HTML5 and great expectations

It's a busy day full of chatter on the InterWeb about HTML 5 and its proposed features. "Could it kill Flash and Silverlight", as supposed by Paul Krill at InfoWorld?

One thing is certain: it is not vaporware. After a short search, I found this example implementing a coverflow-style animation from Adnan Osmani (nice work!!) http://blarnee.com/wp/an-improved-javascript-coverflow-demo-using-canvas-and-html5/ If you have an HTML 5 compatible browser - it worked in my Firefox version 3.0.11 - then give it a try. It replicates the animation you'd previously require Flash to do.

My prediction is not can it kill Flash, but when will it kill Flash. Open standards, after gaining momentum, are unstoppable. I'm sure companies like Adobe realize this, and the newer Flash application software will just generate HTML 5 compatible elements. What provides a heavy anchor on progress is how slow users are to upgrade their web browser to be compatible with the new features.

Today I was waiting for files to upload to the server in order to upgrade Drupal (in fact it is Acquia Drupal) and CiviCRM for Drupal, and while I was waiting I started writing this.

If you don't know what those software programs are, I will not explain here - you can go and read about them. What matters is they are web applications, meaning they run on a web server and you use them in your web browser.

Many computer users - hundreds of millions of them - have from time to time installed software or upgraded software on their own computers. On Microsoft Windows, Mac, and others it's a fairly painless procedure of clicking the "Next" button as you confirm some choices. The computer does some work, you may restart the computer, and 19 times out of 20 you're good to go. Web applications are not that simple.

The general process in upgrading web application software is:

• Backup the files, and the database (basically backup every single thing you can find)
• Delete the old files, which sometimes means all the files
• Upload the new files
• Upload the themes and customizations you had made to your system (sometimes this must be done after running upgrade scripts)
• Run an upgrade script, to upgrade your database. Sometimes, it may even help you merge the customizations you made to templates and files, such as how phpBB3 does).
• Add back all the customizations you may have done (if they were wiped out), and confirm everything is running correctly. You may use some tools to help you merge the new files with the ones you customized.

These usually require any combination of FTP, tar, gzip, zip, ssh, svn, SQL, Linux, and lots of specific knowledge to debug problems that may arise. There's typically no one click install, and if there is, it may work correctly 2 out of 3 times.

The web application software, through its available functions and possible customizations, are very complex. I dread having to upgrade the software because of the possible headaches that arise. But, I don't think it should necessarily any easier to upgrade it. Making it easier would lower the bar of expertise required to upgrade the software, and lowering that bar makes those who should not be upgrading the software themselves to think "it's not that hard, why don't I do it myself".

Unfortunately having a difficult upgrade process causes a steep learning curve for software administrators, and ... it's probably really scary for many people. When a security update is released, they are dependent on hard to find experts to help them.

There are so many adages that apply just as well to computers and software as anywhere else in life: be prepared! An ounce of prevention is.... The wise ones will use prevention to their advantage.

For example, if you run a hosting server, and a users' domain name is about to expire, would you rather:

1. Send a short email or make a phone call to check it will be renewed by them? Or,
2. Wait until the weekend, the web site may or may not go down. If it goes down, you may be scrambling to try and fix it.

You can think of it in terms of a game of chance. Let's say sending the reminder will take 10% of the time and headache a downtime causes. 50% of the time, the domain name doesn't get renewed, and the downtime is caused. Then that ounce of prevention wins.

It's not magic to keep computers running and keep bugs to a minimum, though to some it seems like the "black arts". Developers and IT professionals experience  much they wish never to repeat again. The good ones will learn from their experiences, and if there's a known cause, next time that situation appears they prevent the bad experience from happening again.

As the years go by, this experience and knowledge starts to gel into some kind of wisdom. Wisdom is not just about fixing all the problems you've seen before , but is more like a "gut feeling" about what is a good choice and what is a not so good choice, and what you should or shouldn't do.

A couple years ago I was discussing this with a someone. Well, I was probably just blathering on about how I had fixed something that I noticed was wrong in some software, but this error was not my responsibility. Funny enough, she has done the same thing before, and for the same reason. We didn't fix the error primarily because we care about the customer. We fixed the errors because when it goes disastrously wrong, we will have to fix it, and it will be far more time consuming and frustrating to fix it than to ensure it never happens in the first place.

In software development, this wisdom works in an ongoing basis. We anticipate how the application will evolve over time. Perhaps we'll add a field in a database table for the date a record was created, because there's a good chance we may need it in the future. I'm sure there are hundreds of ways wisdom allows software and computers to run more smoothly. Unfortunately wisdom cannot be taught, so to gain more wisdom always be one the lookout for ways to anticipate disaster (and prevent it).