Friday, May 21, 2004 - Posts
Taken out today: I'll be there... in 14 hours.
# of referenced posts by category: Blogging: 1; Development: 1; Other: 4; SQL: 3; WILY: 2
Line of the day: It was us developers that set the whole thing up in the first place, years ago,
just as an excuse to get the PM’s out of the office for a while.
“Sure, head on out and have a ‘conference’, take the whole week.”
That way we can actually get the real work done.
Post of the day:
Query and Transformation: 2 Sides of the Same Coin?
Remember: the Take Outs FAQ is at: http://sqljunkies.com/WebLog/ktegels/articles/1387.aspx
I'm having a very philosophical day for some reason and this time, the target is the cold ware between DBAs and Developers. I think that such a conflict is a dangerous thing with YukonSQL Server 2005. I try to explain why, and what we can do about it, here.
Your comments welcomed.
Sorry, I couldn't resist just outright stealing this from here. They may be signs you've hired a bad DBA, but they also may very well be signs you've hired a programmer instead.
- Believe that adding a foreign key constraint with the ‘Enforce Relationship’ checkbox unchecked is the best of both worlds, it defines relationships without having to deal with ‘those pesky constraint errors’ when modifying data.
- Thinks adding a new column to a table is a four step process: create a new table, copy all the data over, drop the old table, and rename the new table.
- Is convinced that the error generated by trying to insert a duplicate primary key value, is a bug in SQL
- Thinks a stored procedure is the best way to copy one million records across databases.
- Thinks 9 hours to copy one million records across databases is ‘good performance’
- They prefix all stored procs with ‘sp_’ -- so as to be consistent with the Microsoft naming convention
- Forcefully argue that indexing every field in the databases is the best way to increase performance (right after buying better hardware)
- Believe a Stored Procedure which compiles is production ready
- Think performance tuning = buy better hardware
- They can’t understand why changing column names break the application
Thanks to Pooman Lall for the original posting, and to Ms. Clinton for pointing it out.
I'm going to take a departure from my usual Friday Morning Thoughts format
and instead do a brief chautauqua. There is single topic that's been bothering me for a while. It finally came
to head yesterday. No known quantitative-basis exam can measure a programmer ability to
craft quality applications. You can no more certify somebody as a "good"
developer than you can certify a good artistic painter. The best you can hope to achieve
with any testing process is to measure a persons ability to use a tool to some
subjective standard. But being able to say that "I'm certified an expert hammer
user" doesn't make you any more a carpenter than somebody who isn't "certified"
with a hammer.
Continues here.