[x3d-public] agreement on number of \\\\\\\\ in Java???? X3DJSAIL.

[Michalis Kamburelis]

2017-07-29 22:23 GMT+02:00 Don Brutzman <brutzman at nps.edu>:
> Observation: it is quite important that MFString and SFString get
> represented consistently.  Author clarity and usage needs to be correct (of
> course).  Programmatic consistency is also necessary since SFString value(s)
> may be typecast or composed into an MFString value.

Consistency is an important design criteria, I absolutely agree. But
in the case of backslashes in X3D XML it's too late to make them
consistent now. That was the point of my message "Backslashes in X3D
XMLin SFString" to x3d-public on 2017-05-11,
http://web3d.org/pipermail/x3d-public_web3d.org/2017-May/006690.html .

- All the existing browsers (including my own) treat backslashes as a
special escape character in X3D XML MFString,
- and no browser treats backslash as anything special in X3D XML SFString.

So the implementations have already converged on a consensus, even
when the specification is imperfect. This behavior makes sense if you
consider a typical X3D XML implementation: in case of MFString you
need to parse the contents to split it, while in case of SFString you
just take the value of XML attribute. So it's completely natural  to
*not* treat backslash as "special" in case of X3D XML SFString.

And authors and conversion tools have already learned to cope with
this, as far as I know. So changing it now (or in X3D 4.0) would be
counter-productive, I think. Instead, the specification wording should
be adjusted to make it clear how (all) browsers behave. That's the
underlying theme in my corrections.txt, they not only make the
specification more correct / more precise, they also mostly describe
how existing browsers actually work already :)

Regards,
Michalis