Over the last year I have gotten an education on PHP and MySQL web sites to go along with my existing expertise with ASP.Net and SQL Server.

It turns out that I purchased a web site a little over a year ago that supports gamers who play World of Warcraft (a game I have played for years). The site gets about 100,000 unique users a month with just shy of a million page views a month. The site was written in PHP against a MySQL backend and is just not driving the revenue yet to justify porting it to ASP.Net and SQL Server (though as you will read here the balance of pain is shifting that equation). It turns out that we end up rebooting the system pretty damn often which was a problem with IIS back in the old days, but not one I have had in recent versions.

We have thrown more hardware at the system, brought in professional help and it just seems that at these levels of use the system runs down and needs a kick and sometimes intensive care.

My point here is that it has been an education for me to validate what I suspected, there is no magic with the non-MS stack. It can hang in some regards, but it seems that for really heavy loads, MS has got them beat on stability. I am working on an ASP.Net with SQL Server site now that handles similar traffic and it just doesn't suffer the same issues.

I plan to dig deeper into the tech here if for no other reason to figure out what it takes to port the site to ASP.Net with SQL Server.

I have been working on commercial products for a long time and repeatedly have seen companies compete with similar solutions. Often one is the technology leader and innovates while the other plays catch up and only survives by clever marketing. Sometimes the laggard can become the market leader, but typically only if the innovator makes a mistake (the classic example of a market leader losing ground due to a mistake is when New Coke came out).

When it comes to software products the rule is pretty simple, mistakes in usability are the ones that cost marketshare fastest. Customers are pretty tolerant of technical issues and bugs since all sofware has them, but if the user feels stupid when trying to use your product, they will switch very quickly to an alternative.

Bottom line is that mistakes of ususability are more costly in a competitive market than almost anything else, design wisely.

I just got back from the Microsoft PDC in LA and have been thinking about what I saw there.

It turns out that I have come to a couple of conclusions that I will surely post more about in the future, but for now here is the overview.

First there were several Windows Azure announcements that have swayed me from skeptic to seeing a real chance for Azure to be a contender. Chief among my concerns was the fact that I just didn't see companies doing a big rewrite just to leverage a cloud solution. Now it is much easier to port an existing application to Azure and there is the option to customize the hosted image. I also saw a demo that no one else seems to have noticed (or I was imaging things). I could have sworn I saw a demo where SQL data hosted behind the company firewall was opened up for consumption by an Azure hosted application. I plan to watch that keynote again to make sure I know what I am talking about so consider this a disclaimer.

Second, I am now confident that Microsoft will not abandon either WPF nor SilverLight developers since there were already announcements to make both able to run with the same assemblies. A small step, but when coupled with the fact that VS2010 is built with WPF I think the two technologies are both valid for development (I was worried about the future of WPF until recently).

There was of course more, but those will have to wait for other posts.

For many, many years I have been writing and reviewing contracts between my company and clients. As a result I have some insights into how things can be made to work more simply.

First up, this is not legal advice, just me sharing some experiences. You should always run your contracts by your lawyer to ensure you aren't painting yourself into a corner you did not intend.

Second, I have always tried to standardize contracts as much as possible and educate prospective clients up front as to what our process was for setting up contracts. Often the client will have their own ideas and their own contracts, but life is much better if you get the majority of clients to use your system rather than having to make a project out of every deal. I find that the more reasonable my process and contracts the more likely the client will accept my contracts rather than insist on using their own.

Third, you must always remember that contracts are to govern the relationship between you and the customer when things to wrong. They almost never come up when the project comes off to mutual satisfaction. They are insurance if done well and they are a death sentence if they are done badly in cases where the project goes off the rails.

Fourth, contracts are not personal, they are just part of business. If you are doing business with someone you like and trust then there is a temptation to skip on the contractual completeness or correctness. THIS IS A MISTAKE! Always think in terms of what would happen if the project went sideways and the person you had to deal with was not the one with whom you set things up. This has happened to me on a regular basis and the only defense is to have solid contracts.

I hope to post more information like this in the future.

I am packing tonight to head to the PDC in Los Angeles and wanted to tell anyone else who will be attending that I am hosting a Birds of a Feather session at lunchtime on Thursday on security hype.

The thesis is that we are seeing a steady stream of over hyped security "issues" that tend to remind me more and more of the ads for the evening news that say things like "Your water could be killing your children, details at 11". We plan to discuss how this trend is hurting actual preparedness for the real threats.

Hope to see some of you there.