When I entered the title of this post, a big yellow exclamation mark (inside a dull gray triangle) appeared above my head: "Warning: potential circular reference!" But I think it's useful once in a while to step back and think about what you typically blog about, and how it can impact the community. Simon Sabin recently mentioned three very interesting facts that I hadn't really thought about before:

1. People tend to blog about things that are not only useful to others, but also useful to themselves.

   I think this is absolutely true. As an example, I posted a series of articles a few weeks ago about splitting comma-separated lists. Did I do that to share my knowledge, help you solve your own related problems, invoke arguments, and show you how smart I am? Of course. =) But when I think hard about it, there was probably some subconscious effort to document the issues I came across in my own work, so that I would have an easy place to recall them later (and reuse the code). Since my blog is much easier to scan than trying to find a code example deployed on some server and/or in some source control system that I may no longer have access to.

2. Your most popular blog posts might not be the ones you expect.

   This I can see as well. Typically your blog has a steady or very slowly growing audience, but then if you step out of your "comfort zone" and talk about something that might be a little off-topic for you, you can experience a burst of traffic. Kind of like when celebrities appear in the news for doing a good thing when they're usually doing bad things (or vice-versa). This all depends, of course, on how your content is publicized. Just to highlight, here are my 5 most popular blog posts over the past 3 years here at SQLBlog (in terms of overall views). Note that one of them has very little to do with SQL Server at all.

   • The XP SP3 / MSXML6 SP2 / SQL Server debacle

     There was a very loud outcry earlier this year when people discovered that Windows XP Service Pack 3 changed the way MSXML files are marked by the system, which in turn prevented many SQL Server installations from succeeding. The bulk of the traffic has likely come due to the enormous level of feedback the post has generated: 48 comments as of this writing. In fact the issue is still looming, as Microsoft's SQL Server Support team recently posted some information about the workaround, claiming that if you call them they will tell you how to add code to your setup programs that will enable the SQL Server install to succeed. The automatic fix is simply deleting a registry key, which they have asked me not to disclose. They can't really stop me, but I don't want to make any enemies at CSS. Ho hum.

   • SQL Server 2008 System Views poster is now available for download!

     I can't explain the traffic to this one. It is just an explanation that the PDF version of the poster is downloadable, and there is a follow-up comment that explains that some company online is selling printed copies. I'd imagine the notices that were posted on the SQLMag, PASS and SQLServerCentral web sites would have generated a lot more traffic than me.

   • My stored procedure "best practices" checklist

     This one was floated around on Twitter IIRC, and generated a lot of feedback (46 comments so far). This is likely because a lot of my best practices are not necessarily aligned with your best practices. Things like coding conventions and naming standards are a lot like cars or cell phones: everyone has their favorites, and not everyone who likes a Porsche is also going to like a Bentley. As I say often, just pick a standard way to do things within your group/company, and be consistent; your standard doesn't have to match anyone else's except your own.

   • More info on SQL Server 2005 SP3 + Cumulative Updates

     Back in December, there was some real confusion about whether customers should update their instances to SP2 + CU11, or jump to SP3. The problem was that people who had installed CU10 or CU11 already would *lose* those changes if they installed SP3, since the SP3 branch of code was frozen before CU10 or CU11 came out. After speaking with members of the Service Releases team, I published this post, explaining the differences and what people should do based on the build they were running at the time.

   • Using a Mac in a Windows world

     This is another one where I can't explain the traffic. It is simply a link to a blog post I made over on my company's web site about the ins and outs of choosing Apple hardware and Mac OS even though my job is completely Windows-centric.

3. People are afraid to provide redundant content.

   Simon also suggests that you shouldn't be afraid to blog about what's in the documentation. I concur wholeheartedly. This is primarily because, quite frankly, people don't read the documentation unless you point them to a specific topic. And even then, they don't "read" in a lot of cases, they merely grab the code sample and move on. When I write about a topic that appears in Books Online, I have the opportunity to provide my own commentary and opinion about the feature or topic, point out missing or incorrect information, and link multiple topics together in a way that might not otherwise be obvious. None of these things could be accomplished when I just say, "Go read topic x in Books Online." Don't get me wrong, Books Online is a fantastic resource, and one that everyone working with SQL Server should be familiar with; for the most part, though, it sticks to the "Just the facts, ma'am" mantra.