|
Hello, I'm not looking for an exact answer but some from experience. I have developed/designed small databases before but I have been tasked to reverse engineer a web-based application which includes approximately 900 data elements. I'm assuming this database would require about 30-35 tables. The timeline given preferred is 3-5 days by the customer. Can someone from experience provide me with some information as to how feesible that time line is. I honestly, initially, had given a far much longer time-line. I think the database schema can be erected fairly quickly (1 week) but more time would be required taking into consideration the relationships, indexes, constraints, keys, etc. Any insight will be highly appreciated. Thanks, S
(comments are locked)
|
|
It sounds like you are being set up for failure. I don't know your relationship with the Client but, if possible, it may be best for both of you if you walk away now. Whatever you provide in 3-5 days will be sub-standard and you will always be the one who delivered it, regardless of the unreasonable conditions you were placed under. That many data elements will have very intricate dependancies and relationships. Forget about even attempting to tackle any of the constraints and stored procedures that will be part of the solution. If the Client is looking at retaining the current data your noose becomes tighter. Add access/security on top of all that and you are sure to come up short. Believe me when I say that you want all the relationships defined in concert with your schema design or you will be creating tables for nothing and creating some truly horrendous workarounds later to make it work later as new relationships are discovered. Another thing to add to this timeline is time for tuning and indexing. Tuning alone will exceed 3-5 days.
Nov 01 '09 at 02:23 AM
Håkan Winther
(comments are locked)
|
|
I think anybody asking for that sort of database design in 3-5 days doesn't know the craft. Good design takes thought time - and that time spent thinking about how the system might be used in the future etc will easily outweigh the time spent actually implementing tables etc... and believe me, it will pay dividends...
(comments are locked)
|
|
Why do you need to reverse engineer the database? Isn't there a new application to be developed as well? Don't do the two in isolation. Basically reverse engineer a customersl vendors application to build a customers in-house back-end which they'll use as a contingency plan in case the vendor they deal with pulls the plug on them. They are currently paying the vendor 2.5 mil a month for using this crm type app. Eventually the customer will develop this app using the back.-end database I design..
Nov 01 '09 at 01:48 AM
Slick84
It's the "eventually" bit I have a problem with. It makes no sense to do the database and app as separate pieces of work.
Nov 01 '09 at 04:07 AM
David 1
So... the Client wants to spend ~$3,000 or so to "own" a $30 million/year dB? Tell them for a million or so 10 of us will set aside a month to work on it. Then see what they say. It needs to be put into perspective. It isn't an Excel spreadsheet they're asking for. What value do they derive from this appliation? That is paramount in deciding how much effort needs to be put into this solution. Right now it's a factor of, say, 1/10,000th that they see as adequate. The more details you supply the more I stand by my initial suggestion - walk away from this.
Nov 01 '09 at 11:43 AM
Blackhawk-17
I have walked away. The customer will be out-sourcing this to a company in India where they have promised to reverse-engineer in 3-5 days. I'm looking forward to what the results are.
Nov 02 '09 at 11:51 AM
Slick84
(comments are locked)
|
