I think there are still a lot of doubts on what a successful web application is. There are a lot of successful applications on the web facebook, myspace, youtube to name a few.I am a hardcore engineer. I used to believe that the engineering aspect of a project was more important than anything else. But I have learnt over the past few years that an application is only successful only if it meets the demands of the business. The term application is a combination of the software and the infrastructure.
In my world a successful web application is one which
“Is simple to maintain and manage
Generates revenue
Allows the business owners to do what they need to do to increase the revenue. It could be the ability to make changes at a rapid pace without compromising on stability.
Allows the developers to develop and deploy without a lot of constraints”
The world’s most complicated software will be useless if there is no revenue stream associated with it. The word “complicated” actually negates the success factor. There is so much competition in the world that any delay in marketing a product or inability to make changes swiftly only means that the web application is unsuccessful.A successful business means that all the employees are focused on exploring ways to improve it. The generic term “best practices” is so over rated. I think we should do what is good for the business and that should be the best practice.
I don’t think there is any harm in breaking the rules as long as there is a solid business reason behind it. This is where the employees come into the picture I have seen so many business’s getting into a tangle with employees who just put their interests ahead of the company without realizing that the if the business fails they fail too. I call this the “I” factor. If there are too many engineers with the “I” complex the application/business is doomed to fail. A sense of ownership is great for any business but that has to be backed by a track record of successful and on time delivery. Anyone can build the best software/infrastructure in 2 years and these days the rest of the world would have moved on if we build anything that complex that it takes a long time to be stable.
This is not to say that core engineering values are meaningless but there are so many of them these days that the job of a good engineer should be to pick those which are good for the business like rapid deployments , ease of development , stability etc and not invent something new or reinvent(again)
A solid engineer is one who is not just technically brilliant but one who can use his skill to design simple solutions which are scalable. I think context is very important in the software world. There is no point in building the worlds most secure infrastructure when all you are doing is displaying content but security is important if you are storing user information. I have seen vendors take a financial hit by failing sometimes rather than building a 100% foolproof solution simply because sometimes it is cheaper. Uptime is a classic example.
I think every engineer would have realized by now that sometimes the best code he has written may have not solved anything but something simple would have generated an awesome amount of revenue. As an engineer if you are not able to satisfy the demands of the business and not helping them succeed it is time for you to look for another job.Youtube in my humble opinion had its flaws but it was a big hit because it got there first. I have learnt that the youtube model is a classic evolution model where the focus was one step at a time. I have experience first hand people trying to build a 5 series when a Camry was the order of the day. There is no contesting the fact that everybody wants a 5 series but I think the business can do well with a fully functional Camry to begin with than a flawed 5 series. The ” I will build everything in one shot” will only mean that the business will have to be in catch up mode all the time than making progress. I think this simple concept is so vital that it can cripple a business if it is not followed.
