and Makes Me More Social

Rob Bazinet — Wed, 14 May 2008 01:10:00 GMT

It is pretty easy when you work independently, and spend a lot of time in a home office as I do, to feel isolated from the outside world.

I have been using Twitter for the past 6 months or so and have found it to be very rewarding and an integral part of my day-to-day networking. I blogged a few months ago about finally "getting" Twitter and its value to me has only grown since that time.

I use Twitter to:

Provide a quick update to readers - Sometimes there is not time for a blog post or the content is just too short for a post, so shouting out a quick update to what I am working on at this moment.
Ask a question to my followers - Who mostly have similar interests and backgrounds as I do, to get a piece of information which is usually returned very fast.
Find leads on articles to write about for my work on InfoQ - I follow around 300 individuals who provide a constant stream of great information. I get news faster on Twitter than any other way.
Discover new things related to my profession - Tips or blog links to ways to do things better is always part of the daily stream, really just a free education.
Network and talk to those I may not often get to see face-to-face - Conferences are a great way to meet and talk with peers, getting to one isn't always easy but the ability to interact with those same people is great way.

I find Twitter to be most valuable to me as a way to build my network of like-minded folks.

I have benefited from Twitter in the contacts I have made by:

I found Ninject - a great dependency injection container for .NET
I became part of The Lounge network and now a member of an open source project, Graffiti-Redirect.
Met the organizer of a new Rails conference, acts_as_conference, attended it and met many people I knew only online.
And so many others.....

If it wasn't for Twitter I would not likely have come across any of these folks and not gained by what they have to offer.

I think everyone who spends time on the Internet using services like Facebook, MySpace and others should sign up for a free Twitter account and explore what the chatter is all about. It is very easy look up people you admire or you may have interest in, follow them and see who they follow. A network of valuable contacts in only a few follows away.

If you don't already follow me on Twitter, please do so now. I will return the favor.

Technorati Tags: Twitter,Socializing,Networking

Posted to Technology

IronRuby, Mono and My Mac

Rob Bazinet — Mon, 12 May 2008 16:00:00 GMT

I have been developing Ruby applications for the past couple years, not as my only work but part of my overall work. The IronRuby project is a large part of my work as well, not as a committer, but from a consumer standpoint. I am looking forward to the day where we can run Ruby on Rails on IronRuby and therefore IIS. I also use IronRuby for the book I am working on with a friend for Wrox Press, which I talked about previously.

I have always been a Windows developer and use Windows systems for all of my daily work. This change in the past year when I jumped on the Apple Mac bandwagon and started using a MacBook Pro. I have to say I really love it from a hardware perspective as well as from the perspective of the usability of OSX, it is just a pleasure.

This leads me to the reason for this post, getting IronRuby to build and run on a Mac under Mono.

Mono

The first part of getting IronRuby to run on the Mac is to install Mono. Mono is needed to have a .NET 2.0 available on the Mac, its current version at the time of this writing is 1.9.1. The download for the Mac is a Universal DMG Image extracting to a PKG which you just double-click and install, pretty straight forward.

I followed much of Seo Sanghyeon's brief tutorial on the details of IronRuby on Mono but needed to make some adjustments to actually get it to work on my system. For example, I did NOT install Mono from source, I used the pre-built package.

IronRuby Source

The IronRuby team is constantly make changes to the source and not all revisions are compatible with this built process. I am sure as we get closer to release there will be more stable base of code. For right now we are on Revision 100, so the Subversion command to get the source looks like this:

svn co -r 100 http://ironruby.rubyforge.org/svn/trunk ironruby

IronRuby Build Process

The build process is where it gets a little tricky to make sure everything is lined up. The first thing is changing to the directory where IronRuby source is located:

cd ironruby

A patch needs to be applied for Rake to work properly. The patch can be found on Seo's web site here, notice it is build revision specific and new ones will likely be available. Next up run Patch:

patch -p0 < patch-mono-r100

The next step was the key to my success, telling the Mono compiler where the needed libraries are located. The build would never work without it.

export PKG_CONFIG_PATH="/Library/Frameworks/Mono.framework/Libraries/pkgconfig/"

This export I put in my ~/.bash_profile file and closed and reopened my Terminial window. Next, run Rake:

rake compile mono=1

If all goes well the build will succeed, with likely some warnings, and you can run IronRuby:

mono build/mono_debug/ir.exe

You should see a nice prompt showing a happy IronRuby interpreter running.

Conclusion

The process to get IronRuby on the Mac under Mono is pretty straightforward and will likely get easier. The key for me getting this to work was how great the community is around IronRuby. The members of the IronRuby-Core mailing list do such a great job.

Technorati Tags: IronRuby,Mono,Mac

Posted to IronRuby

ASP.NET MVC Recent Code Drop Updates Cause Some Pain

Rob Bazinet — Tue, 06 May 2008 10:49:00 GMT

Microsoft released the ASP.NET MVC Framework updates to CTP 2 in the middle of April and for those using it you may have noticed some changes to the way certain things are done.

I am working on a project where I am using the CTP to get a web site built and learning ASP.NET MVC along the way. I started using the CTP 2 for the project and moved over to the latest code drop as soon as it came out. Sure, I expected some changes so it was not a big deal, but the ones I encountered were interesting and caused a little head scratching about something that worked before but not now.

Project Wizard

The CodePlex April 16 MVC Interim Source Code Release includes two downloads, the MVC code itself and a set of Visual Studio MVC Templates. When I first saw the templates I figured they were just some additional ones but they are above and beyond those in the original CTP2. You can see there are two templates now when creating a new project.

One is shown under Visual Studio installed templates, called ASP.NET MVC Web Application and the other one and newer is under My Templates and is also called ASP.NET MVC Web Application. Choosing one over the other creates two very different sets of code.

One benefit of having two sets of templates is you can create applications for both sets of code bases, CTP2 and the interim drop. One other thing that is different when using the new templates is you don't get a nice test project to go in your solution. It can be added of course.

Controllers

One of the changes which threw me off right away is the way Controller action methods have changed. You can see from the screen shot just below that the controller action method Index() returns a void.

But the controller from the interim code drop controller action now returns an ActionResult, not a void. This obviously breaks existing code. Changing my existing code to work was not a big deal but was a surprise when trying the first build with the new code drop.

After fixing this code and heading over to Scott Guthrie's blog, he explains the reasoning behind the changes:

The MVC feature team is experimenting with a few ideas in this week's drop and are trying out some new ideas:

Action methods on Controllers now by default return an "ActionResult" object (instead of void). This ActionResult object indicates the result from an action (a view to render, a URL to redirect to, another action/route to execute, etc).

The RenderView(), RedirectToAction(), and Redirect() helper methods on the Controller base class now return typed ActionResult objects (which you can further manipulate or return back from action methods).

The RenderView() helper method can now be called without having to explicitly pass in the name of the view template to render. When you omit the template name the RenderView() method will by default use the name of the action method as the name of the view template to render. So calling "RenderView()" with no parameters inside the "About()" action method is now the same as explicitly writing "RenderView('About')".

So why change Controller action methods to return ActionResult objects by default instead of returning void? A number of other popular Web-MVC frameworks use the return object approach (including Django, Tapestry and others), and we found for ASP.NET MVC that it brought a few nice benefits.

Please take a look at Scott's post on this to see all of the benefits which I won't go into here.

Routing and Default.aspx

In the CTP2 release there is a default.aspx page that was really just a placeholder page that had no code-behind, but building with the new bits causes the default route to not behave the same way and produce a page not found error. In order to fix this some code behind is added to default.aspx which fixes the problem for CTP2 code. Below is what I added and incidentally is the same code added by the new templates.

The changes are needed because of helper method called routes.MapRoute() which eases the syntax of creating routes. Anyone who created routes with routes.Add will understand how it can be a bit simpler:

The recent code drop (4/16) templates fix all these issues and new projects just work without having to deal with tweaking code.

Conclusion

I think these changes are keeping us on course in the right direction but are a couple of gotchas if you are not aware of them.

Technorati Tags: Microsoft,ASP.NET MVC

Posted to ASP.NET MVC

Upgrading My Development System to Quad Core

Rob Bazinet — Tue, 06 May 2008 02:48:00 GMT

OK, well the title pretty much tells it all, I set out to upgrade my 1 year-old system's processor from an Intel E6600 (Core 2 Duo) to a new Intel Q9300, quad core 45nm processor. The task is easier said then done.

I ordered the Q9300 from my favorite component reseller, Newegg, and had it my hands in a couple days. The process fits into the same 775 socket as my previous chip but my motherboard would need a BIOS update to support the new chip, which I did. My motherboard is a Asus P5W-DH Deluxe which is capable of processor upgrades as long as the CPU is a 775 socket.

The motherboard is probably the most sophisticated and feature-packed motherboards I have ever owned. It certainly has the most slots and connectors on it. Anyway, I unpacked the CPU, removed the old one and installed the new one with ease. I had thought to myself, how can it be this easy? Well, I spoke too soon. Right after turning on the system I got the the post screen and right after received an error message stating the CPUID was not recognized. After much searching around the Asus support site and several other forums such as Anandtech, I was told the 975 chipset does not support the new 45nm chips and I would probably be the first person on the planet to have it working if I could pull it off.

I am not the type to play with hardware for all hours trying to prove I can do something. I ended up talking to a few people who have the same CPU I have and had it running successfully using motherboards from Abit and in particular the Abit IP35 Pro, which has great support for the 45nm chips. I was lucky to find a great deal from Newegg.

Once I found out I needed a different motherboard to support my new CPU, I decided to upgrade my RAM from 4G to 8G, order a new video card and install Vista x64 to take advantage of the full 8G of RAM. The prices of RAM have come down so much that I paid half of what the original 4G cost.

So the order consisted of:

The video card is real incredible with a huge set of features:

Chipset Manufacturer: NVIDIA
Core clock: 500MHz
Stream Processors: 96
Memory Clock: 1600 MHz (effective)
DirectX: DirectX 10
OpenGL: OpenGL 2.0
DVI: 2
TV-Out: HDTV / S-Video O

The video card will be very capable of off-loading some future video processing, Silverlight work or who knows.

I am expecting to get everything setup and going by the weekend which I will post some follow-up to my experience. I am very anxious to get back up on Vista x64 now that there is iTunes support in iTunes 7.6+.

Technorati Tags: Abit,Intel,Hardware

Posted to Hardware

Similar Posts

Joined the Infozerk Lounge Network

Rob Bazinet — Fri, 02 May 2008 21:18:00 GMT

I don't believe in using my blog as a generator of revenue but as a way to share information in the community. As of May 1 I have joined The Lounge by InfoZerk. I have been a long-time follower of James Avery, reading his blog and seeing him on Twitter. We were introduced when I started using GraffitiCMS as my blogging engine and his Graffiti.Redirect plugin (which I am now a contributor).

I decided to join The Lounge for several reasons:

The network consists of many people I respect in the development community.
The advertisers are top-notch organizations and am proud to have them represented on my blog.
Self-promotion, getting the work out about the products and services I provide.
The payment we receive being part of the network can go toward server hosting fees.

All in all, I think it is definitely a win-win for everyone. You can read the announcement about the new Lounge members on the Infozerk blog.

Technorati Tags: The Lounge,Infozerk

Posted to Technology

Back in the Vista Saddle Again

Rob Bazinet — Tue, 29 Apr 2008 17:09:00 GMT

I have wanted to be a Windows Vista user and developer since before it was released, but there always seemed like something didn't work or some application I used kept crashing. The pure pain of the performance of Vista alone was not worth the "upgrade" from Windows XP.

I have tried both the 32 and 64-bit versions of Vista and had a horrible time, consistently running into problems trying to get the iPhone working with iTunes under Vista as well. I wrote about some issues before giving up.

Once Vista Service Pack1 was released I decided to update a couple of my virtual machines (VMWare, no Virtual PC here) and was a bit disappointed by my observations, the VMs were not noticeably faster. I had heard SP1 was significant performance improvements over Vista RTM, but this was not so.

I had the opportunity to get my hands on a Vista disk with SP1 already part of the install, this way you don't to install Vista and then install the service pack later. I did this first in a fresh virtual machine with 1G of RAM, the install was clean and the VM runs amazingly fast on VMWare Fusion on my MacBook Pro with 4G of RAM.

Since Vista ran so well in a VM on my Mac I decided to go for it and replace a flaky Windows XP installation on my main development system. The installation went very smooth and I did not have to install any third party drivers for my Asus motherboard.

Setting out to install all the necessary software packages like Visual Studio 2008, SQL Server 2005, iTunes and a ton of other packages has resulted in a very fast and stable platform. All of my issues with the iPhone and iTunes as well as the ASP.NET development server are gone. One of my biggest complaints, speed, seems to be a non-issue for me today.

After a few days of pretty heavy use I have to say I am pleased with Vista at this point. It shouldn't matter installing Vista and the service pack later but it certainly seems to be better doing it this way.

If I look back to the days of Windows XP it did take Microsoft a couple service packs for users to adopt XP and be happy with it. There were performance complaints, stability issues and compatibility problem that were all overcome. I think this may be very much the same.

I think Vista is very good at this point and I have to admit it took me a lot to say it. I have been down on it since the release and all of the troubles I have faced. I have given a good deal of razzing to the Microsoft consultants I work with, they get paid to love Vista and poking at them has been a lot of fun.

Technorati Tags: Microsoft,Vista,VMWare

Posted to Microsoft

Simple Subverison Repository Setup with VisualSVN Server

Rob Bazinet — Tue, 29 Apr 2008 12:00:00 GMT

I have been using some form of source control for what seems an eternity. I started out using the copy-to directory method because source control did not really exist at the time. More recently I have used both Microsoft Visual SourceSafe, SourceGear's Vault and Team Foundation Server, all of which I use because of client-specific needs.

I have a few projects I am developing for my own company and choosing source control has really not needed much discussion. I use Subversion for everything because it's free, works really well, has great community support and support a wide-variety of clients on many operating systems. The other products such as Vault and Team Foundation Server use Microsoft SQL Server and that is just one more license and database to maintain.

Since this is Subversion and my clients of choice are pretty open I use TortoiseSVN so I can access the repositories from Windows Explorer. I also use VisualSVN which supports both Visual Studio 2005 and 2008 and integrates into the development environment, letting me checkout and check-in when the same way other source control clients work in Visual Studio.

Much of the time users of TortoiseSVN may be accessing an existing repository for an open source project or something existing for their job. Most developers don't have to setup their own Subversion repository, but when you do....there is a simple way from the creator of VisualSVN. They have a free product called VisualSVN Server with will aid you in setting up and configuring your Subversion repositories.

VisualSVN Server

VisualSVN Server is a package that contains everything you need to install, configure and manage Subversion server for your team on Windows platform. It includes Subversion, Apache and a management console.

Installation

The download is only about 5MB and is a simple MSI setup application. Running the setup you are faced with minimal questions, which can be answered easily. The defaults seem to work pretty well.

The only data needed:

Location - Where is the executable parts of Subversion going to live.
Repositories - Where all items under source control will be stored. This would be a great directory, and all of its subdirectories, to add to your automated backup solution.
Server Port - The two options are 8443 (default) and 443. I just stuck with the default.
Use secure connection checkbox - will create its own certificate if this is selected. I used this and it works very well giving the user a nice secure connection.

Installation takes less than a minute and you are done.

Server Manager

Once installation is complete, running the VisualSVN Server manager presents the user with the screen above. It gives a nice status screen telling us the health of the VisualSVN server, in this case it says running.

We are able to now add Users, Groups and Repositories.

The first thing I did was to add myself as a user, nothing difficult just a user name and password. One of the options here is the ability to add Groups which would help with issues of security. I don't currently have a need for groups but who knows in the future. This could be use for an open source project having a group for contributors and another with read-only access for those just interested in getting the source.

Creating the first repository is simple too, right-click on Repositories in the manager and select Create New Repository..., enter the name you want and leave the default structure checked and that's it. The new repository is created.

Once it's created it is easy to see the default structure we expect in our Subversion repositories with branches, tags and trunk. You could probably have a simpler structure but I think why go away from what has emerged as a standard for the structure of a repository.

Selecting a repository and choosing Security from the menu gives you a chance to setup any type of security here. You could really lock it down or give access to previously created groups. Again, my needs are really simple here so I have myself and will set everyone to No Access for now.

Connecting from TortoiseSVN

A Subversion repository is not very useful with having clients to connect to it. I use two, as I mentioned before, TortoiseSVN and the VisualSVN client for Visual Studio.

Connecting to the new repository with TortoiseSVN gives us a prompt where we put in the user name and password created earlier and we are connected. Granted, we don't have a project in there yet but this shows it's all working.

Brining down the only items in the repository, 3 empty directories.

Connecting in Visual Studio with VisualSVN Client

I am not here to show anyone how to use source control from within Visual Studio. I just want to demonstrate some coolness with setting up something so important as source control and have it all just work. Connecting from Visual Studio to our existing repository is easy and as can be seen, it just works. Plugging the URL to the repository, either local or remote and having a local place to put the files and you are done.

Finally

I looked at Git before settling on Subversion. Git seems to be the hot new kid on the block but I am not sure of its value to a project if it is not open source. Git seems to have distributed power but not much better in any other way.

Scott Hanselman had a good podcast about Git, called Distributed Source Control with Git.

I think Subversion is a step above CVS from the open source side of things and different than SourceGear Vault and Team Foundation just because I didn't have to deal with setting up and maintaining another SQL Server.

Technorati Tags: VisualSVN,Subversion,Source Control

Posted to Programming

CartoonMe and Twitter

Rob Bazinet — Mon, 28 Apr 2008 10:40:00 GMT

Have you ever wondered how some of your Twitter friends have those cool cartoon versions of themselves? I was put onto one such service from one my Twitter friends.

The source I found recently was from a web site called CartoonMe, and for $5.69 you can have one too.

Here is a pic of my mug before becoming a cartoon:

And here is me after CartoonMe got done with it:

I think I look better as a cartoon than I do in real life but now have a cool Twitter icon for a while.

Technorati Tags: Twitter,Cartoonme

Posted to General

Similar Posts

Finding Your Way Around Web APIs

Rob Bazinet — Thu, 24 Apr 2008 12:00:00 GMT

Any developer who is tasked with the job of consuming the services from some of the popular web sites has probably gone a little mad trying to make it all happen. Each web site, whether it be Google, Digg, Flickr, Yahoo or Microsoft, has their own set of Application Programming Interfaces (APIs) which allow developers to interact with their services. The raw data pushed back and forth to each service may simply be XML in the end, but knowing what to expect from the service may be a few sessions of trial and error when implementing.

It is the developer's good fortune to have a couple web sites to help us sift through the gauntlet of API specifications.

APIFinder

The first one I came across was APIFinder where I introduced it in a news item for InfoQ. I wanted to talk a bit about it here even though I covered some of the details there.

The site is pretty intuitive and allows users to search for the web service they are interested in. A search for Amazon S3 led me to this information about Amazon S3. There is a pretty good overview presented to the user and clicking the Download the API button takes us right to the Amazon S3 page. I think this is good so we get the latest information and don't have to worry about APIFinder having stale information. The only confusing part is the button says Download the API which leads me to think I am actually going to download something instead of being sent to the source site.

Each page dedicated to the API, like the S3 one above, does give some good information about the service, what it supports and any cost that may be associated with it. There is enough information given to let me make the decision if I want to go to the web site for additional information. It would be a bit helpful here to see some code samples that users added, the article may cover such things but some snippets here would be good.

gotAPI.com

GotAPI.com seems to have a bit of a rougher user interface, even though I was told it was easier to use. UI aside it does offer some nice features over the other API sites, giving users the ability to see a list of detailed methods which are selectable and give a nice big of detail about its use. Below is a screen shot of the level of detail given:

Selecting any web method from the left-hand navigation provides equally good information. Notice too, the information is coming from Flickr itself so we should not see any stale information.

I don't know how long this site has been up but they are lacking detail on some common APIs I would expect to see from Google, Microsoft, Amazon and Yahoo. I hope to see additions here in the future as this is the one-stop-source I like to see when trying to find information on the web.

Finally

Finding developer information about consuming and interacting with public web sites is typically available on the web sites of the companies providing the service, sometimes easy to find but not others. These two sites provide a great service to developers and I am happy to have found them. They will be of use to me in the future and I hope you as well.

Scott Hanselman has a great post in one of his Weekly Source Code post about consuming the web services of some of the more popular web sites. The bottom line is, it's all about XML. Scott's post goes into nice detail about how to work with each.

Technorati Tags: API,APIFinder,GotAPI,Web Services