It is not a very good idea. Why do you think that serious intra day traders don't use web based platforms. Too inflexible and too slow.
Depends on the context I guess. I have a Flex-based dashboard/control panel for my ATS infrastructure - I can load it up anywhere with a web browser (usual web-based security mechanisms in place) and take a look at what's going on, monitor status, positions and do some management functions. I don't need charts at this level but there's nothing in the technology to preclude that from being incorporated too. In the context of automated trading (as per this thread) what else do you need? For discretionary trading I would say your argument is more valid. Again, in the context of the thread, I see nothing wrong with using semi-thick-client technology that can be embedded in a web browser i.e. Flex/Silverlight as a front-end for an ATS. It's a matter of determining the scope of functionality for the front-end to see if it will be feasible. BMe.
Put it this way, does your broker provide web-based trading as an alternative to a desktop client? Retail-level brokers like Interactive Brokers, thinkorswim, OptionsXpress etc. do for example. Anything you come up with can be as secure as your brokers web-based trading platform if you want it to be. You may want to get an SSL certificate for starters.
You can generate your own self signed certificate. Don't need an external trusted CA to sign it if you aren't trying to convince anybody else of the authenticity of your site. No less secure.
I am also automating my trades at the moment. If you want to be as flexible as possible I suggest you write everything yourself. With that I mean that you dont use 3th party software, you almost always come to a point where you want to do stuff in 3th party software that is not possible but is a very doable in your own code. Hiring someone or work together with a good programmer is also fine but better is to write the code yourself so that the system is you. It will take longer though. Begin with a system that is based on reading historical data from flat files and fake ordering. When you want to go live you replace the historical stream with a live stream. Make sure that you can switch easy between historical and live stream. Don't fixate on brokers, live order implementation or datastream implementation yet. It is very comfortable and important to have a system that can easily test ideas first. Once you have something that works on historical and live data you choose a broker, implement ordering and start trading on a papertrade account. Once that is stable you start trading on real money.
Again, thanks for the replies. I think that the best way is to start with some of the user-friendly ideas that you have recommended. I know what i need to get done, but i dont know how, so maybe i can start with some baby-steps and see how it goes. However, i still think i will work with a programmer, because my ideas need a gray box strategy with some discretionary action/execution. The programmer/partner friend of mine is looking to get into the trading business, and perhaps this is a good start for both of us. Again, just for the story, i AM a profitable trader, and my strategy is a "gimmick" that works well until now, but dont know until when of course... So basically i dont want a pure order execution and position covering program. And i need an operator as well, who will be taking positions, and i cannot do that because i will be trading my own account while the ATS is taking action on a second/third account. Simply, i need an extra hand (or two), to apply my strategy to more symbols than i can usually handle. I want to multiply my strategy, capitalize!
noone3 multiplying ATS's in cyborg is quite easily done, you can run the same or different strategies on any number of products at the same time. They also have a basket feature allowing to launch ATS on sector or correlated stocks. The strategies you can build are very customizable and you can interact with them if need be.