are new benchmarks needed?

[maarten]

So your saying,

when you go buy some vegetables you travel back in time to verify that
they indeed did not use any pesticide as the package mentions?
I can see it now, just because the scale in the store says it's 500
grams of tomatoes, you're the guy that takes out his own scale and
measures it again.
And really, you shouldn't be using any library or API before you
thoroughly tested each function to see if it does exactly what the
documentation says right!?

The documentation and performance test is exactly there to tell me if it
suits my needs.  Only if the description of the test or the
hardware/situation/... is different enough from my own situation may I
chose to do testing for myself.  I don't like wasting my time to do
something someone else already did, probably even better than me at
that.

I'm not saying you should never test anything, but that is closer to
reality than testing everything yourself.

regards,
Maarten

P.S.: You're at a store for a new phone that costs 100$, but your friend
stops you and tells you that that other store has the same phone for a
much cheaper price.
You go over to that store and find it costs 99.99$.  0.01$ less.  I
would punch (friendly punch) my friend at that point, won't you?

On Wed, 2010-12-29 at 10:57 +0100, Hongli Lai wrote:
> On Wed, Dec 29, 2010 at 3:21 AM, Charles Kerr
> <charles at transmissionbt.com> wrote:
> > Oh come on... "A is faster than B" is *clearly* not the same as "No
> > practical difference between A and B."  Saying they're equivalent just
> > because they're both true doesn't even pass the laugh test. :)
> 
> I say there clearly is a difference. I even gave you an analogy with
> numbers which can only be objectively true.
> 
> It seems like you're under the assumption that documentation can
> replace the need to do your own performance tests. They can't. There's
> no silver bullet and no easy way. You must always test things yourself
> and see whether the benefits are applicable to your situation, there's
> no need to blame the author for your own lack of research. As with
> pretty much everything in life, YMMV.
> 
> _______________________________________________
> libev mailing list
> libev at lists.schmorp.de
> http://lists.schmorp.de/cgi-bin/mailman/listinfo/libev