Hi guys,

first of all, i would like to thank you all for your help. Your opinions 
are wise enough to confirm me what I was expecting. So, i'll stick with 
the current solution for this project, leaving open the possibility for 
a change on a near future. Not because I'm afraid of the time needed to 
learn and apply the knowledge on these subjects (Spring, Hybernate, Red5 
or openAMF), but because there is a risk associated with my inexperience 
with them. Spring and Hybernate do not scare me at all, what scares me 
most is what will happen when unexpected things occur. Especially when i 
do not have experience with Java on the web - this may not seem a big 
deal, but it's important to be comfortable with specific stability and 
security issues, among others, to know from the start that the server 
won't crash, or act unexpectedly.

After finishing onFashion, i'll dedicate my time to change the current 
framework to use Spring, Hybernate and openAMF, apply it to future 
projects, and refactor the server-side code of onFashion. When Red5 
seems stable enough for production server, i'll replace openAMF for Red5.
The hurry with using this already, was because:

- i am completely tired of the code to access the RDBMS, convert it to 
objects, and vice-versa, which takes a lot of development time not 
focused on the business rules or usability - so i would use an ORM 
solution, like Hibernate, which would require Java
- comparing to PHP, Java is much more mature solution, and preferable 
when we desperately need maintainability - so i would change PHP for Java
- once you use UML live bi-directional synchronization with the code, 
you cannot go back. I've used it with Java, two or three years ago, and 
i said i was never going to stop using it. I break my word when i chose 
PHP for the server side code of my RIAs ..... I want to go back to the 
bright and shiny days of UML - so i would change PHP for Java
- onFashion will have videos (and this is an important part of the 
project), and it's preferable to have streaming to serve them - so i 
would use Red5
- onFashion will have a chat at a later time - so a push server 
technology is vital, just like Red5
- CastingOffice (another project) and onFashion on certain sections 
require to have the clients synchronized - so a push server technology 
is vital, just like Red5
- I believe that if the server can handle it, a push server will be the 
future for enterprise RIA's, and very dynamic web-sites - .....Red5

and the last one:

- assuming all those premisses, it would be preferable to invest time 
developing onFashion already using the referred technologies, than to 
wait, develop the framework and refactor onFashion server-side code, 
which would take more time and resources

Anyway, you are completely right. Time is not enough to take the risk of 
having to deal with unexpected issues.

Unfortunately i feel like a little kid playing with big toys. I do have 
initiative, will, and enough knowledge to try to use the best techniques 
around there - improving our Agile development environment is what i 
love to do, and that's why with so little resources we normally give a 
great response. But i have finished my graduation two or three months 
ago, i'm starting on my own, and as you can imagine money is something i 
do not have for now so it's impossible to make investments like 
recruiting more people to work. Worse, life in Portugal it's not easy 
for who's starting, and definitely there is lack of knowledge among our 
developers - i risk to say that only about 5% of the Portuguese 
developers knows how to use a design pattern.

Thanks a lot once again for your opinions, they surely helped me to make 
my decision. :)

Best of lucks for Red5, and also for Chris Allen and his beloved one. :)

João Saleiro



_______________________________________________
Red5 mailing list
[email protected]
http://osflash.org/mailman/listinfo/red5_osflash.org

Reply via email to