williamweber.net

Let me start by admitting that I have a problem. I like iPhone Apps. I like downloading them. I like using them. I like bitching about how shitty they are. I like deleting them and leaving 1 start ratings. I pretty much like everything about them. I’ve downloaded hundreds, come really close to maxing out my 11 home screens and I’m always looking for new ones to play with. So the idea of an app that recommends other apps to me is pretty much crack dipped in chocolate to me. My new favorite is Chomp.

Chomp is one of a handful of recommendation apps that I’ve found. Others that I’ve tried, and ultimately discarded, are Chorus, AppMiner, BargainBin, and PandoraBox. The latter three are primarily discount and sale finders so they may not be directly comparable, but Chorus is billed as a recommendation network very similar to the Chomp. I had high hopes for Chorus. You had a friends list so you could get recommendations from friends. There’s a concept of “feeds” that my assumption was you could subscribe to and get turned on to things you may not find otherwise. There’s a feed for friends and an All feed that originally I figured was just any and all recommendations by Chorus users, sort of like the Twitter public timeline. That’s not really the case though. I really have no idea what the All feed is in Chorus. But in general it’s crap. The interface design on Chorus is mostly poo as well. It takes about twenty seven clicks to see a list of reviews of a specific app. The app uses a white on black design that always puts me off from the start. And the “App Mavens” in general, are douche bags.  Basically Chomp succeeds in just about all of the areas that Chorus fails for me.

Read more…

Recently I’ve been looking into new ways to deal with extremely large datasets. With Twitter and Digg making announcements recently that they’re migrating, or have migrated their operations to Cassandra, I’ve started to take a keen interest in NoSQL. NoSQL is a term that casts a wide net. In general it encompasses data stores that eschew the traditional relational database model in favor of distributed data stores. Google, Amazon, Facebook, LinkedIn and a host of others have all thrown their hats in the distributed data ring. I mention those companies specifically because all of them have built their own distributed database offerings to satisfy their particular needs. Google has BigTable, a propriety technology that has an open source flavor under the Apache Software Foundation project Hadoop called HBase. Amazon has Dynamo (even more proprietary). Facebook created Cassandra and open sourced it thankfully. And LinkedIn developed Voldemort (a great name if I do say so).  Twitter and more recently Digg.com have gotten behind Cassandra which has given that particular project (not even up to version 1.0 yet) a lot of press lately. I’ve started to dabble with Cassandra and HBase (really the whole Hadoop project) in an attempt to get familiar with technology but I keep coming back to something in my mind. That is that “NoSQL” is probably the stupidest possible name that could have been coined to describe all of these amazing technologies.

Read more…

Is a Service Level Agreement (SLA) with your users a waste of time. As far as I’m concerned the answer is yes. I’m sure there are folks that will come up with all sorts of “it depends on the situtation” answers to the question, but I’d say that for the most part those folks are, either detached from the day to day maintenance of whatever is trying to be measures, or untrusting of the people who are charged with those duties, or perhaps a little of both.

Just so everyone is on the same page let me start with some assumptions. An SLA is an agreement with your customers (whomever they may be) that you will maintain a certain level of service. It’s a contract that both sides agree on, one side attempts to maintain the terms and another enforces those terms. When dealing with a vendor, an ISP or hosting service (I come from a tech background so we’ll stay in that field) an SLA may be necessary. It gives the customer potential recourse if the vendor fails to meet the terms of the contract. But when talking about departments within an organization creating SLAs with each other I’d argue that any time spent on creating and attempting to montior or enforce an SLA is wasted time better spent doing just about anything else.

Read more…

I did a presentation at my office recently about the basics of XNA 3.0 and what’s possible with it. During that presentation someone asked me if XNA was being used by any major developers to create games. At the time I didn’t have any specific examples but I did say that I believed the platform had the capability to do so. Yesterday I found an article that talks about the new Rock Band Network a joint venture between MTV (which owns Harmonix, the company that developed Rock Band) and Microsoft to create a community driven platform for adding content to Rock Band.

This is pretty much what I’ve been expecting from one of the two music games from the beginning. Some way for emerging artists and bands to throw their master tracks up and have something spit out the appropriate colored buttons on virtual fret board. I actually think the Rock Band Network is an even better solution than some sort of algorithm. Now, there’s a way for labels, bands, or studios to put a track on the Network and have a dedicated community help create the virtual track.

Now, being that this Network is actually a subset of XNA Creators Club it’s unsurprising that the tracks created through this network will only be available on the Xbox 360. Bummer for those PS3 owners, but in the grand scheme of things a venture like this is only going to improve the music game genre which, in my opinion, has become incredibly stale and over-saturated. Maybe I’ll even dig my plastic instruments out of the garage and turn the game on again.

Here’s an excerpt from the article:

The Rock Band Network is the result of a 16-month development process with a number of partners. Most important was Microsoft. Tracks released through the Rock Band Network will only be available at first to Xbox 360 users, as it relies on Microsoft’s XNA game development platform and its Creators Club online community of developers.

The Creators Club allows freelance developers or hobbyists to make their own games and sell them on the Xbox Live Marketplace. Games created through this process must first be submitted to the Creators Club community for game-play and content review before they are added to the marketplace.

The Rock Band Network marks the first time that XNA and the Creators Club have been used to outsource the development of expansions to an existing game, according to Dave Mitchell, the Microsoft product unit manager in charge of overseeing the two programs.

The Creators Club reviews about 30-50 games per month. Because MTV and Microsoft expect the number of “Rock Band” submissions to quickly dwarf these totals, the software giant took the unprecedented step of creating a custom version of the Creators Club for Harmonix, complete with a customized set of review procedures specific to music games-including checking for copyright infringement-which Harmonix will host separate from the existing Microsoft site.

We use partitioning quite heavily at my current job. And in order to take advantage of switching data from a partitioned table into a staging table all indexes must be aligned on the partition scheme. So, I was frustrated to find out that when you attempt to script out the create statement for an index in SQL Sever Managment Studio (SSMS) 2008 it does not include the clause to create the index on the partition scheme that the index was created on.

Thankfully, Microsoft wasn’t so dense as to remove this funtionality entirely, they simply turned it off by default (which I would argue is ridiculous, but they aren’t really going to listen to little old me). So, here’s the fix.

Open up Tools -> Options then change the setting at: SQL Server Object Explorer -> Scripting -> Script Partition Schemes to TRUE. (it’s at the very bottom of the list)

I’ve wanted to tinker around with creating videogames for years. I’ve been a “gamer” for most of my life. We had an Atari 2600 when I was very young because my dad thought it was cool. The Nintendo Entertainment System was the first games system that I could really call my own and I’ve owned just about every system released since then. I’ve built my own MAME joystick layout (that’ll be coming in a future post) but I’ve never been able to break into the creation realm. The price for entry was just too high. The level of programming knowledge, or the tools necessary were always a little out of reach for the amount of time and effort I had to spare. However, with Microsoft’s XNA offering the whole prospect seems to finally be within reach. Read more…

I wanted a more granular level of detail when looking at disk space usage of our databases. I wanted to see the disk usage of each table. Since we use partitioning pretty heavily I also wanted to see how that usage was broken out by partition and filegroup. In our implementation we have a separate filegroup for each table partition but in a configuration that has multiple parititons on a partition this script can tell you usage for either metric.

Normally I’ll take this query and take the results over to excel and pivot the data so that I can filter and sum data in whatever way I want. I also have this running in a SQL Agent job and inserting into a table so that I connect to that table and do growth trend analysis.

Read more…

Here’s a quick script to find the recent history of file auto growth on a MSSQL system. I work with a large OLTP database that’s partitioned monthly. If we get files auto-growing toward the end of a month when the partition is already upwards of 60GB, timeouts start to crop up. Sometimes finding out where those timeouts are coming from aren’t readily apparent. This script helped us out when we were searching for recent file growth.

SET NOCOUNT ON
 
DECLARE  @path VARCHAR(max), @cmd VARCHAR(4000);
 
SELECT @path = CONVERT(varchar(max),value)
FROM fn_trace_getinfo(DEFAULT) WHERE property = 2
SET @cmd = 'dir /b "' + @path + '"'
-- PRINT @cmd
 
DECLARE @files TABLE
(
	fn VARCHAR(64)
) 
 
INSERT @files
EXEC master..xp_cmdshell @cmd
 
DELETE @files
WHERE fn IS NULL
 
SELECT
	e.DatabaseName
	, e.[FileName]
	, e.SPID
	, e.Duration
	, e.StartTime
	, e.EndTime
	, CASE e.EventClass
		WHEN 92 THEN 'Data'
		WHEN 93 THEN 'Log'
	END AS 'FileType'
	, f.fn AS 'TraceFile'
	, te.name
FROM @files f
	CROSS APPLY fn_trace_gettable(@path, DEFAULT) e
	JOIN sys.trace_events te
		ON e.EventClass = te.trace_event_id
WHERE
	e.EventClass IN (92,93)
ORDER BY
	e.StartTime DESC