My eyes just teared up (no, really...I'm still kinda all emotional inside) watching the Nintendo E3 press conference intro video...wow, incredibly moving.

Amazing.

Update: Man, the tears keep coming...I dunno...such a fanboy I guess.  Glorious day!

I'm not going to go into a detailed review with performance specs and what not (they're floating around the web already, no?), but just a quick take.

Purchased the notebook last week from Geared2Play (you can find details in this thread: http://forum.notebookreview.com/showthread.php?t=52014).

Price was right and service from G2P was good (wife said the man on the line (Eddie?) was friendly and knowledgeable).

Notebook was shipped Friday and I received it today in the afternoon with a great deal of anticipation.

First of all, the W5F is gorgeous in person. Incredibly sexy and attention grabbing. Coming from an S1A previosly, the W5F is an increase in weight (S1A was incredibly light, though). However, the weight is put to good use; whereas the S1A would creak under torsional stress, the W5F is solid. This is one of the most solid notebooks I've felt in a long time. No creaking; very rigid chasis.

The screen was a big surprise. The S1A really suffered in this respect as the screen was of very poor quality in terms of brightness and dot pitch. The W5F screen is beautiful. Incredibly vivid and sharp. The pictures came out a bit grainy due to the high ISO, but trust me, the screen is incredibly sharp and the contrast is excellent. I have to admit that the widescreen format makes it seem like the screen real estate is much larger than it actually is.

Haven't tried out many of the features yet as it will be wiped and OS reinstalled (wife needs to use it at work (elementary school) on a domain). The bluetooth worked out of the box with a bit of setup (hint: to get the mouse to connect, press and hold the reset button at the bottom). The mouse was a great addition, Bluetooth no less. Unlike the S1A, W5F does not ship with a carrying case. Not a big deal as we have tons at home

Setting up the network was a trial. It was pretty confusing (see the screenshots below) as I could get a signal and an IP from the router, but I couldn't access the network (no ping response from the router even). I had to fiddle around and reboot the machine to get it to work. The Intel software didn't help much (see screen).

(Some shots are grainy from high ISO)

Waited for UPS all day for this!

A box inside the box...

Yet another box....

Finally, the goods!

Size comparison to V3 Razr.

Not as slim as the S1A, but S1A had a modular CD-R/W drive.

Widescreen goodness.

Orange light is badass.

Keyboard has excellent texture and feedback. The touchpad texture is nice too (although it'll probably wear off after a while I assume).

Screen is very nice; much better than anticipated.

Very nice "soft" LED lights. The touchpad is actually textured (those little gray dots are bumps).

Incredibly vivid; the green shows the contrast much better.

Viewing angle is not bad! Much better than S1A and even better than my Chembook (Compal).

From the other side...

Vertical viewing angle suffers a bit, but acceptable.

Uh....what "Wireless On" option???

Widescreen looks good.

Another shot...

Open from the side.

All in all? Better than expected. I had high hopes (rightfully so for a $1700 notebook), but this notebook (so far) has exceeded them. Beautiful looks, snappy performance (primarily office usage and web browsing), decent weight, and great build quality.

Just started.

Turns out that it requires "HTML Help 2 Compiler" or "VSHIK".

Having used the previous version of NDoc, I first turned to my VS2005 install CD to see if the help utilities were included but not installed by default (but I swear I looked through that thing when I installed it specifically for help utilities).

Not there

Googling led me to this sie: http://www.helpware.net/mshelp2/h20.htm, which has info on how to obtain VSHIK and how to set it up.  The site also has a good primer on MSHelp2.

Will report on how NDoc 2 is coming along...

Update: First run looks good.  The HxS output format is still a bit confusing to me...will have to dig into that.

I was tired of working with tools that used XmlSchemaImporter and XmlCodeExporter (XSD.exe and WSCF) to generate code from XSD files.  I'm sure I'm not the only one, as there are other developers that are awaiting a .Net 2.0 version of the tool as well.

Unfortunately, the execellent XsdObjectGen.exe tool hasn't been updated to take advantage of .Net 2.0 features like generics and partial classes (the most significant change).

After looking at Dingo a bit (already open source and using a templating method as well) and searching for other options, I came to the conclusion that the only sensible thing to do would be to fix the codebase for XsdObjectGen.exe and rewrite the code generation logic.  Dingo was a bit too dense for its own good and there really weren't any other alternatives that I could find.

Took me about 45 minutes to get it up and running with generic lists.  I emailed Dan Rogers to see if it would be acceptable to post this code and/or the compiled binaries and executable (same exact command line parameters).  I know that the generated code refers to an EULA.doc, but I checked the installation directory and found no such document.  So we'll see how that turns out...

So I've been working with Enterprise Library 2.0 (EL2) Logging Application Block recently and I've come across some quirks that are puzzling me.

First, I've been using log4net for most of my logging in the past.  Recently, I've taken a look at NLog due to the fact that log4net is currently under "incubation" and has been inactive for a loooong time.  The developers are still active as shown by the activity in the the mailing lists, but otherwise, the codebase has been kinda sitting there for quite some time (until recently) with no date on when it'll exit incubation.

Anyways, after checking out some performance numbers on EL2 vs. log4net, I was sold.  Easy configuration via the configuration GUI, easy to understand, tons of documentation, and it's first party Microsoft (easy to get team members and managers to buy into it).

So here I am working with it today and setting up my test code to automatically regenerate the database before each run and my application code crashes when the logging fails (exception).  I had mis-typed the path for one of my SQL files and the database wasn't created for the logging block, but still, I don't think that the right thing for EL2 to do is to allow that logging error to bubble up to the application code.  With log4net, if the connection to the log database is broken, the AdoNetAppender will simply fail but not cause the rest of the application code to fail. [Update: can't reproduce it, but I know this is what cause the error since as soon as the database was there, it was happy, but it's running fine now even without a database. Ugh, totally puzzling...]

Weird design choice.  I guess it's useful to know that your logging block is failing.  But what the heck, isn't that why there are multiple listeners so that if one fails, you have a fallback (i.e. log all critical errors to database, event log, and flat file)?

Secondly, as I'm looking at the database scripts for creating the procedures and database tables for logging included with the EL2 source code, I'm puzzled by the design choice.

Take a look at the code for adding a category:

CREATE PROCEDURE [dbo].[AddCategory] -- Add the parameters for the function here @CategoryName nvarchar(64), @LogID int AS BEGIN SET NOCOUNT ON; DECLARE @CatID INT SELECT @CatID = CategoryID FROM Category WHERE CategoryName = @CategoryName IF @CatID IS NULL BEGIN INSERT INTO Category (CategoryName) VALUES(@CategoryName) SELECT @CatID = @@IDENTITY END EXEC InsertCategoryLog @CatID, @LogID RETURN @CatID END

First of all, why are the categories stored in a seperate table?  My guess is that the designers wanted to save some space in the log entry row by taking out the category from the log entry???  I can't seem to come up with another good reason for it since it's not like the categories in the category table are associated with an application identifier (and they must all be unique category names).  Profiler tells me that it requires at least 14 reads to write one entry into the log.

Not only that, the code to execute adding the category and adding the log entry are two seperate calls from the client since the WriteLog procedure doesn't receive category information.  I'm going to go out on a limb and say that the only reason that EL2 logging is able to outperform log4net is due to .Net 2.0 related optimizations.

So I think it's back to log4net for me.  I don't know how the rest of the team will take it, but it seems to be a better choice.

Well, my pineapple died recently.

Yes, that's right, pineapple.  Sometime last year, I decided to start growing a pineapple plant after I picked up a weird interest in growing tropical plants (fruits in particular).

Why pineapples?  Well, I found out from various sites online that the pineapple has an amazingly rich history.  Plus, I love the taste of pineapple   Besides, I figured it would be easier to try to raise a potted plant as opposed to, say a mango tree.

So I found some directions online on how to grow a pineapple and started to grow one from the healthiest crown I could find.

At first I was a bit skeptical as to how hardy the plant would be after reading the directions.  As it turns out, though, pineapples are amazingly hardy with regards to water and temperatures.  The plant survived sub-70 degree weather in my townhouse and office during the winter and wasn't any worse for wear.  But alas, it was not hardy enough for sub-50 degree temps.  I recently moved and since the weather was getting nicer, I started to keep the pineapple outside.  A couple of nights ago, I forgot to bring it inside only to wake up in the morning to find frost (!!!) on my car window.  Frost == dead pineapple

It lived for over a year; not bad.  Quite an attractive plant, too.  I started to grow two more yesterday; we'll see how these two go.

On an unrelated note, I'm starting to suffer with "only" 1GB of RAM on my notebook (unfortunately, the max it can accept).  With VS.Net 2005 and ReSharper 2.0b running on a small project, it's taking up roughly 350 MB of memory.  To complicate matters, all of the 2005 apps (VS, SQL Management Studio, etc.) are such resource hogs, that there is noticeable lag when I use any of them.  I've had to shut down SQL Server for the most part (unless I really need it) and a lot of other services as well.

Merom can't come soon enough...

Came across a great blog while looking for images of praying mantids.

Great links.

Including this one and this one.

Bookmarked!

Hey now! Our inaugural post for May!

Let's start things off with a tip for TortoiseSVN users out there.

First of all, did you know that you can right-drag objects in the Windows GUI? No? Neither did I. I felt like such a noob upon discovering this functionality as I've never, in my several years of Windows usage, ever known about this. Well, in any case, you can, and it's way better than the default left drag that we all know and love.

What's this got to do with Tortoise? Well, I went onto the SVN IRC channel trying to figure out if there was a way to move files in a "SVN-aware" fashion. After asking around and googling, I came across this article that mentioned right dragging.

Note that you will only get the proper context menu commands if the target folder is already under version control; if not, you will not get an option to perform a SVN-aware file/folder move.

This is by far, much more convenient than left dragging and then manually adding the file to the repository in a new location (and removing the file from the old location).

Must see.

I think this pretty much speaks for itself....

Wow.  Trac just impresses me more each day.

Today, it's the peer review plugin.

If you looked up the definition of awesome in the dictionary, I'm almost positive that you'd find this mentioned somewhere in the entry.

It's a component that plugs into Trac and allows you to start peer reviews on code blocks in your repository.  Users assigned to the peer review can view the source code in the browser and make comments on particular lines of code.  There's also  a pass/fail mechanism to vote on how to proceed with the code.  The truly badass part about it is the GUI implementation.

Here's a screenshot of it in action:

If this isn't the definition of awesome, I just don't know what is.

Great use of AJAX.  I did something similar for a different app I wrote recently:

I love this approach for web application design since it's so much less intrusive than opening a new window and offers far better customization of the look and feel of the overall UI.  The big problem is that it's probably not accessibility standards compliant; many screen readers will not properly pick up the change in the UI and the newly inserted HTML content.  If such domains are not a concern, then such a visual design pattern is definitely the way to go if you're building a web based GUI which requires multiple subscreens.