Just a quick point about a post on Pete-s blog, which in turn references a comment on Tom Kyte’s blog. Pete is totally on the money! Here are a few thoughts related to this:
- You can throw as much hardware as you want at a bad application and it will still be bad. If you design and code applications properly you can run them on less/cheaper hardware and everyone is happy!
- Vendor neutral = lowest common denominator = crap!
- Take a good technology and use it badly and you can easily convince yourself and everyone else that the technology is bad.
- Sometimes, things that sound good in theory are terrible in practice. Did I hear someone mention J2EE? 😉
Over the years I’ve found these to be true time and time again. Remember, it’s just my opinion 🙂
On a different note, my rating super-fan is back again. The DBMS_EPG article once again went to 5 stars for 1 vote, now it’s 3.5 stars for 4 votes. Who is this mystery super-fan?
I recently found out that a couple of friends/acquaintances now read my blog. The majority of my readership for both the blog and the website are based in Asia and the USA. Having some people so close to home seems kinda weird 🙂
Cheers
Tim…
>>Vendor neutral = lowest common denominator = crap!
Tim:
If my mathematics are not too rusty, then It should be “Greatest Common Divisor/Denominator”. The lowest common denominator for all integers is 1.
On the contrary, there is the Least/Lowest Common Multiple.
Of course, I fully agree with the spirit of your assertion.
Cheers,
Carlos.
Hi.
I was using it in this context:
The term lowest common denominator is often used to indicate a lowering of quality resulting from a desire to find common ground for many people: “This fall’s TV programming finds the lowest common denominator of taste.”
That quote is taken from:
http://www.answers.com/topic/least-common-denominator
Look at the section below “lowest common denominator”.
Don’t ask me if it’s mathematically correct. I wouldn’t have a clue 🙂
Cheers
Tim…
You can throw as much hardware as you want at a bad application and it will still be bad.
but you can’t take a well designed application and take away hardware.
The culture clash between J2EE developers & Oracle DBA’s always fascinates me. The former think that databases are bit buckets and all that matters are beautiful designs (without objective criteria), the latter think J2EE is useless and we should all write everything in PL/SQL …
you can write a bad application in any framework; .NET, J2EE, C++, PHP, Perl… PL/SQL… there’s been plenty of skeletons in each of those closets, I guess J2EE is just the current whipping boy.
Tim,
I see this exact situation and it’s maddening. I recently had a client with a HORRID design and it was doing 6-way joins when 2-way would have done the job.
The system was I/O-bound, and we speced-out a redesign of the whole application (600 tables, they paid $4m for it) at $500k and six months.
They wound-up buying solid-state disk for $150k instead. . . .
Same lousy app, but it ran 5x faster and they fixed it practicallly overnight.
Elegant, heck no. . . .
Smart, um well, if the manager had acknowledge that they bought a crapo app, they would have lost their job . . .
“Smart, um well, if the manager had acknowledge that they bought a crapo app, they would have lost their job . . .”
You just defined in one single sentence what is wrong with IT.
For once: cannot agree more.
The application will always be bad once it is done bad. A lot of hardware can never save it nor some additional features. Still, the application should be carefully selected as to have maximized results.