[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

XPath: yes. Which CC to use?



I have reviewed the XPath grammar and think that XPath is a worthy
addition to any
XML effort.

XPath appears to be a key utility in several of the other XML specs and
is a query language on its own.  It appears, as well, that it could be
some months before we see any more published work from the XQL working
group and that an XML query language will probably something along the
lines of  "XPath++"

I will be finishing up my current (for pay) project in the next 5 days
or so and I plan to spend a couple of weeks after that working (if
possible) on an XPath parser.

The grammar and specifications for XPath are considerably more
professional than the same for XQL.  (Every existing XPath
implementation I can find has some problem with it for Ozone's
purposes.)

I am considering which Java "compiler-complier" to use.  ANTLR would be
nice---but it doesn't support the required Unicode AT ALL.  JavaCC is
not open-source but *will* do the job, and is "freely available".
SableCC is very nice and is GPL'd although I am not sure how well its
going to be maintained in the future (SableCC is listed on the GNU site
and might work).

I can also hand-write a parsing engine---but this makes it more
difficult to alter and maintain for others in the future.  Under any
circumstances, the CC must be able to construct a "parse tree".

Speed of the parsing operation is less important than the ease of
construction and maintenance, for now.

Comments on the choice of a CC technology (wrt to the above) will be
greatly appreciated.

Once I have finished the parser, I should have an idea what Ozone API's
I will need to implement the XPath selections.  I suspect that this
effort will be useful in the larger effort to build a compliant XML
query language when the final spec gets done.  If not, XPath is still
useful on its own.

"Speak now or forever hold your peace."

Best Wishes,

Ann Tecklenburg
email:  at@ingenuity-sw.com