xml-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Eric Hodges <harmo...@swbell.net>
Subject RE: [spinnaker] Announce
Date Mon, 10 Jul 2000 20:44:13 GMT


> -----Original Message-----
> From: costin@eng.sun.com [mailto:costin@eng.sun.com]
> Sent: Monday, July 10, 2000 12:54 PM
> To: general@xml.apache.org
> Subject: Re: [spinnaker] Announce
>
>
> > Besides, if anyone (and people from Sun in particular since I expect
> > them to be the best experts in the matter) would help improving Xerces
> > performance over Hotspot it would be great!!!
>
> I think the main problem here is the lack of clean interfaces and modular
> structure in xerces, not the quality of hotspot.
>
> I think JDK1.1 optimizations are _great_, but should be used only with
> JDK1.1, and not with JDK1.3 ( IBM or Sun).

To me, this is a major failing on Sun's part.  For a long time, Sun's Java
compilers and VMs did a horrible job of optimizing code.  Simple
optimizations that most developers expect from a modern compiler were
missing.  But Java was so cool that we started to do the optimizations on
our own.

Now that the compiler and VM are doing what they ought to, our code is hurt
by our attempts to work around the missing optimizations.  We never wanted
to worry about optimization in the first place.

It's as if Sun thought we wouldn't use Java for the first 5 years.  Now they
want a do-over.  Sun can't expect the world to re-write all code for each
new release of the VM.  One of the selling points of Java is that code won't
have to be constantly re-written.

Why can't HotSpot figure out that it's slowing down execution and put itself
into -classic mode for specific classes?


Mime
View raw message