Lists Home |
Date Index |
On Feb 21, 2005, at 8:39 AM, Elliotte Harold wrote:
> Kevin Jones wrote:
>> We clearly have very different objectives here. I think maybe at the
>> core is that a LCD model is always going to compromise performance. I
>> am tied of working with models that do this because they make my job
> As Donald Knuth taught us more than 30 years ago, "Premature
> optimization is the root of all evil in programming." I find it hard
> to believe that the conversion of an existing data model to a system's
> internal model is likely to be the performance bottleneck in an XQuery
> engine. I suspect this is true whether you use an LCD interface or the
> native interface.
Arguments based on belief and random speculation lead nowhere.
Perhaps you should rather concentrate on eliminating the quadratic time
complexity algorithms in your Jaxen XPath wrapper.
That way it might perhaps become useful for more than toy demos. Even
simple benchmarks will demonstrate the current fiasco.