keep in mind it's a matter of scale. If the application will take several years to 
grow beyond the current requirements, then you're fine. If the requirements are going 
to grow rapidly, then you may still want to use Oracle for your mined data.
If you're dealing with only a couple thousand files and the growth rate is managable, 
then file system should last for a while. On the otherhand, if you know within the 
first two years the number of files will double each year, you may want to create an 
abstraction layer for yourself. This way you can store the summaries locally for now. 
then when the dataset size grows past a certain point, you can implement oracle and 
move the storage of the mined data in oracle. That should give you enough flexbility 
to grow without rewriting huge percentage of the app.
 
peter
 
 Swapneel Dange <[EMAIL PROTECTED]> wrote:yeah i know what ur saying PENG , but i am 
still at a very basic stage in 
the FILESYSTEM implementation , thats why before i move any further, i was 
making sure as to whether considering the ROBUST structure and all that 
stuff into account will it be better if i implement the SQLPLUS or stick to 
the FILESYSTEM!

comments awaited !



Swapneel Dange
505-642-4126
http://www.cs.nmsu.edu/~sdange






---------------------------------
Do you Yahoo!?
Yahoo! Tax Center - forms, calculators, tips, and more

Reply via email to