Boris Zbarsky <[EMAIL PROTECTED]> wrote in message 
 
> I decided to look into the XForms implementations you pointed out.  I can't 
> download the Novell one, since they want accounts and such, but I looked at 
> X-Smiles and FormsPlayer.  The former has a download size of 8MB 
> (which actually 
> includes some docs), the latter has a download size of 2.2 MB.  To give XForms 
> the benefit of the doubt, I will focus on the smaller one.

FormsPlayer is closed source, so all we can infer from it's footprint
is what an upper bound would be for mozilla. No one can tell without
reviewing their code whether its the XForms spec or Novell's
implementation of it that causes the large footprint? Also, how could
we have any idea what percent of Novell's code duplicates existing
mozilla functionality?

I took a look at the X-Smiles source tree:
http://www.x-smiles.org/cvs.html
I found about 1.1MB of X-Smiles java code under
xsmiles-0.81/src/xsmiles/fi/hut/tml/xsmiles/mlfc/xforms

This represents about 20% of the whole X-Smiles source. The final
xsmiles.jar is about 2.8MB, so a conservative estimate would put the
xforms part at 600K. I think generally java jars are a bit bigger than
the corresponding binary  compiled from C++, but I don't have a good
number to use.

If we use this 600K number and your codesize comparisons, then we are
looking at something at most double the complexity of XPath and XSLT,
and probably less. Incidentally, the X-Smiles xfomrs code all seems to
be written by one guy.
_______________________________________________
mozilla-layout mailing list
[EMAIL PROTECTED]
http://mail.mozilla.org/listinfo/mozilla-layout

Reply via email to