Re: Performance optimization vs satisficing (was Language Oriented Programming)

On Fri, 22 Jul 2005 16:42:28 GMT, "H. S. Lahman"
<h.lahman@xxxxxxxxxxx> wrote:
>I suspect the problems you end up fixing to get the first 300% are often
>the result of developers simply not caring about performance because of
>that obvious hardware gap. IOW, they look at the hardware gap and
>naively assume hardware will take care of everything.

I'd say the most common causes are:

* Some maintenance operation dropped the primary key and nobody
noticed.

* Someone felt the table didn't deserve a primary key, which might
even have been true at some time zero for all I know, though in my
world everything has a real PK plus or minus an identity/surrogate.

* They make some code change and never even bother to check
performance, which may (if checked) even seem OK especially in test
environments but may degrade exponentially over time on production
volumes.

--

IOW, *really* dumb stuff.

J.





.