I've received some great feedback regarding the entry on setting up a MediaWiki install on Windows. Many of the comments were kind words of thanks (which I really appreciate. Knowing that it helps people is my greatest motivation), and others helpfully suggested improvements to the instructions.

As an example of comment-driven improvements, my instructions have you installing the GNU diff utilities, in particular for the diff3.exe utility, however the MediaWiki setup scripts don't properly find it (e.g. as the instructions are currently written the GNU diff utilities are completely unused, although they can still be useful in your day-to-day travails). This is because a prior revision included fairly involved changes to the MediaWiki config/index.php script so it would properly locate diff3 on the Windows platform, as it is currently Unix-centric and doesn't look for the proper executable, not to mention that it parses the PATH environment variable incorrectly . After receiving two comments that those steps were a little too complex, however, I removed that section.

My goal was to get people experimenting with MediaWiki, or even just wikis in general, so diff3 functionality really wasn't critical. I pared the instructions accordingly. Similarly one early draft included the building and installation of a PHP memory cache to improve performance, but that too is unnecessary to simply try out the product.

Another line of comments involved asking:

• Why would I give instructions for Windows. People should just set it up on Linux and go with its native home.
• -or- Why would I recommend a wiki product that largely caters to the open source crowd. Instead I should be pushing Sharepoint, or something properly anointed by the Microsoft camp, enabled with all of the latest Microsoft buzzwords.

To answer this I really need to describe the philosophy of this blog, along with my resistance to "technology alliances".

In the byline of this blog I describe my philosophy as "pragmatic software development", and this really drives my recommendations. In this case there are a lot of development shops that are Windows-centric, with little or no UNIX/Linux experience, yet MediaWiki is one of the best, most featurer rich, "standard" wiki products out there. Choosing a solution that leveraged what shops already know with the best solution is a pragmatic approach.

Which brings me to my general philosophy towards Microsoft, as comments indicating that I'm either a Microsoft hater, or a Microsoft drone parroting the corporate line, have hit my inbox over the short history of this blog.

I am not subservient to Microsoft.

Unlike many Microsoft technology advocates (I truly love both SQL Server, and .NET, and I think they're remarkable solutions), I have no desire to ever work for Microsoft (Microsoft has some top notch, world-class talent, and I've met and worked with a lot of great talent from there, but they also have their share of both jerks and duds). I'm not going to praise their every move in hopes that I'll get noticed. yafla, my consulting/ISV company, has chosen to avoid any partnerships or tying to the Microsoft brand because we don't want to become another drone "consulting" company single-mindedly acting as a third-party sales force for Microsoft, desperately racking up Microsoft partner points by pushing less-than-optimal solutions on customers. We didn't choose to use .NET for our software because we're hoping to nestle into the Microsoft family -- we chose it on technical merit, and a pragmatic analysis of our current and prospective clients.

We work for our clients and ourselves, not Microsoft. This is a very important mantra for our services, and for the technology of our software, and if Microsoft wants their products to get recommended to our clients, and their technology to the foundation of our software, they need to make great products at competitive prices. No sales gladhanding, or sad career dreaming, is going to change that.

Am I saying that Microsoft solutions are second rate? Of course there are examples of Microsoft products that are terrible, and customers are being misled into buying buzzword-laden atrocities because a Microsoft partner is hoping to get invited to the next Microsoft dinner party. Yet there are also Microsoft solutions that are extraordinary. Windows 2003R2 is a superlative operating system, and where you need the breadth of its functionality, it can be well worth the money. Microsoft Small Business Server can be an amazing package of value for some small organizations, within the constraints of the product. Other times, however, if you have the appropriate skills, a Linux machine is the best choice, along with a stack of the many available free or close to free server products on that platform. Sometimes IIS 6 is the superior solution for a problem, while other times Apache would be your best bet. Sometimes PHP and MySQL is a great solution, and other times C#/ASP.NET with SQL Server is the perfect combo.

I don't blindly assume the Microsoft product to be the best, but neither do I automatically presume it to be second rate. Instead I evaluate on merit, and propose solutions based upon the customer and their needs.

To do otherwise would be just biased noise, and wouldn't be to the service of clients and peers.

Tagged: [Software Development], [Programming], [Software-Development]