Re: [xsl]Identifying patterns within texts

[Wendell Piez <wapiez@xxxxxxxxxxxxxxxx>]

Jim,

At 05:04 PM 11/29/2007, you wrote:
> The two responses I got (thank you....) reiterated the problem that I
> identified at the beginning of this project.  How do you identify
> "math"?  Since I am working on an educational tool where I am taking an
> old format (strings within xml tags)and converting to a new
> format(strings within new xml tags), it is tough to identify
> What is considered math.  1/2/99  Is that math or a date?  Basically, I
> am to decide what represents a mathematical expression and place it
> within its own element/tag.  Therefore, the software that processes it,
> will be able to display it in a 2-dimensional format....So the algorithm
> I come up with must be flexible and expandable.  And it may not be
> perfect.... Spaces between text will be a killer....  I guess what is
> acceptable will be up to the Systems guys......

This sort of squishiness isn't unusual for this kind of problem, if 
that's any comfort.

> However, I have went on the path to choose the XSLT implementation that
> was used on the GNOME project (xmlsoft.org) which implements XSLT 1.0.
> The engine I chose must be easily added to an existing dll and later
> ported to a MAC library (now that Mac is very much Unix :) ).  I needed
> something that was free, something the lawyers would approve the
> licensing, and something that would be portable among those two
> platforms.  I have seen some Java and C++ (Xalan with Xerces)
> implementations, but I did not want the added tasks of integration (JNI
> and C++ bindings).  Please comment on my logic if you see flaws.

If Java isn't a realistic option for you, the reasoning seems sound enough.

If Java is conceivable, you should at least consider Saxon8, which 
will give you all the 2.0 features and more. But I'm certainly not 
qualified to say whether it should be.

> Therefore, the idea of using xsl:analyze-string element or regular
> expressions in XSLT 2.0 is not an option right now.

If it isn't, then you basically have two choices:

1. Embrace the "fun" of setting out to be a killer XSLT 1.0 
programmer, devising various sorts of amazing trickiness in a 
language not designed for the task at hand. In this case I'd 
recommend getting a copy of Jeni Tennison's "XSLT and XPath On the 
Edge", which covers this sort of thing along with much else. You will 
become skilled in recursive templates, arcane tricks with the 
translate() function, and other sorts of madness.

If you want to stay sane, however ...

> I guess I could use a package like Boost/regex to post process my
> converted
> Xml.  I assume I can generate the XML from the result tree in memory and
> then parse that looking for math using C.

This sounds like a worthwhile option.

Just to make sure there isn't an intermediate course, you might 
investigate what sort of extensibility your processor of choice 
offers. Maybe you could manage the requirement by writing your own 
function library, and if you're lucky, maybe some of it has already 
been done for you.

Cheers,
Wendell


======================================================================
Wendell Piez                            mailto:wapiez@xxxxxxxxxxxxxxxx
Mulberry Technologies, Inc.                http://www.mulberrytech.com
17 West Jefferson Street                    Direct Phone: 301/315-9635
Suite 207                                          Phone: 301/315-9631
Rockville, MD  20850                                 Fax: 301/315-8285
----------------------------------------------------------------------
  Mulberry Technologies: A Consultancy Specializing in SGML and XML
======================================================================