| Newsgroups |
gmane.comp.embedded.carlsbad-cubes |
| Message-ID |
<[email protected]> |
I'm with Steve in feeling that using standard Java libraries is a better
choice than using non-standard, but possibly slightly better, packages.
Choosing the absolutely best library for data structures, regular
expressions, logging, etc where there is a reasonable standard one would
seem to yield many incremental benefits, but I believe that the result for
most projects is actually negative, with version issues, package footprint,
mental footprint, etc increasing with each "optimal" addition.
Altho I've switched most of my program to the standard packages, I've
sometimes just liked something so much that I didn't want to change. After
all, this should be fun. If Wolf is doing most of the work and JAXP makes
him unhappy, even tho Steve has provided an alternative, then I say
happiness is worth some price. But otherwise, stardard is better.
-- Fred
PS: Please don't argue ironic self-contradiction in that my pro-XUL approach
is non-standard.
Steven Pearson writes:
> Hi,
>
> I've modified SwiXML to use the JAXP parser shipped as standard with
> j2se1.4.2.
>
> My motivation was to reduce the deployment footprint if SwiXML by no longer
> needing the jdom.jar to be distributed.
>
> I implemented it using a factory so it's possible to switch between the
> normal JDOM parser and my JAXP parser.
>
> Is there any interest in such a modification in the SwiXML community?
>
> I've talked briefly with Wolf and he was concerned about performance so I
> ran a test with a 23Kb SwiXML file.
>
> Rendering the file 100 times with both JDOM and JAXP revealed
>
>>40% in javax.swing / java.awt
> 11% in various JAXP packages
> 7% in JDOM
>
> The actual time to render the document without profiling on average was ~200
> millis for both JDOM and JAXP.
>
> Thanks,
>
> Steve.
>
>
>
>
> _______________________________________________
> Forum mailing list
> [email protected]
> http://carlsbadcubes.com/mailman/listinfo/forum_carlsbadcubes.com