[jdom-interest] Internal DTD subset verification

O'neil, Jerome joneil at cobaltgroup.com
Thu May 9 09:25:44 PDT 2002


> But the I/O is relevant to anybody using JDOM who is going to have 
> read in the files form somewhere, possibly disk, possibly a network. 

I think that's Dennis' point.  I/O is going to be an equal factor regardless
of the source of the document.  Your disk and network are going to perform
as they will regardless of what API you are using. Using that as a measure
of JDOM (or any other API) speed would invalidate what you are testing, i.e.
build performance.  

> My claim is that in that more realistic scenario, JDOM performance is 
> a much smaller fraction of the total. Thus we really shouldn't be 
> wasting out time optimizing it, and we certainly shouldn't be 
> worrying about optimizing it at the expense of correctness.

I agree.  If runtime speed is that critical, most applications should be
using something else to begin with.  I'd much rather have it be right and
not-as-fast-as-X than have it be faster-than-thou and not guarantee
correctness.

--
Jerome O'Neil
Sr. Software Engineer
The Cobalt Group
206.219.8008
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://jdom.org/pipermail/jdom-interest/attachments/20020509/7c4ac4e7/attachment.htm


More information about the jdom-interest mailing list