I must start with a declaration that what im about to write is my own perspective on things and experience of working in startup company’s that failed or prevailed. that expeirence showed me the weak points of bad planning and the selection of a proper technology to start with is an ingredient and even in some cases, the difference of having a successful application and a doomed one.

I’ll start of by showing an example of a case i’v witnessed of improper selection and planning ahead with the right technology. When choosing a technology at the first stage of blueprints you always have to think ahead what your going todo. for example. Company ABC decided to build a social app with a strong backend server that serves a client and a web app. naturally they started building the strong server at first with a core of server team developers that did an excellent job of doing so by choosing java as their main technology. In the process of building the server they started developing the web app in the proper web technology of opensource php. That was a proper selection. problem is php and java dont mix. thinking ahead of selecting the technology would have avoided the trashing of month’s of work down the drain and rewriting everything from scratch using jsp applets. Also the fact that choosing conformist technology like zend to bridge between the technology was declined. Its very important to chose widespread technology in order to avoid critical mistakes.

Some might say php is too old and you might want togo with django or ruby on rails and other exotic technology’s. i say no. html has been rocking since the 80’s. why go for something that didnt pass the 1.5 version and build production web apps on it because its trendy? When choosing a technology you always have to concider the work force that is out there if you need to hire some. just because you know one good ruby coder doesnt mean its widespread. fact is that most hosting companys dont offer ruby containers at all and stick to the regular LAMP configuration. but we’ll get to that later on.

When you start off with your project always do a case study and think over everything around the technology and not the technology itself. when java can do amazing things you will have some hard time finding good coders for it. if java puts out 244,000,000 results on google then php puts out 2,960,000,000 results showing off the widespread of opensource communities that offer alot of work and frameworks preready for you that will save you alot of time with time consuming mindless development of stuff that people already developed. if you want to go for rentacoder like i did for a few times it would be alot easier and cheaper to find a small team of developers in php then in ruby,django or java all together. check market popularity for technology before going forward with it.

To just graze the top of the hardware side. there is no proper web server built for serving java as a web technology. jboss and its children were meant to serve as an application server. tomcat isnt proper for that task either. and you will not find ANY company that offer hosting service, dedicated that is ready for java or ruby or django. all of them require special planning ahead and knowledge of linux for installation of all the modules. the costs are accordingly. you can save alot of money renting a dedicated server preinstalled then buying a server and renting rackspace for it then hiring a personal that can install all the stuff for you.

Try to simplify things, dont try to reinvent the wheel when it comes to technology. if everybodys doing A dont try doing A+B+C+D. someone probably already tried doing that and failed. you will too or just spend alot of money getting where other have passed because they went conformist with their technology.
Ready for your dev kick off? do another small test. go thru all the major websites and check what they offer API wise. if fickr offers a php api pack before a java one it means you need togo with the php one. checking what the market is set on will help you cut down on costs for developing alot of work arounds for other web app’s api’s.
Do a test and try to change or add some line of text or button somewhere on your web app. if it takes more then an hour you got something crooked and your on your path to oblivion. websites need to be dynamic like the market. if one day purple rules and trendy and then other day white is you need to be dynamic and fast enough to respond to the market. think of a broker that trys to sell stocks with a 3 day delay. its critical and will lead to disaster if not.
i have summed up a few rules i use for pre planning any project that i go with:

1. Check what everybody is doing.
   innovation is nice but your car runs on fuel for a reason, dont try sticking milk because you think it might work. the market already chose what to use and what is good.
2. Never develop anything before you check if someone already did that.
   if you want a blog use wordpress or other opensource software that reached maturity. your blog will suck compare to other blogging platforms out there. Use online community even like wordpress.com or typepad for your corporate blog. you might even get users thru there also. use opensource.
3. Never let a coder decide on technology.
   most of them are niched coders that will code that they know best and not what is good for you. hire a proper Chief technology officer to fiddle around with technologies and let him decide on what is best. if you think you can do it do it. it aint that hard.
4. Dont be paranoid.
   Its good to have your guards up but dont think that linux is a security hazard because its opensource or that wordpress holds backdoors that can risk your company. if they were it would already be out there. your not the only one using that. People can do stuff better then you.
5. Stop making things complicated more then they are.
   If you can do it the old fashion way dont try changing world orders because its kewl or that it might make you look good for your investors. At the end you will eat what you cooked when it all gets so complicated you wont be able to hire anyone that can work for you because your the only one that understands the ultra complicated matrix system you developed to publish a blog.
6. Be receptive to technology.
   dont think you know better. being arrogant will lead you to failure when choosing a technology and developing a successful web app.
7. Be dynamic and choose wisely.
   If you chose a technology for your web app that requires you to compile stuff and have complicated scripts to update a website your on the wrong path. websites need to be dynamic. if you want to change the design or add a new feature it needs to be done quickly and not weeks but hours or days.
8. Always wear sunscreen.
   its good for you.

Hope i helped in something and remember this is my perspective so if your a java addict. go comment else where. like on slashdot or sun.com.

   

• BROWSE / IN TIMELINE
• « Safari hangs while logging into Gmail
• » Cross browser testing done right

• BROWSE / IN General New start? service web 2.0 model social web savvy security linux software google develop
• « Safari hangs while logging into Gmail
• » Cross browser testing done right